Rab32 is essential for maintaining functional acidocalcisomes and for growth and infectivity of Trypanosoma cruzi

Page 1

RESEARCH ARTICLE

Rab32 is essential for maintaining functional acidocalcisomes,and for growth and infectivity of Trypanosoma cruziSayantanee Niyogi1, Veronica Jimenez1,*, Wendell Girard-Dias2, Wanderley de Souza2,3, Kildare Miranda2,3

and Roberto Docampo1,‡

ABSTRACTThe contractile vacuole complex (CVC) of Trypanosoma cruzi, theetiologic agent of Chagas disease, collects and expels excess wateras a mechanism of regulatory volume decrease after hyposmoticstress; it also has a role in cell shrinking after hyperosmotic stress.Here, we report that, in addition to its role in osmoregulation, the CVCof T. cruzi has a role in the biogenesis of acidocalcisomes.Expression of dominant-negative mutants of the CVC-located smallGTPase Rab32 (TcCLB.506289.80) results in lower numbers of less-electron-dense acidocalcisomes, lower content of polyphosphate,lower capacity for acidocalcisome acidification and Ca2+ uptakethat is driven by the vacuolar proton pyrophosphatase and theCa2+-ATPase, respectively, as well as less-infective parasites,revealing the role of this organelle in parasite infectivity. By usingfluorescence, electron microscopy and electron tomographyanalyses, we provide further evidence of the active contact ofacidocalcisomes with the CVC, indicating an active exchange ofproteins between the two organelles.

KEY WORDS: Acidocalcisome, Calcium, Contractile vacuole,Polyphosphate, Vacuolar pyrophosphatase

INTRODUCTIONTrypanosoma cruzi (Clark, 1959), the etiologic agent of Chagasdisease, together with Leishmania spp. (Figarella et al., 2007) and anumber of monogenetic trypanosomes (Baqui et al., 2000; Linderand Staehelin, 1979), possess a contractile vacuole complex (CVC)involved in osmoregulation. In T. cruzi, the CVC has been shown tobe important for regulatory volume decrease (RVD) afterhyposmotic stress (Rohloff et al., 2004) and for shrinking of thecells when subjected to hyperosmotic stress (Li et al., 2011). Inaddition, we have recently reported a role for the CVC in traffickingglycosylphosphatidylinositol (GPI)-anchored proteins to theplasma membrane (Niyogi et al., 2014). Previous studies in T. cruzi(Hasne et al., 2010) and Dictyostelium discoideum (Heuser et al.,1993; Moniakis et al., 1999; Sesaki et al., 1997; Sriskanthadevanet al., 2009) suggest that some soluble (Sesaki et al., 1997;Sriskanthadevan et al., 2009) and membrane (Hasne et al., 2010;Heuser et al., 1993; Moniakis et al., 1999) proteins can also be

transported through the CVC to the plasma membrane. Thepresence of Rab11, a small GTPase that localizes in recyclingendosomes in most cells – including Trypanosoma brucei (Jeffrieset al., 2001) – and in the CVC of T. cruzi (Ulrich et al., 2011) andD. discoideum (Harris et al., 2001), suggests that the CVC could bean evolutionary precursor to the recycling endosomal system inother eukaryotes (Docampo et al., 2013; Harris et al., 2001).

In a previous proteomic and bioinformatic study of the CVC ofT. cruzi, we identified a number of proteins that have roles intrafficking, among them SNARE 2.1 and SNARE 2.2, VAMP1 (anortholog of mammalian VAMP7), AP180 and the small GTPasesRab11 and Rab32 (TcCLB.506289.80) (Ulrich et al., 2011). Rabproteins mediate tethering of incoming vesicles to the correct targetorganelle through cycling between a GDP-bound inactive and aGTP-active form (Zerial and McBride, 2001). They have also beenimplicated in vesicle budding and in the interaction withcytoskeletal elements (Zerial and McBride, 2001). Different RabGTPases are localized to different organelles, and this represents animportant determinant of the identity of each organelle (Bright et al.,2010; Munro, 2002; Pfeffer, 2001; Seabra and Wasmeier, 2004;Turkewitz and Bright, 2011). Rab32 and its close homolog Rab38are predominantly expressed in cells that produce lysosome-relatedorganelles (LROs), such as melanocytes and platelets (Bultemaet al., 2012), and it has been suggested that these Rabs could be thespecificity factors that work in concert with the ubiquitoustrafficking machinery to direct transport toward LROs (Bultemaet al., 2012). It has been proposed that LROs arise through thedelivery of specific cargoes from the early endosomal network,comprising sorting and recycling endosomes (Delevoye et al., 2009;Raposo and Marks, 2007).

