Acidocalcisome is required for autophagy in Trypanosoma brucei

Abstract

Lysosomes play important roles in autophagy, not only in autophagosome degradation, but also in autophagy initiation. In Trypanosoma brucei, an early divergent protozoan parasite, we discovered a previously unappreciated function of the acidocalcisome, a lysosome-related organelle characterized by acidic pH and large content of Ca²⁺ and polyphosphates, in autophagy regulation. Starvation- and chemical-induced autophagy is accompanied with acidocalcisome acidification, and blocking the acidification completely inhibits autophagosome formation. Blocking acidocalcisome biogenesis by depleting the adaptor protein-3 complex, which does not affect lysosome biogenesis or function, also inhibits autophagy. Overall, our results support the role of the acidocalcisome, a conserved organelle from bacteria to human, as a relevant regulator in autophagy.

Introduction

Trypanosoma brucei (African trypanosome) is a blood-borne, flagellated protozoan parasite that causes sleeping sickness in humans and Nagana in cattle. A total of 50,000 to 70,000 new cases and ∼12,000 disabling and fatal cases were reported every year, threatening human health and agricultural economy in Africa.¹ Trypanosoma brucei has a complex life, proliferating in the midgut of tsetse fly (procyclic form) or the blood stream of mammals (bloodstream form). Each life stage exhibits different morphologies and catabolic or biosynthetic capacities, adapted to distinct nutrients availability, pH, and temperature of host environments.² Autophagy, the bulk degradation pathway to clear macromolecules or whole organelles through double-layered membrane-bound autophagosomes,³ is involved in the transformation⁴ and starvation responses⁵ in T. brucei. But the lack of detailed molecular signaling pathway precludes understanding the mechanisms and targeting this process for therapy.

The process of autophagy is complex and not completely understood. To date, more than 30 ATG (autophagy-related) genes have been identified,⁶ and orthologs to approximately half of these genes have been found in the T. brucei genome.⁷ In general, autophagy is triggered upon inhibition of the target of rapamycin (TOR), a conserved protein kinase.⁸ Phosphorylation and activation of ATG13 leads to complex formation with ULK1/2 (Atg1 in yeast),⁹ which further stimulates the formation of other protein complexes including phosphatidylinositol 3-kinase (PtdIns3K), whose catalytic subunit is termed PIK3C3 in mammals and Vps34 in yeast, PIK3R4 (Vps15 in yeast), BECN1/Beclin1 (Vps30/Atg6 in yeast) and ATG14. PIK3C3 produces phosphatidylinositol 3-phosphate (PtdIns3P) at certain cellular locations, which serves as a signal to recruit other proteins for phagophore formation.¹⁰ The phagophores elongate and engulf cytoplasmic components, forming double-membrane autophagosomes. Finally, autophagosomes are transported to and fused with lysosomes, where the engulfed materials are degraded.

In addition to degradation of autophagic materials, lysosomes are also required for autophagy initiation.¹¹ Lysosomal vacuolar type (V)-H⁺-ATPase is linked to the MTOR pathway in mammals, which acts upstream of autophagy induction. Autophagic activity is physiologically associated with the lysosome, through a link between the lysosome and V-H⁺-ATPase-LAMTOR-RRAG-MTORC1 signaling axis.^8,12-15 Suppression of lysosomal function triggers autophagy via MTORC1 downregulation;¹¹ during autophagy, MTORC1 suppression allows upregulation of lysosomal functions, thus accelerating autophagosome degradation.^16,17

In yeast and mammalian cells, bafilomycin A₁ (BafA₁), a selective V-H⁺-ATPase inhibitor,^18,19 suppresses not only the lysosome function, but also the MTORC1 activity,¹¹ thus leading to an increase in autophagosome number. In T. brucei however, BafA₁ completely inhibits autophagy, while chloroquine (CQ) blocks autophagy flux, leading to increased autophagosome number.⁵ In pursuit of the elucidation of this unusual effect of BafA₁ in trypanosomes, we discovered an unexpected function of the acidocalcisome in autophagy regulation. The results supported a strong correlation between acidocalcisome acidification and autophagy.

Results

Bafilomycin A₁ and monensin treatments inhibit autophagosome formation in trypanosomes

We have previously established procyclic T. brucei cell lines stably expressing BB₂-TbATG8.2 or YFP-TbATG8.2, which provide useful markers to monitor autophagy in fixed or live parasites.⁵ In yeast and mammalian cells, disruption of lysosome acidity using BafA₁, CQ, or monensin (an ionophore) prevents autophagosome fusion with lysosomes,^20-22 consequently inhibiting autophagic flux and resulting in an increase in autophagosome number.²¹ In T. brucei, treatment with CQ leads to an increase in average autophagosome number per cell,⁵ consistent with its role in blocking autophagosome fusion to lysosomes found in other organisms. Interestingly, both BafA₁ and monensin treatments completely inhibited starvation-induced autophagosome formation in T. brucei (Fig. 1A and B), displaying effects distinct to those observed in other eukaryotes. Since the V-H⁺-ATPase that is required to maintain lysosome acidity²³ is also found present on the acidocalcisomes in T. brucei,²⁴ it raised an interesting possibility that both lysosomes and acidocalcisomes play a role in autophagy.