T. cruzi possesses organelles that have similarities to LROs ofmammalian cells, known as acidocalcisomes (Docampo et al.,2005, 1995; Docampo and Moreno, 2011). Like LROs of humanplatelets (Ruiz et al., 2004; Smith et al., 2006) and mast cells(Moreno-Sanchez et al., 2012), acidocalcisomes have roundedmorphology, are acidic, and are rich in Ca2+, pyrophosphate (PPi)and polyphosphate (polyP). In addition, adaptor protein complex-3(AP-3), the system known to be involved in the transport ofmembrane proteins to LROs of mammalian cells (Theos et al.,2005), is also involved in the biogenesis of acidocalcisomes(Besteiro et al., 2008; Huang et al., 2011). Interestingly, electronmicroscopy analyses have previously provided evidence of thefusion of acidocalcisomes to the CVC of T. cruzi (Montalvetti et al.,2004) and D. discoideum (Marchesini et al., 2002). Also, underhyposmotic stress, acidocalcisomes fuse to the CVC, which resultsin translocation of an aquaporin [T. cruzi (Tc)AQP1] (Rohloff et al.,2004). In this work, we demonstrate that the expression ofdominant-interfering TcRab32 mutants alters osmoregulation,acidocalcisome morphology and content, as well as parasiteinfectivity. The results suggest that the CVC and TcRab32Received 29 January 2015; Accepted 30 April 2015

1Department of Cellular Biology and Center for Tropical and Emerging GlobalDiseases, University of Georgia, Athens, GA 30602, USA. 2Laboratorio deUltraestrutura Celular Hertha Meyer, Instituto de Biofısica Carlos Chagas Filho andInstituto Nacional de Ciência e Tecnologia em Biologia Estrutural e Bioimagens –

Universidade Federal do Rio de Janeiro, Rio de Janeiro 21941-902, Brazil. 3Diretoriade Metrologia Aplicada a Ciências da Vida, Instituto Nacional de Metrologia,Qualidade e Tecnologia (INMETRO), Xerem, Rio de Janeiro 25250-020, Brazil.*Present address: California State University, Fullerton, CA, USA.

‡Author for correspondence ([email protected])

2363

© 2015. Published by The Company of Biologists Ltd | Journal of Cell Science (2015) 128, 2363-2373 doi:10.1242/jcs.169466

Journal

ofCe

llScience

Page 2

participate in the trafficking of proteins involved in acidocalcisomebiogenesis, and reaffirm the role of the CVC as a trafficking hub.

RESULTSThe localization in the CVC of TcRab32, a Rab usually associatedwith LROs (Bultema and Di Pietro, 2013), suggests that the CVC, inaddition to its role in osmoregulation, could be involved in thebiogenesis of acidocalcisomes. We first confirmed the localization ofTcRab32 in the CVC using specific antibodies and then investigatedwhether expression of dominant-negative TcRab32 affectedosmoregulation and the biogenesis of acidocalcisomes. We analyzedthe interaction between the organelles, as well as the enzymaticactivities (vacuolar H+-pyrophosphatase and Ca2+-ATPase),composition and number of acidocalcisomes, and finally therelevance of this interaction for the infectivity of the parasites.

Localization of TcRab32 at different T. cruzi stagesWe have reported previously that N-terminal tagging of TcRab32with green fluorescent protein (GFP) results in fluorescent labelingof the CVC of epimastigotes and additional punctate staining(Ulrich et al., 2011). We confirmed this localization by usingindirect immunofluorescence analysis with specific affinity-purifiedantibody against TcRab32, which was raised in mice againstthe recombinant protein (supplementary material Fig. S1A).Supplementary material Fig. S2A,B shows that TcRab32 localizedto the CVC of wild-type epimastigotes, trypomastigotes andamastigotes, as evidenced by the circular staining close to theflagellar pocket; additional punctate staining was also observed,especially in epimastigotes and trypomastigotes. Western blotanalysis of parasite lysates, using the same antibody, revealed aband of∼26 kDa, corresponding to the native protein (supplementarymaterial Fig. S2C). A double band was detected in amastigotelysates, indicative of some cross-reaction with another protein orpost-translational modification that occurs at this life-cycle stage.Control experiments using pre-immune serum were negative. TheCVC localization of GFP–TcRab32 was also confirmed by usingimmunogold electron microscopy and antibodies against GFP(supplementary material Fig. S2D,E), which was negative whenwild-type cells were used.

In vitro prenylation studies of TcRab32TcRab32 possesses the sequence CSC at the carboxyl terminus(supplementary material Fig. S1B), and it is known that Rabprenylation at cysteine residues of the carboxyl end retains Rabs atmembranes (Jean and Kiger, 2012). To examine whether TcRab32is geranylgeranylated, we performed in vitro prenylationexperiments (Fig. 1A) using recombinant TcRab32 as substrate inthe presence of a cytosolic epimastigote extract as the source ofprenyltransferases. When tritiated geranylgeranyl pyrophosphate([3H]GGPP) was used as the isoprenoid donor, His-tagged TcRab32was efficiently geranylgeranylated as shown by the labeled band of42 kDa that was detected, corresponding to the His-tagged protein.The intensity of the prenylated band was strongest at 30 min, theoptimum incubation time. Conversely, when tritiated farnesylpyrophosphate ([3H]FPP) was used as the donor, we were unableto detect prenylation of recombinant TcRab32 (data not shown).Therefore, TcRab32 is specifically geranylgeranylated. Previousstudies of recombinant T. cruzi protein geranylgeranyl transferase I(GGTI) using a panel of mammalian and yeast protein substratesreport that two mammalian Rab-family GTPases containing theC-terminal CXC sequence do not serve as substrates for thisenzyme, as expected (Yokoyama et al., 2008; Nepomuceno-Silva

et al., 2001). Accordingly, Prenylation Prediction Suite (PrePS)predicts that geranylgeranyl transferase II (GGTII) is the enzymeinvolved in the prenylation of this protein.

Localization of TcRab32 mutantsTo examine the role of the prenylation motif in the targeting ofTcRab32 to cell membranes, we generated mutants in whichalanine replaced the cysteine residues at the prenylation motif, andwe studied the effect of this mutation on the localization of theprotein. The geranyl-geranyl tails of Rabs tether the proteins tocell membranes and help to restrict free diffusion through thecytoplasm (Rak et al., 2003). In transfected T. cruzi epimastigotes,GFP–TcRab32 mainly localized to the CVC (Fig. 1B), as reportedpreviously (Ulrich et al., 2011), although it also showed somelocalization to the perinuclear endoplasmic reticulum (ER), whichis probably a product of its overexpression, whereas the prenylationmutant GFP–TcRab32C241A/C243A exhibited cytosolic locali-zation (Fig. 1C). We also engineered an expression plasmidencoding a TcRab32 mutant that mimics the GDP-bound form(dominant-negative, TcRab32T24N) (Fig. 1D). In transfectedT. cruzi epimastigotes, the dominant-negative GFP–TcRab32exhibited cytosolic localization, which included localization to thecytoplasm along the flagellum. Taken together, these resultsindicate that TcRab32 localizes to the membrane of the CVC in aGTP- and geranylgeranyl-dependent manner. The morphology ofthe CVC remained unaffected when the mutant proteins wereexpressed. We confirmed tagging of the mutant proteins usingwestern blot analyses (Fig. 1E). The anomalous migration of theGFP–TcRab32 prenylation-motif mutants could be due to lack of

Fig. 1. TcRab32 is geranylgeranylated in vitro. (A) Radiolabeled proteinswere analyzed using SDS-PAGE on a 15% gel followed by autoradiography.Lane 1, in the presence of all reactants – rTcRab32, epimastigote extract and[3H]GGPP. Lanes 2 and 3 are negative controls. The enzyme assay wasperformed for 30 min. A radioactive band of 42 kDa was observed,corresponding to the His-tagged protein (arrow). (B) GFP–TcRab32-expressing epimastigotes (GR) show preferential localization in the contractilevacuole (arrow) and perinuclear region (arrowhead). (C) GFP–TcRab32prenylation-motif mutants have a cytosolic localization. (D) Mutant dominant-negative GFP–TcRab32, which mimics the GDP-bound state of the protein,has a cytosolic localization. Scale bars: 10 µm. (E) Western blot analysis oflysates from epimastigotes expressing GFP–TcRab32 (GR) or expressing theprenylation mutant (PM) or dominant negative (DN) TcRab32 using anantibody against TcRab32. Tubulin (antibody dilution 1:10,000) was used as aloading control. Arrow shows the molecular mass of GR and dominant-negative TcRab32, as well as the endogenous Rab32 in the PM lane. Theprenylation mutant has a higher molecular mass.

2364

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2363-2373 doi:10.1242/jcs.169466

Journal

ofCe

llScience

Page 3

prenylation or other post-translational modifications of this proteinin the absence of prenylation (Beranger et al., 1994).

Colocalization of GFP–TcRab32 with VP1 under osmoticstressIt has been reported that mammalian (Tamura et al., 2009) andXenopus (Park et al., 2007) Rab32 partially localizes tomelanosomes,which are LROs. We therefore investigated whether TcRab32partially co-localizes with the acidocalcisome marker vacuolarproton pyrophosphatase (V-H+-PPase, VP1). We did not observeany substantial overlap between the labeling of VP1 and GFP–TcRab32 under isosmotic conditions (Fig. 2A). The observedadditional punctate staining of GFP–TcRab32 could correspond toendosomes. However, under hyposmotic (Fig. 2B) or hyperosmotic(Fig. 2C) conditions, we observed that staining of VP1 overlappedwith that of GFP at the CVC region, in agreement with the previouslyreported fusion of these organelles under osmotic stress (Montalvettiet al., 2004; Rohloff et al., 2004). The lack of colocalization betweenGFP–TcRab32 and the acidocalcisome marker under isosmoticconditions could be attributed to the dynamic nature of the interactionbetween these two organelles. These results were confirmed byanalyses using cryo-immunogold electron microscopy. Afterhyposmotic stress, it was possible to detect colocalization of VP1staining and GFP–TcRab32 in the contractile vacuole bladder and inthe tubules of the spongiome (Fig. 2D,E).We have previously reported that GFP-tagged VAMP1 (the

ortholog of mammalian VAMP7, here on referred to as TcVAMP7)localizes to the contractile vacuole bladder of epimastigotes thathave been submitted to osmotic stress (Ulrich et al., 2011).Immunogold electron microscopy analyses of GFP localizationalso revealed labeling of the spongiome, flagellar pocket, smallvesicles and plasmamembrane (supplementarymaterial Fig. S3A,B).No labeling was observed when wild-type cells were used. Correcttagging of the protein was demonstrated using western blot analysis