Figure 1.

Both BafA₁ and monensin inhibit starvation-induced autophagy. (A) Cells stably expressing BB₂-TbATG8.2 were cultured in medium, or starved in gHBSS in the absence or presence of 100 nM BafA₁ or 10 μM monensin for 2 h. Cells were fixed, immunolabeled with anti-BB₂ (red), and stained with DAPI for DNA (blue). Scale bar: 5 μm. (B) The average number of autophagosomes per cell was quantified for cells starved in the absence or presence of BafA₁ and monensin. Values denote means ± s.d. (n = 3). At least 200 cells were counted for each condition.

Starvation and vanadate-induced autophagy is accompanied by acidification of acidocalcisomes

To monitor acidocalcisomes and lysosomes during starvation-induced autophagy, control and starved cells were labeled with antibodies specific to acidocalcisomes (anti-TbVP1)²⁵ or lysosomes (anti-trypanopain);²⁶ as well as pH-sensitive dyes LysoTracker Red and fluorescently labeled CQ (BODIPY-CQ or coumarin-CQ), which have been shown previously to accumulate in acidic organelles such as lysosomes and vacuoles.^25,27,28 Upon starvation, the size and number of acidocalcisomes were not perceptibly changed as compared with control cells according to anti-TbVP1 staining (Fig. S1A). The lysosome marker trypanopain however, stained a larger region between the nucleus and the kinetoplast (mitochondrial DNA aggregate), consistent with accumulation of autophagic materials for degradation (Fig. S1B).

Interestingly, under normal cultivation conditions, both LysoTracker Red and fluorescent CQs were found to stain punctate structures inside of the cells, colocalizing with acridine orange (AO) that stains acidocalcisomes²⁴ (Fig. S2A). Staining by these pH-sensitive dyes was likely due to the acidic nature of acidocalcisomes. Interesting still, the staining of BODIPY-CQ, AO or LysoTracker Green in the acidocalcisome puncta enhanced significantly upon starvation (Fig. 2A), suggesting an increase in acidocalcisome acidity in starved cells.

Figure 2.

Acidocalcisome pH decreases upon autophagy induction. (A) Cells cultivated in medium or starved in gHBSS were stained with 6 μM AO, 10 μM LysoTracker Green or 2 μM BODIPY-CQ. All 3 pH trackers showed enhanced staining in acidocalcisome puncta upon starvation. For each fluorescent pH tracker, control and starved cells were stained and imaged using identical conditions. All fluorescence images shown were acquired using 1,000 ms exposure, whereas the inset shown under BODIPY-CQ treatment was acquired using 100 ms exposure. (B, C) Whereas vanadate and starvation both induced acidocalcisome acidification, BafA₁ and monensin treatments inhibited acidocalcisome acidification upon starvation. Cells were incubated in medium with or without 1 mM vanadate, or starved in gHBSS with or without 1 mM vanadate, 100 nM BafA₁ or 10 μM monensin, respectively for 2 h, then stained with BODIPY-CQ for another 30 min. BODIPY-CQ accumulation was monitored by fluorescence microscopy (100 ms exposure, B) and flow cytometry (C). The inset shown under vanadate treatment during starvation was acquired using 20 ms exposure. Scale bar: 5 μm. (D, E) Vanadate treatment induces autophagy in medium. YFP-TbATG8.2 or BB2-TbATG8.2 cells were incubated in medium or starved in gHBSS for 2 h, in the absence and presence of 1 mM vanadate. Cells were fixed and autophagosome formation was monitored by fluorescence microscopy (D). Scale bar: 5 μm. The average number of autophagosomes per cell was quantified and shown as mean ± s.d. (n = 3) (E). At least 200 cells were counted for each condition.

To further verify that the pH-tracker positive punctate structures were indeed acidocalcisomes, LysoTracker Red staining was performed on cells stably expressing YFP-TbATG8.2 (that labeled the autophagosomes) or YFP-p67 (YFP fusion to a T. brucei lysosomal protein;²⁹ Fig. S2B). Note that BODIPY-CQ and AO could not be used in this case, due to their overlapping emission wavelength with YFP. Upon starvation, the punctate structures stained by LysoTracker Red did not colocalize with the autophagosomes but did overlap in part with the lysosome structure (Fig. S2C and D).