(supplementary material Fig. S3C).We now report that when taggedwith GFP at its C-terminus, TcVAMP7–GFP localizespredominantly to the acidocalcisomes, as revealed by itscolocalization with VP1 (Fig. 3A) and by using cryo-immunogoldelectron microscopy (Fig. 3B). TcVAMP7–GFP also weaklylabeled the CVC, as shown by using immunofluorescenceanalyses (Fig. 3D). When the cells were submitted to hyposmoticstress, an increase in TcVAMP7–GFP labeling of the CVC wasobserved (Fig. 3C,E,F), and once regulatory volume decrease(RVD) had been completed, acidocalcisome labeling predominatedagain (Fig. 3G), suggesting transient fusion of both organelles.Live-cell microscopy analyses of epimastigotes that expressedTcVAMP7–GFP and were labeled with BODIPY–ceramide(Fig. 3H,I; supplementary material Movies 1 and 2) shows how,after hyposmotic stress, acidocalcisomes that had been labeledwith TcVAMP7–GFP make contact with the CVC, reinforcingthe hypothesis that these two organelles interact and exchangeproteins.

Additional evidence of contact of acidocalcisomes with theCVC complex under osmotic stressBy using electron tomography to examine the CVC underhyposmotic stress, we observed an organized network ofinterconnected tubules (spongiome) that was attached to thecentral vacuole or bladder (Fig. 4A), confirming previous results(Girard-Dias et al., 2012). Reconstruction of the whole volume ofthe organelle by using serial electron tomography showed that theCVC was surrounded by acidocalcisomes, which were identifiedby their remaining electron-dense content that was observedalong the depth of each organelle (Fig. 4A–C; supplementarymaterial Movie 3). Inspection of the virtual sections through thetomogram provided further evidence for the close apposition ofacidocalcisomes with the bladder of the CVC under hyposmoticconditions (Fig. 4C,D–F; supplementary material Movie 3). Taken

Fig. 2. Colocalization of GFP–TcRab32 with staining of VP1 under osmotic stress. (A) There is no colocalization between VP1 (red) and GFP–TcRab32(green), as detected with antibodies against T. brucei VP1 and GFP under isosmotic conditions. (B,C) Overlap between signals for VP1 (red) andGFP–TcRab32 (green) in the CVC occurs under hyposmotic (B) or hyperosmotic (C) stress conditions (arrows, and enlarged insets in the merged images).(D,E) Cryo-immunogold electron microscopy analysis of epimastigotes subjected to hyposmotic conditions (150 mosm/l) and incubated with rabbit anti-GFPantibodies (1:25) and mouse anti-VP1 antibodies (1:25, 12-nm gold particles, arrows) at 4°C overnight, and then treated with secondary antibodies (goatanti-mouse conjugated to 12-nm colloidal gold, arrows) and goat anti-rabbit 18-nm gold for 1 h at room temperature. The images show colocalization in thecontractile vacuole (CV) bladder and the spongiome (Sp). Ac, acidocalcisome; FP, flagellar pocket; F, flagellum; K, kinetoplast. Scale bars: 10 µm (A, appliesto A–C); 2 µm (D,E).

2365

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2363-2373 doi:10.1242/jcs.169466

Journal

ofCe

llScience

Page 4

together with the results presented thus far, these results aresuggestive of fusion events (Fig. 3H,I and Fig. 4; supplementarymaterial Movies 1–3).

Response of dominant-negative Rab32 mutants tohyposmotic and hyperosmotic stressFusion of acidocalcisomes to the CVC is important for the responseof the parasites to osmotic stress (Rohloff et al., 2004). We thereforeinvestigated whether epimastigotes that expressed dominant-negative GFP–TcRab32 were deficient in their response tohyposmotic and hyperosmotic stresses. Wild-type epimastigotesand epimastigotes expressing GFP–TcRab32 or the dominant-negative mutant were subjected to hyposmotic stress, and their RVDwas measured using the light-scattering technique, as describedpreviously (Niyogi et al., 2014). Dominant-negative mutants wereless able to recover their volume after hyposmotic stress than wild-type cells, whereas recovery was faster in GFP–TcRab32-expressing cells (Fig. 5A), probably as a result of overexpressionof this protein. When subjected to hyperosmotic stress,epimastigotes expressing dominant-negative TcRab32 shrank less,whereas cells expressing GFP–TcRab32 shrank more than controlcells (Fig. 5B), and in all cases, they did not recover their volumeduring the time of the experiment, as has been reported before (Liet al., 2011).

Deficient acidocalcisome biogenesis and function in cellsexpressing dominant-negative TcRab32VP1 is a marker of acidocalcisomes in T. cruzi, and its activity wasinvestigated in digitonin-permeabilized epimastigotes by measuringthe uptake of Acridine Orange, which was induced by the additionof PPi. The decrease in fluorescence, after a delay that was due toplasma membrane permeabilization, indicated increasing vacuolaracidity (Fig. 6A). The vacuolar pH was neutralized, and AcridineOrange was released by the K+/H+ exchanger nigericin. The initialrate of PPi-induced acidification was greatly decreased inepimastigotes expressing dominant-negative GFP–TcRab32(Fig. 6A).

Acidocalcisomes are capable of taking up Ca2+ using a plasma-membrane-type Ca2+-ATPase (PMCA) (Lu et al., 1998). Ca2+

uptake by digitonin-permeabilized epimastigotes was measuredusing Calcium Green-5N (Fig. 6B). The initial rate of Ca2+ uptakewas greatly decreased in cells that expressed dominant-negativeGFP–TcRab32 (Fig. 6B). Ca2+ was released by the addition ofnigericin (Fig. 6B). Ca2+ uptake under these conditions measuresnot only Ca2+ uptake by acidocalcisomes but also by thesarcoplasmic-endoplasmic reticulum-type Ca2+-ATPase (SERCA).We therefore evaluated whether cells expressing dominant-negativeGFP–TcRab32 were deficient in acidic Ca2+ content. In previouswork with T. cruzi epimastigotes, we noticed that epimastigoteswere deficient in esterases that were able to hydrolyze Fura-2/AM,which is commonly used to detect Ca2+ pools in live cells, but thiswas not the case with infective stages (Docampo et al., 1993;Moreno et al., 1992). We therefore loaded amastigotes that had beenobtained by differentiation of wild-type epimastigotes andepimastigotes expressing dominant-negative GFP–TcRab32, asdescribed inMaterials andMethods, and examined their acidic Ca2+

content using ionophores, as reported previously (Moreno et al.,1992). Addition of ionomycin to Fura-2-loaded amastigotes in Ca2+-free medium (with the addition of 100 µMEGTA) resulted in Ca2+

release from neutral or alkaline compartments because ionomycinbinds to essentially no Ca2+ below pH 7.0 and it cannot carry Ca2+

out of acidic compartments. However, further addition of nigericin,which alkalinize acidic compartments, resulted in a greater releaseof Ca2+ (Fig. 6C). If the order of addition was reversed, nigericincaused a transient increase in Ca2+, which was greatly increasedafter ionomycin addition (Fig. 6D). The amount of Ca2+ released by

Fig. 3. Translocation of TcVAMP7–GFP to the CVC under hyposmoticconditions. (A) Colocalization of VAMP7–GFP (green) with VP1 (red) using apolyclonal antibody against GFP and a monoclonal antibody against TcVP1,respectively. (B,C) Cryo-immunogold electron microscopy analyses usinganti-GFPantibodies show predominant labeling of acidocalcisomes (Ac) underisosmotic conditions (B) and of the CVC under hyposmotic conditions(C). FP, flagellar pocket; F, flagellum; K, kinetoplast; N, nucleus.(D–G) Immunofluorescence analysis of cells imaged under isosmoticconditions (D) after 30 s (E), 2 min (F) and 5 min (G) of hyposmotic shock(150 mosm/l) using anti-GFP antibodies (green). DAPI staining labels DNA(blue). (H) Labeling of TcVAMP7–GFP-overexpressing parasites (green) withBODIPY–ceramide (red). The time indicated on each frame corresponds to10, 30, 110, 120, 130 and 170 s (top row, left to right then bottom row, left toright). Note the dilation of the contractile vacuole bladder. (I) Overlay ofdifferential interference contrast with the green and red channels at the sametime points as those in H. In H and I, yellow arrows show acidocalcisomes, andgreen arrows show the localization of the CVC. Scale bars: 10 µm (A,D–G);2 µm (B,C); 5 µm (H,I).

2366

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2363-2373 doi:10.1242/jcs.169466

Journal

ofCe

llScience

Page 5

the combination of ionomycin and nigericin was considerably lowerin cells expressing dominant-negative GFP–TcRab32 (Fig. 6C,D),suggesting that the acidocalcisomes of these mutant cells containless Ca2+.Most PPi and polyP in trypanosomes accumulate in

acidocalcisomes (Docampo et al., 2005; Docampo and Moreno,2011). It is unknown whether PPi is taken up from the cytosol orsynthesized inside acidocalcisomes, whereas synthesis of polyP isthrough the activity of polyP kinases, such as that formed by thevacuolar transporter chaperone (VTC) complex (Lander et al., 2013;Ulrich et al., 2014). This is a complex of at least two subunits intrypanosomatids, VTC1 and VTC4, both of which localize to themembrane of acidocalcisomes, and VTC4 is the catalytic subunit(Lander et al., 2013; Ulrich et al., 2014). We hypothesized thatif TcRab32 is important for the biogenesis of acidocalcisomes,these organelles will have a reduced ability to synthesize these

compounds, and this is what we found. Expression of the dominant-negative form of TcRab32 led to a significant reduction in the levelsof short-chain polyP (∼80%) (Fig. 6E) and PPi (∼50%) (Fig. 6F).There was, however, no significant change in the content of longchain polyP (>300 up to 700–800 phosphate units) (Fig. 6G),suggesting that only the activity of the VTC complex, which ismainly involved in the synthesis of short-chain polyP (Lander et al.,2013; Ulrich et al., 2014), is affected in these mutants. The resultswere further verified by visualization of short-chain polyP that hadbeen extracted from the above cell lines, resolved by urea-PAGEand stained with Toluidine Blue (Fig. 6H).