Furthermore, the adaptor protein-3 (AP-3) complex that is essential for acidocalcisome biogenesis in T. brucei,³⁰ was ablated by inducible and inheritable RNA interference (RNAi) of either the β3 (Tbβ3, Tb927.11.10650) or the δ (Tbδ, Tb927.5.3610) subunits (Fig. S3A). Consistent with previously reported,³⁰ depletion of either subunits resulted in significant reduction in the number of acidocalcisomes (Fig. 3A), though only moderately affected cell growth at 72 h or 96 h postinduction (Fig. S3B). Also consistent with the previous report that Tbβ3- and Tbδ-RNAi did not affect lysosome formation and function,³⁰ the presence of lysosomes was confirmed by staining with anti-trypanopain²⁶ (Fig. 3B). Upon starvation, no accumulation of LysoTracker Red, BODIPY-CQ or AO was observed in either Tbβ3- or Tbδ-depleted cells (Fig. 3C). Weak labeling in a single punctate structure could be observed in some cells (Fig. 3C, arrows), possibly due to staining of lysosomes. This observation further attested the nature of the pH tracker-positive puncta as acidocalcisomes. Lysosomes, on the other hand, had little or only minor contribution to pH-tracker accumulation under starvation conditions.

Figure 3.

Blocking the biogenesis of acidocalcisomes inhibits autophagy. (A, B) Depletion of either of the AP-3 subunits, β3 and δ, inhibited acidocalcisome but not lysosome biogenesis. Cells uninduced (tet-) and induced for TbAP-3 RNAi for 72 h (tet+) were fixed with 4% PFA and immuno-stained with anti-TbVP1 antibody for acidocalcisomes (A) or anti-trypanopain for lysosomes (B). Scale bar: 5 μm (C) Depletion of AP-3 subunits β3 or δ, inhibited LysoTracker Red, BODIPY-CQ or AO accumulation in starved cells. Arrows, residual staining by pH-trackers in acidocalcisome-depleted cells. Scale bar: 5 μm. (D and E) AP-3 depletion inhibits autophagosome formation. Starvation-induced autophagosome formation was imaged (D) and quantified (E) in control and induced cells. Scale bar: 5 μm. The average number of autophagosomes per cell was quantified for at least 200 cells for each condition and shown as means ± s.d. (n = 3).

It was interesting to note that BODIPY-CQ and coumarin-CQ did not act the same as CQ. Whereas both fluorescent CQs accumulated in the acidocalcisomes, CQ may act on lysosomes thereby inhibiting autophagic flux.⁵ Preincubation with coumarin-CQ greatly inhibited BODIPY-CQ accumulation in starved cells, but preincubation with CQ did not affect BODIPY-CQ accumulation (Fig. S4A). This result suggested that the fluorescence labeling may indeed have altered the property of the original CQ, allowing specific accumulation of the fluorescent CQs into the acidocalcisomes.

The accumulation of BODIPY-CQ during the course of starvation was also monitored in live cells by flow cytometry. Significant BODIPY-CQ accumulation was observed as early as 5 min after starvation, and acidocalcisome acidification maximized in 15 min (Fig. S4B). Acidocalcisome acidification thus appeared to precede autophagosome formation, which starts ∼30 min after starvation.⁵

In addition to starvation, autophagy can also be induced with 1 mM sodium orthovanadate in procyclic T. brucei (Fig. 2D and E). Similar to starvation-induction, increased BODIPY-CQ accumulation and acidocalcisomes acidification were also observed upon vanadate treatments (Fig. 2B and C), further supporting the connection between autophagy and acidocalcisome (and perhaps also lysosome to a lesser extent) acidification.

Treatments that result in acidocalcisome alkalinization lead to autophagy inhibition

As both BafA₁ and monensin inhibited starvation-induced autophagic activity (Fig. 1A and B), their effects on BODIPY-CQ accumulation and acidocalcisome acidification were also examined. Both BafA₁ and monensin treatments abolished BODIPY-CQ accumulation in the acidocalcisomes, as shown by fluorescence microscopy or flow cytometry (Fig. 2B and C). This was not a BODIPY-CQ-specific effect as monensin treatments also blocked AO and LysoTracker Red accumulation in the acidocalcisomes during starvation (data not shown), suggesting that the inhibitory effects of BafA₁ and monensin on autophagy may be due to their inhibition of acidocalcisome (and/or lysosome) acidification.

A close relationship between acidocalcisome acidification and Ca²⁺ uptake has been well established in T. brucei.³¹ Next we tried to modulate the acidocalcisomal pH through alteration of the acidocalcisome-stored Ca²⁺ flux. In T. brucei, ionomycin alone releases a relatively small amount of Ca²⁺, but significantly more Ca²⁺ is released from acidocalcisomes when ionomycin is added together with either nigericin or NH₄Cl.²⁵ Compared with starvation control, NH₄Cl only slightly reduced the acidification of acidocalcisomes. Ionomycin, however, drastically inhibited acidification. Treatment with both ionomycin and NH₄Cl further alkalinized the acidocalcisomes (Fig. 4A). Consistently, ionomycin and NH₄Cl treatment inhibited autophagosome formation under starvation conditions, and reduced autophagosome number per cell from 1.15 puncta/cell to 0.58 (ionomycin treatment), 0.94 (NH₄Cl treatment), and 0.23 (both ionomycin and NH₄Cl treatment) (Fig. 4B). The effects of ionomycin and/or NH₄Cl on acidocalcisome acidification correlated precisely with their inhibitory effects on autophagy.