Changes in acidocalcisomeelectron-density and thenumberof cells expressing dominant-negative GFP–TcRab32In previous work (Mendoza et al., 2002; Urbina et al., 1999),electron microscopy techniques have been used to demonstrate thattreatment of fixed trypanosomes with yeast pyrophosphatase resultsin loss of the electron-density of acidocalcisomes, as observed inwhole unstained cells, suggesting that PPi (complexed with cations)is the main electron-dense material in these organelles (Urbina et al.,1999). In agreement with the considerable decrease in PPi and shortchain polyP content in epimastigotes expressing dominant-negativeGFP-TcRab32, we detected an increase in the presence of emptyand dilated vacuoles in intact unstained epimastigotes expressingthe mutant protein by using direct transmission electron microscopy(compare Fig. 7A,B). There was also a substantial reduction in thenumber of acidocalcisomes per cell (Fig. 7B,G). Of the cellsexpressing the dominant-negative protein, 84% had empty dilatedvacuoles (Fig. 7H) with an average of ∼10 empty vacuoles perdominant-negative cell. Similar results were obtained withtrypomastigotes (compare Fig. 7C,D) and amastigotes (compareFig. 7E,F) expressing dominant-negative GFP–TcRab32.

Cells expressing dominant-negative GFP–TcRab32 havereduced growth and infectivityThe growth rate of the epimastigotes expressing dominant-negativeGFP–TcRab32 was reduced as compared to that of controlepimastigotes expressing GFP alone (Fig. 8A).

To study the infectivity of cells expressing dominant-negativeGFP–TcRab32, we fully differentiated them into cell-derivedtrypomastigotes, as described in Materials and Methods. Invasionof culture cells with parasites expressing dominant-negative

Fig. 4. Close apposition of anacidocalcisome with the CVC, indicatinga fusion event. (A–C) Virtual section (1-nmthickness) sequence of a tomogram showingthe anterior region of the parasite. The CVC isrepresented by the central vacuole of bladder(CV) and the spongiome (Sp).Acidocalcisomes (Ac) in the neighboring regionare observed in close contact with the CVC. Inthe left lower corner, the section number isshown. In C, it is possible to observe a closeapposition between acidocalcisome and CVCmembranes, which are suggestive of a fusionevent (arrow) between the two organelles.(D–F) 3D models of the CVC (blue) and itsclose contact with an acidocalcisome (orange).(F) Tilted view of the 3D model at 45° aroundthe x axis. Spongiome (Sp) and flagellum (‘F’)are shown. Scale bar: 200 nm.

Fig. 5. Effect of TcRab32 mutation on the parasite response tohyposmotic and hyperosmotic stress conditions. Epimastigotes werepre-incubated in isosmotic buffer for 3 min and then subjected to hyposmotic(final osmolarity=150 mosm/l) (A) or hyperosmotic (final osmolarity=650 mosm/l)(B) stress. Monitoring absorbance at 550 nm by light scattering was used tofollow the relative changes in cell volume. As compared with wild-type cells(WT), cells expressing dominant-negative GFP–TcRab32 (DN) failed to fullyrecover their volume after hyposmotic stress and shrank less afterhyperosmotic stress, whereas cells expressing GFP–TcRab32 (GR)recovered their volume faster after hyposmotic stress and shrank more afterhyperosmotic stress. Values are means±s.d. of three independentexperiments. Asterisks indicate statistically significant differences, P<0.05(Bonferroni’s multiple comparison ‘a posteriori’ test of one-way ANOVA), at alltime points after the induction of osmotic stress (after 4 min).

2367

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2363-2373 doi:10.1242/jcs.169466

Journal

ofCe

llScience

Page 6

GFP–TcRab32 was substantially reduced as compared with that ofcontrols that had been infected with parasites transfected withGFP alone or with wild-type parasites (Fig. 8B,C). Cytosoliclocalization of cells expressing dominant-negative GFP–TcRab32was maintained when epimastigotes were differentiated intotrypomastigotes and intracellular amastigotes (supplementarymaterial Fig. S4). These results suggest a role for TcRab32 ininfectivity.

DISCUSSIONWe show here that expression of a dominant-negative form of theGTPase TcRab32 results in alterations in the morphology andcontent of acidocalcisomes, and in a deficient response to osmoticstress, growth in vitro and invasion of host cells. The results suggestthat the CVC, where TcRab32 is located, is involved in thebiogenesis and function of acidocalcisomes.We have reported previously that GFP-tagged TcRab32 localizes

to the CVC of T. cruzi epimastigotes (Ulrich et al., 2014). We nowconfirm those observations using antibodies against the protein andfind it distributed in the CVC at different stages of the parasitelife cycle, with additional punctate staining in epimastigotes

and trypomastigotes. Dominant-negative GFP–TcRab32 or theTcRab32 protein lacking the prenylation motif, however, have acytosolic localization, indicating that localization to the CVC isgeranylgeranyl- and GTP-dependent. Dominant-negative TcRab32mutants might act by blocking or reducing the function ofendogenous TcRab32 by competing or sequestering Rab32effector proteins.

TcRab32, like other Rab32 proteins, contains amino acidsequences that are shared with only a small number of other Rabsequences (Hirota and Tanaka, 2009). For example, threonine in theWDTAGQE sequence (GTP binding site), which is conserved inalmost all Rab proteins, is replaced by isoleucine (WDIAGQE). Asimilar replacement is found in Rab38 and Rab29/Rab7L1 ofmammalian cells, in RabE from D. discoideum (Hirota andTanaka, 2009) and in Tetrahymena thermophila Rab32 (Brightet al., 2010), but there are no orthologs of any of the other Rabs inT. cruzi (Berriman et al., 2005). TcRab32 also possesses three aminoacids – Gly-75, Gln-76 and Val-80 – that are only conserved in theswitch II region of Rab32 and Rab38 alone and not in anyof the other58 Rabs ofmammalian cells (Tamura et al., 2011). Val-80 is requiredfor binding ofmammalianRab32 to its effector VPS9-ankyrin-repeat

Fig. 6. Deficient PPi-dependent H+ uptake, ATP-dependent Ca2+ uptake, acidic Ca2+ stores, short-chain polyP and PPi levels in epimastigotes

expressing dominant-negative GFP–TcRab32 in comparison with wild-type and GFP–TcRab32-expressing epimastigotes. (A) Wild type (black) orepimastigotes expressing dominant-negative GFP–TcRab32 (red, DN) (5×107/ml) were added to the standard buffer containing 125 mM sucrose, 65 mM KCl,10 mM Hepes, pH 7.2, 2.5 mM potassium phosphate, 1 mM MgCl2, to which 3 µM Acridine Orange and 40 µM digitonin were added. PPi (0.1 mM) and nigericin(1 µM, Nig) were added where indicated. (B) Epimastigotes, as above, were added to the standard buffer (in A), to which 1 µM Calcium Green-5N and 40 µMdigitonin were added. ATP (1 mM) and nigericin (1 µM) were added where indicated. The bar graphs (A,B) show quantification of the initial rate of H+ (A) or Ca2+

(B) uptake after adding PPi (A) or ATP (B) post-permeabilization. Three independent experiments were used for quantification, and results are expressed asmeans±s.e.m. (C,D) Amastigotes (5×107/ml) loaded with Fura-2 were incubated in buffer A in the presence of 100 µM EGTA, and, where indicated, 1 µMionomycin (Ion) or 1 µM nigericin was added. Bar graphs in C,D show quantification of the amount of Ca2+ released by the combination ionomycin-nigericin asmeans±s.e.m. of three independent experiments. Asterisks in A,B,D indicate that differences were significant (P<0.05). In panel C, the difference was marginal(P<0.07). Extracts from epimastigotes expressing dominant-negative GFP–TcRab32 (DN) showed ∼80% reduction in short-chain polyP levels (E) and ∼50%reduction in PPi levels (F) with no significant changes in long-chain polyP levels (G) in comparison with wild-type (WT) or GFP–TcRab32-expressingepimastigotes (GR). Values are means±s.e.m. of three independent experiments. *Differences are statistically significant as compared with respective controls,P<0.05 (0.041, DN versus GR, and 0.031, DN versus WT for E; 0.028, DN versus GR, and 0.003, DN versus WT for F) (Student’s t-test). (H) Extracts of short-chain polyP produced by epimastigotes expressing GFP–TcRab32 (GR) or dominant-negative GFP–TcRab32 (DN). Two samples were resolved by using urea-PAGE and visualized using Toluidine Blue. ORG represents migration of Orange G dye. Bands corresponding to short-chain polyP of different lengths are shownas a series of bands in the lower part of the gel; these bands had lower intensities in lanes corresponding to the dominant-negative mutant (two samples) incomparison with those of cells transfected with GFP–TcRab32 (GR) and of control wild-type cells (WT) (two samples).

2368

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2363-2373 doi:10.1242/jcs.169466

Journal

ofCe

llScience

Page 7

protein (Varp)/Ankrd27, and this interaction is important fortrafficking of tyrosinase-related protein 1 to melanosomes (Tamuraet al., 2011). By contrast, TcRab32 [as well as T. thermophila Rab32(Bright et al., 2010)] has a phenylalanine residue at amino acidposition 194 instead the alanine residue in mammalian Rab32(supplementary material Fig. S1B). This alanine is an anchoringdeterminant for regulatory subunit IIα (RIIα subunit) of proteinkinase A and is responsible for mammalian Rab32 interaction withmitochondria (Alto et al., 2002). In agreement with those studies, wefound that TcRab32 does not associate with mitochondria incolocalization experiments using Mitotracker (data not shown).Interestingly, other authors have also been unable to confirm theassociation of human Rab32 with the mitochondria of COS cells(Hirota andTanaka, 2009), andT. thermophilaRab32 associateswiththe phagolysosomal system (Bright et al., 2010).