Figure 4.

NH₄Cl and ionomycin inhibit acidocalcisome acidification and autophagy. (A) Cells cultivated in medium or starved for 2 h in the absence or presence of 1 μM ionomycin, 10 mM NH₄Cl or both, were stained with BODIPY-CQ and analyzed by flow cytometry. (B) Quantification of autophagosomes in BB₂-TbATG8.2 cells cultivated in medium or staved in gHBSS for 2 h, in the absence or presence of 1 μM ionomycin (ion), 10 mM NH₄Cl, or both. *P < 0.05, **P < 0.01, relative to control. The average number of autophagosomes per cell was shown as mean ± s.d. (n = 3). At least 200 cells were counted for each condition.

Inhibition of acidocalcisome biogenesis blocks starvation-induced autophagy

The specific involvement of acidocalcisomes in autophagy was further evaluated in cells deficient in acidocalcisome biogenesis. In both Tbβ3 and Tbδ-depleted cells, autophagosome formation was inhibited from 2.02 ± 0.23 (Tbβ3-RNAi control) to 0.44 ± 0.11 (Tbβ3-RNAi induced) or from 1.83 ± 0.31 (Tbδ-RNAi control) to 0.36 ± 0.14 (Tbδ-RNAi induced) autophagosomes/cell (Fig. 3D and E). These results indicated that the presence of acidocalcisomes is essential for starvation-induced autophagy in T. brucei.

Starvation-induced acidocalcisome acidification is V-H⁺-ATPase and V-PPase-independent

Proton pumps are responsible for the generation of vacuolar H⁺-gradient or proton motive force. Similar to the plant vacuole,³² trypanosome acidocalcisomes contain at least 2 proton pumps, V-H⁺-ATPase and V-PPase.^24,25 The V-H⁺-ATPase is a multisubunit complex comprising of a V₀ sector that forms the proton pore and a V₁ sector that mediates ATP hydrolysis.³³ T. brucei genome encodes all the subunits found in the yeast V-H⁺-ATPase (Fig. S5). Depletion of subunit α (TbVPH1, Tb927.5.1300) of the V₀ sector or subunit A (TbVMA1, Tb927.4.1080) of the V₁ sector (Fig. S6A), both inhibited cell proliferation at 48 h postinduction (Fig. S6B). Both depletions also led to decreased LysoTracker Red accumulation in the acidocalcisomes, suggesting alkalinization of the acidocalcisomes under normal growth condition (Fig. 5A, left panel). Acidocalcisome acidification upon starvation, however, was not affected (Fig. 5A, right panel). The average number of autophagosomes per cell increased from 1.86 ± 0.19 (TbVPH1-RNAi control) to 3.1 ± 0.32 (TbVPH1-RNAi induced) and from 1.73 ± 0.23 (TbVMA1-RNAi control) to 4.04 ± 0.69 (TbVMA1-RNAi induced) (Fig. 5B). These results suggested that while the V-H⁺-ATPase functions in maintaining acidocalcisome acidity under normal cultivation conditions, it is however, not required for acidocalcisome acidification or autophagosome formation upon starvation.

Figure 5.

Starvation-induced acidocalcisome acidification is V-H⁺-ATPase- and V-PPase independent. (A) Depletion of V-H⁺-ATPase subunit TbVPH1 and TbVMA1 neutralized acidocalcisomes in medium, but did not inhibit starvation-induced acidification. Uninduced control and cells induced for 72 h were cultured in medium or starved in gHBSS for 2 h and then stained with LysoTracker Red. LysoTracker Red staining was monitored by fluorescence microscopy with 1,000 ms (cells in medium) or 100 ms (starved cells) exposure. Scale bar: 5 μm. (B) V-H⁺-ATPase depletion increased autophagosome number in starved cells. Autophagosome formation was monitored with YFP-TbATG8.2 stably expressed in the V-H⁺-ATPase RNAi-treated cells (lower panels). The average number of autophagosomes per cell was quantified and shown as means ± s.d. (n=3). At least 200 cells were counted for each condition. (C to E) TbVP1 depletion alkalinizes acidocalcisomes in medium, but does not affect acidocalcisome acidification and autophagosome formation in starved cells. Control and cells treated with TbVP1 RNAi for 72 h were monitored for anti-TbVP1 immunostaining (C), LysoTracker Red accumulation in acidocalcisomes (D) and autophagosome formation (E). Scale bars: 5 μm. (F to H) Double depletion of either V-H⁺-ATPase subunits with TbVP1 does not inhibit starvation-induced acidification of acidocalcisomes, or starvation-induced autophagosome formation. Control and double-RNAi-treated cells were monitored for LysoTracker Red accumulation in acidocalcisomes by fluorescence microscopy (F) or flow cytometry (G), and autophagosome formation (H). Scale bar: 5 μm. The average number of autophagosomes per cell was quantified and shown as means ± s.d. (n = 3).

Depletion of the other proton pump, V-PPase (or TbVP1) by inducible RNAi (Fig. 5C) did not affect cell proliferation (Fig. S6C) as described previously.²⁵ Similar to V-H⁺-ATPase, the TbVP1 depletion alkalinized the acidocalcisomes in the medium but had no effect on starvation-induced acidocalcisome acidification (Fig. 5D) or autophagosome formation (Fig. 5E). To exclude the possibility that V-H⁺-ATPase and V-PPase may play redundant roles to control the acidification of the acidocalcisomes upon starvation, we performed double RNAi of TbVP1 with TbVPH1 or TbVMA1 (Fig. S6D and E). Again, the double depletions did not block starvation-induced acidocalcisome acidification or autophagosome formation, shown by fluorescence microscopy or by flow cytometry (Fig. 5F to H). Furthermore, to rule out the possibility that V-H⁺-ATPase and V-PPase expression may be subject to regulation under starvation conditions thus offseting the RNAi effects, the expression of V-H⁺-ATPase and V-PPase were examined in cells before and after starvation using immunoblots (for V-PPase with anti-TbVP1) or reverse transcriptase (RT) –PCR (for V-H⁺-ATPase due to the lack of specific antibody). Neither V-H⁺-ATPase nor V-PPase exhibited significant upregulation upon starvation at the protein or mRNA level, respectively (Fig. S6F and G). It should be noted that these methods could not distinguish if V-H⁺-ATPase or V-PPase activity (not necessarily expression levels) may be upregulated upon starvation. Nevertheless, these results suggested that starvation-induced acidocalcisome acidification and autophagosome formation may require mechanisms other than V-H⁺-ATPase and V-PPase.

PtdIns3P, the product of TbVPS34, associates with acidocalcisomes upon starvation

In our previous work, we have found that wortmannin, an inhibitor of Vps34 (the catalytic subunit of the class III phosphotidylinositol 3-kinase, or PtdIns3K) that produces PtdIns3P and plays an essential role in autophagy in yeast and mammalian cells,³⁴ also inhibits starvation-induced autophagy in T. brucei.⁵ The precise role of Vps34 in autophagy is still not clear. However in Candida albicans, Vps34 interacts with Vma7, a subunit of the V-H⁺-ATPase. Vps34 knockout leads to vacuolar acidification defects due to the defective proton transport into the vacuole.³⁵ T. brucei cells treated with wortmannin, in medium or upon starvation, did not exhibit detectable changes in acidocalcisome acidification (Fig. S7A and B), suggesting a postacidification function of PtdIns3K. Using a PtdIns3P-specific probe GFP-2×FYVE, PtdIns3K was localized to punctate structures in procylic T. brucei cultivated in medium. This labeling pattern is distinct to PtdIns3P localization in the endosomal system in bloodstream-form trypanosomes.³⁶ It is also different from the endosome labeling previously observed by Alexa Fluor 488 dextran uptake in procyclic T. brucei.³⁷ In addition, GFP-2×FYVE did not colocalize with glycosomes, lysosomes, Golgi, ER, or acidocalcisomes, as shown by immunofluorescence studies (Fig. S8 and Fig. 6A). However, upon starvation, a subset of GFP-2×FYVE positive puncta became associated with acidocalcisomes. This association could be disrupted by BafA₁ treatment, which inhibited autophagy (Fig. 6A and B). GFP-2×FYVE did not colocalize with autophagosomes upon starvation (Fig. 6C), which was consistent with our previous observation that autophagosomes do not colocalize with acidocalcisomes.⁵ It is tempting to speculate that TbVPS34 relocalized to acidified acidocalcisomes and produced PtdIns3P, which further initiated the phagophore formation and autophagy process. No phagophore marker has yet been characterized in T. brucei. A single T. brucei ortholog to phagophore marker ATG5 (TbATG5)³⁸ (GeneID:Tb927.6.2430) with 31.9% and 24.8% similarity to Leishmania³⁹ or yeast Atg5, respectively, was identified in the T. brucei genome, fused to YFP, and stably expressed (Fig. S9). In cells cultivated in medium, both YFP-TbATG5 and TbATG5-YFP exhibited cytoplasmic distribution. However, upon starvation, neither showed relocalization to punctate structures, suggesting that the TbATG5-YFP fusions could not be used as a reporter for phagophore formation. Thus a direct link between acidocalcisome-resident PtdIns3P and phagophore formation could not be yet established.

Figure 6.

A pool of PtdIns3P localizes to acidocalcisomes upon starvation. (A) Cells expressing GFP-2×FYVE were cultivated in medium or starved in gHBSS with or without 100 nM BafA₁ for 2 h, then immunostained with anti-TbVP1 for acidocalcisomes (red). GFP-2×FYVE became associated with TbVP1-positive acidocalcisomes upon starvation (arrowheads), and BafA₁ inhibited this association. Scale bar: 5 μm. (B) The percentage of GFP-2×FYVE puncta associated with acidocalcisomes was quantified and shown as box plot. (C) TbVPS34 was not associated with the autophagosomes. BB₂-TbATG8.2 cells expressing GFP-2×FYVE were cultivated in medium or starved in gHBSS for 2 h. The cells were immunostained with anti-BB₂ (red) to monitor autophagosome formation and their localization relative to TbVPS34-positive puncta (green). Scale bar: 5 μm.

Discussion

In this study, we discovered an unexpected role of acidocalcisomes in T. brucei autophagy regulation. When the adaptor protein-3 complex was depleted to specifically block acidocalcisome biogenesis but not lysosomes,³⁰ starvation-induced autophagy was completely inhibited. Furthermore, autophagy showed a strong correlation with acidocalcisome acidification, using BODIPY-CQ, LysoTracker Red or Green, or AO as acidocalcisome pH indicators. Upon starvation or vanadate treatment to induce autophagy, the pH-sensitive dyes accumulated further in acidocalcisomes, suggesting acidocalcisome acidification. Alkalinization of acidocalcisome with BafA₁, monensin or ionomycin, and NH₄Cl treatments, all inhibited autophagy. Lysosomes, the other known acidic organelle in T. brucei, played only a minor role in pH-tracker accumulation in starved procyclic cells.

Surprisingly, RNAi depletion of neither V-H⁺-ATPase nor V-PPase showed an inhibitory effect on T. brucei autophagy. Whereas both proton pumps contributed to the maintenance of acidocalcisome acidity under normal growth conditions, neither was required for acidocalcisome acidification upon starvation. In the case of V-H⁺-ATPase-depletion cells, autophagosome number even increased upon starvation. It is possible that depletion of V-H⁺-ATPase, which is also required to maintain lysosome acidity in trypanosomes,²³ affected lysosome function and thus blocked autophagosome degradation via lysosomes. These results were further confounded by the observation that the V-H⁺-ATPase inhibitor BafA₁ efficiently suppressed autophagy. Whereas it is possible that BafA₁, a specific V-H⁺-ATPase inhibitor in many organisms⁴⁰ may act upon other targets in the highly divergent T. brucei; a V-H⁺-ATPase/V-PPase-independent pathway may be present in T. brucei and facilitate acidocalcisome acidification under starvation conditions.

Besides proton pumps, there is biochemical evidence for the presence of Ca²⁺/H^{+ 31} and Na⁺/H⁺ antiporters^41,42 on the acidocalcisomes, which can regulate ions and H⁺ homeostasis. Additionally, T. brucei acidocalcisomes contain high concentration of free amino acids, of which 90% are basic amino acids arginine and lysine^43,44 and a high content of pyrophosphate (PPi) and polyphosphate (polyP).⁴⁵ The PPi and polyP are present in all kingdoms of life and function as direct energy donors or regulators of stress and survival.⁴⁶ Whether PPi or polyP metabolism has a role in autophagy regulation is also of interest. Further molecular characterization of acidocalcisome metabolic activities and membrane transporters, particularly under stress conditions, will be crucial to understanding its cellular functions.

Another puzzling observation is the lack of direct association between the acidocalcisomes and autophagosomes, at both light microscopy and electron microscopy levels.⁵ Using GFP-2×FYVE as a probe, a population of PtdIns3P was found associated with the acidocalcisomes upon starvation and acidocalcisome acidification. Inhibition of acidocalcisome acidification by BafA₁ also inhibited PtdIns3P association with acidocalcisomes, whereas inhibition of PtdIns3P production by wortmannin inhibited autophagy but did not affect acidocalcisome acidification. Together, these observations raised an interesting possibility that upon acidocalcisome acidification, a pool of PtdIns3P became associated with acidocalcisomes and initiated phagophore formation. Verification of this possibility requires characterization of T. brucei phagophore markers, which has not been reported yet.

Overall, our study provides evidence supporting acidocalcisome acidification as a key regulator of autophagy initiation in T. brucei. We have shown that the presence of the acidocalcisomes is essential for autophagosome formation. Furthermore, acidocalcisome acidification strongly correlates with autophagy induction, though the acidification is V-H⁺-ATPase and V-PPase independent (Fig. 7). Combining with the observations that reducing lysosome acidity inhibits autophagosome degradation, we propose a model of autophagy regulation in trypanosomes at different stages through acidocalcisomes and lysosomes (Fig. 7). Under normal growth condition, V-H⁺-ATPase and/or V-PPase are involved to maintain the acidity of acidocalcisomes and lysosomes. Upon starvation, the acidity of acidocalcisomes was further increased via an unknown, V-H⁺-ATPase and V-PPase-independent mechanism. Following acidification, a pool of PtdIns3P, the products of PtdIns3K, were transported to the vicinity of acidocalcisomes and may initiate phagophore formation via an unknown mechanism. The phagophore matured to an autophagosome and finally was transported to and fused with a lysosome to form the autolysosome, in which the engulfed materials were degraded. V-H⁺-ATPase present on the lysosomes may be involved in regulation of this final step. Whether the acidocalcisome-mediated regulation is specifically evolved in autophagy regulation in T. brucei remains to be examined. Given that the acidocalcisomes and similar structures are ubiquitously found from bacteria to human,⁴⁴ this may also represent a conserved mechanism in other eukaryotes.

Figure 7.

A model of autophagy regulation in T. brucei. Under normal growth conditions, V-H⁺-ATPase, which is found present on both acidocalcisomes and lysosomes, was required to maintain their acidity. V-PPase was additionally required to maintain acidocalcisome acidity. Upon starvation, the acidity of acidocalcisomes was further enhanced, via an unknown, V-H⁺-ATPase and V-PPase independent mechanism. Following the acidification, a pool of PtdIns3P was transported to the area of acidocalcisomes and furthered phagophore and autophagosome formation. Mature autophagosomes then fuse with lysosomes to form the autolysosome where the engulfed materials are degraded, as has been proposed in other organisms.

Materials and Methods

Cell culture, drugs, and chemicals

All experiments were performed on insect stage, procyclic T. brucei YTat1.1 and 29.13 cells (wild type or expressing YFP-TbATG8.2), that were maintained in Cunningham's medium.⁴⁷ Bafilomycin A₁ (100 nM; B1793), monensin (10 μM; M5273), wortmannin (2 μM; W1628), ionomycin (1 μM; 10634), acridine orange (6 μM; A6014) and sodium orthovanadate (Na₃VO₄, 1 mM; S6508) were purchased from Sigma. Coumarin- and BODIPY-labeled chloroquines (coumarin-CQ and BODIPY-CQ) were purchased from BioLynx (BL004RUO and BL005RUO, respectively). LysoTracker Green, Red and Deep Red (10 μM, L7526, L7528 and L12492, respectively) were purchased from Invitrogen. All other chemicals were purchased from Sigma. Antibodies used in this study are shown below.

Plasmid construction and transfection

For ectopic expression of YFP-tagged lysosome-marker p67,²⁹ full length cDNA coding for T. brucei p67 (Tb927.5.1810) was amplified with primers listed in Table S1, digested with BamHI and cloned into pHD1034 vector. For stable transfection, DNA construct was linearized with NotI and transfected into YTat1.1 cells. The stable transformant was selected with 10 μg/ml puromycin (Sigma, P8833).

To express the N- or C-terminal YFP-tagged TbATG5, the full-length cDNA (Tb927.6.2430) was amplified with primers listed in Table S1, digested with BamHI and EcoRI, or with HindIII and NheI, respectively, then subcloned into pXS2-YFP vector. For stable transfection, the constructs were linearized with MluI and transfected into YTat1.1 cells. The stable transformants were selected with 10 μg/ml blasticidin (Invitrogen, R210-01).

To knock down Tbβ3, Tbδ, TbVPH1, TbVMA1, or TbVP1 by inducible and inheritable RNA interference (RNAi) in trypanosomes, suitable fragment were selected using RNAit (http://trypanofan.path.cam.ac.uk/software/RNAit.html) to avoid off-target effects. The selected fragment was amplified with the primers listed in Table S1. The PCR fragment was digested with XbaI and subcloned into p2T7 vectors.⁴⁸ For double knockdown of TbVPH1 and TbVP1 or TbVMA1 and TbVP1, the TbVP1 fragment was amplified with primers containing HindIII restriction sites (Table S1), then subcloned into the p2T7-TbVPH1 and p2T7-TbVMA1 vectors. For stable transfections, DNA constructs were linearized with NotI, and electroporated into 29.13 or cells stably expressing YFP-TbATG8.2 or BB₂-TbATG8.2.⁵ Stable transformants were selected with 10 μg/ml phleomycin (Sigma, P9564).

To track TbVPS34 localization, the pHD1034-GFP-2×FYVE vector was linearized with NotI and transfected into procyclic YTat1.1 cells and stable transformants selected with 10 μg/ml puromycin.

Starvation and drug treatments

For starvation, T. brucei cells were washed once with gHBSS (137 mM NaCl, 5.4 mM KCl, 0.25 mM Na₂HPO₄, 0.44 mM KH₂PO₄, 1.3 mM CaCl₂, 1 mM MgSO₄, 4.2 mM NaHCO₃, 1 g/L glucose, pH 7.3),⁵ and then resuspended in gHBSS at a concentration of 5 × 10⁶ cells/ml and incubated at 27°C for 2 h unless otherwise indicated. To monitor drug effects on autophagy, cells stably expressing YFP-TbATG8.2 or BB₂-TbATG8.2 ⁵ were pretreated with different drugs (with final concentrations indicated under Cell culture, drugs and chemicals) for 30 min, washed once with gHBSS, followed by starvation in the presence of drugs for another 2 h. The formation of autophagosomes was monitored by fluorescence microscopy.

Monitoring the acidification of acidocalcisomes

For wild-type or cells stably expressing BB₂-TbATG8.2, 2 × 10⁶/ml control or drug-treated cells were loaded with 2 μM coumarin-CQ or BODIPY-CQ, 6 μM AO, or 10 μM LysoTracker Green or Red for 30 min at 27°C. The cells were washed once with serum-free medium or gHBSS, with or without drugs. The staining in live cells was then immediately evaluated with an Axioplan2 inverted fluorescence microscope (Carl Zeiss MicroImaging, Germany) equipped with a CoolSNAP HQ² camera (Photometrics, USA) and a Plan-Apochromat 63×/1.40 oil DIC objective using DAPI (for coumarin-CQ staining), eGFP (for BODIPY-CQ and LysoTracker Green staining) or dsRed (for AO and LysoTracker Red staining) channels. For colocalization studies, live cells stained with a single dye were also used to optimize imaging conditions and eliminate bleed-through effects. Flow cytometry analyses were performed on a Guava® EasyCyte Plus flow cytometer (Millipore, USA) with 488 nm excitation and 515 to 545 nm emission (for BODIPY-CQ staining) on live T. brucei cells. For YFP-TbATG8.2-expressing cells, 10⁶ cells were loaded with 10 μM LysoTracker Deep Red for 30 min, washed once, and the staining was evaluated by a FACSAria II flow cytometer (BD Biosciences, USA) with 633 nm excitation and 650 to 670 nm emission. To quench the acidity of acidocalcisomes, cells were pretreated with 20 μM coumarin- or BODIPY-CQ for 1 h.

Fluorescence microscopy

Trypanosoma brucei cells were washed and resuspended in serum-free medium or gHBSS with or without drugs, and attached to coverslips. Cells were fixed with 4% paraformaldehyde (PFA, Sigma, P6148) in phosphate-buffered saline (PBS, 1^st BASE, BUF-2040), pH 7.4 at room temperature, permeabilized with 1% NP-40 (Igepal CA-630, Sigma, I8896) and blocked with 3% BSA (Sigma, A7906) before antibody staining. Monoclonal anti-BB₂ ⁴⁹ was used at 1:100 dilution to label BB₂-TbATG8.2. Polyclonal anti-TbVP1²⁵ was used at 1:500 to stain the acidocalcisomes. Polyclonal anti-trypanopain²⁶ was used at 1:2,000 to stain the lysosomes. Polyclonal anti-TbGRASP⁵⁰ was used at 1:1,000 to stain the Golgi. Polyclonal anti-SKL⁵¹ and anti-TbBiP⁵² were used at 1:800 to stain the glycosomes and ER, respectively. Nuclear and kinetoplast DNA was stained with 2 μg/ml DAPI (Invitrogen, D1306). Fixed cells were observed using an Axioplan2 inverted microscope (Carl Zeiss MicroImaging, Germany) equipped with a CoolSNAP HQ² camera (Photometrics, USA) and a Plan-Apochromat 63×/1.40 oil DIC objective. For colocalization analyses, the cells were observed using a laser scanning confocal microscope 510 META (Carl Zeiss MicroImaging, Germany) equipped with a Plan-Apochromat 63×/1.40 oil DIC objective.

RT-PCR

The total RNA was extracted from 5 × 10⁷ control or treated cells with TRIzol Reagent (Life Technologies, 15596-026) following the provided protocol. After treatment with RNase-free DNase I (Roche, 04 716 728 001), the first-strand cDNA was synthesized with M-MLV Reverse Transcriptase (Invitrogen, 28025-013) using the standard protocol and the PCR was performed with the primers listed in Table S1.

Western blotting

Cell lysates were prepared from 5 × 10⁶ control or starved cells, then separated using 12% SDS-PAGE gels. The proteins were immobilized onto PVDF membrane (Bio-Rad, 162–0177) and probed with anti-TbVP1 or anti-PFR1⁵³ as loading control. Following incubation with appropriate secondary antibodies conjugated to horseradish peroxidase (Bio-Rad, 170–6015 for goat-anti-rabbit IgG and Sigma, A9037 for goat-anti-rat IgG), the blots were processed using the Clarity Western ECL system (Bio-Rad, 170–5060) and documented by ImageQuant LAS 4000 mini Luminescent Image Analyzer (GE, USA).

Statistical analyses

For autophagosome quantification, 3 independent experiments were performed and statistical analyses were performed using the 2-tailed, equal variance Student t test. P values <0.05 and <0.01 were determined to be statistically significant (*) and highly significant (**), respectively.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Funding

This study is supported by the National Research Foundation of Singapore (NRF-RF001–121).

Supplemental Material

Supplemental data for this article can be accessed on the publisher's website.