Rab proteins participate in membrane trafficking events thatinvolve membrane fusion, fission and motility. Although our data donot distinguish between these events, the localization of TcRab32 inthe CVC, as well as the deficient morphology and content ofacidocalcisomes upon expression of its dominant-negative form,suggests that the CVC acts as an equivalent to the early endosomes ofmammalian cells where TcRab32 functions as a tether to facilitatecargo loading into fused vesicles (Bultema and Di Pietro, 2013). Thefusion of CVCwith acidocalcisomes would facilitate the exchange ofmembrane proteins between the organelles, such as translocation ofTcAQP1 from acidocalcisomes to the CVC (Rohloff et al., 2004), orof membrane enzymes and transporters involved in acidification, aswell as Ca2+ transport from the CVC to the acidocalcisomes. Insupport of this hypothesis, in previous studies we have presented theresults of electron microscopy analyses showing the localization ofVP1 (Montalvetti et al., 2004) and a Ca2+-ATPase (Lu et al., 1998) inlarge vacuoles of T. cruzi, which were comparable in the size and thelocation of the CVC. The fact that acidocalcisomes exhibiteddeficiencies in the transport of H+ and Ca2+ in epimastigotesexpressing dominant-negative GFP–TcRab32 is compatible withdeficiencies in the transfer of these proteins to the organelles. Ourelectron tomography results provide direct evidence of contact ofacidocalcisomes with the CVC under osmotic stress. Transientcolocalization of VP1 with GFP–TcRab32 under hyposmotic andhyperosmotic stress, and the transfer of TcVAMP7–GFP fromacidocalcisomes to the CVC under hyposmotic stress support thoseresults. CVC-located TcRab32 is likely to function as a tether in orderto attract acidocalcisomes, facilitating fusion that is mediated by theCVC SNAREs (Ulrich et al., 2011) and the acidocalcisomeTcVAMP-7, leading to an exchange of proteins. This model isconsistent with that in the mammalian system, where Rab32 effector

Fig. 8. Reduced infectivity of GFP–TcRab32 mutant trypomastigotes.(A) Growth rate of epimastigotes expressing dominant-negative (DN) GFP–TcRab32 in comparison to GFP-expressing epimastigotes (GFP). Values aremeans±s.d. from three independent experiments. (B,C) The effect ofexpression of dominant-negative GFP–TcRab32 (DN) on the invasion bytrypomastigotes of host cells in vitro in comparison with wild-typetrypomastigotes (WT) or GFP-expressing trypomastigotes (GFP). Values aremean±s.d. (n=3). *Differences are statistically significant, P<0.05 (one wayANOVA with Bonferroni post-hoc test).

Fig. 7. Reduction in electron-dense acidocalcisomes and a considerableincrease in empty vacuoles in cells expressing dominant-negativeTcRab32 in comparison with wild-type cells. (A,B) Direct transmissionelectron microscopy image from whole unstained and unfixed epimastigotesexpressing dominant-negative GFP–TcRab32 (DN) show the presence ofnumerous empty vacuoles (B) in comparison with wild-type epimastigotes (A).(C–F) A similar increase in the number of empty vacuoles was observed intrypomastigotes (C,D) and amastigotes (E,F) expressing dominant-negativeGFP–TcRab32 (D,F) as compared with that of wild-type cells (C,E). (G) Thenumber of acidocalcisomes per epimastigote was counted in 70 random cellsfrom two independent experiments, and the numeric distribution ofacidocalcisomes showed that the majority of epimastigotes expressingdominant-negative TcRab32 (TcRab32DN) had <10 or between 11 and 20electron-dense acidocalcisomes. The results in wild-type (WT) cells are similarto those reported previously for trypanosomatids. (H) An increase in thepercentage of epimastigotes showing empty vacuoles. In order to quantify thephenotype of empty vacuoles, by using transmission electron microscopy, weexamined 50 random wild-type parasites, ‘C’, and parasites expressing thedominant-negative protein (DN), and counted the number of parasites withempty vacuoles for each. In comparison with wild type, we found that therewasa significant increase in the percentage of parasites with empty vacuoles whenthey expressed dominant-negative TcRab32. *Differences between controland the dominant-negative protein are statistically significant, P<0.05. Scalebars: 2 µm (A,B); 1 µm (C–F).

2369

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2363-2373 doi:10.1242/jcs.169466

Journal

ofCe

llScience

Page 8

proteins interact with VAMP7 (a vesicle SNARE protein that isinvolved in vesicle fusion; Tamura et al., 2011, 2009) and theδ-subunit of the AP-3 complex (Martinez-Arca et al., 2003) (acomplex known to be involved in the biogenesis of acidocalcisomes;Besteiro et al., 2008; Huang et al., 2011). Other proteins could beinvolved in these interactions. For example, myosin heavy chain(myosin VI andmyosin VII), which is an actin-basedmotor, has beendetected in proteomic analyses of the CVC of T. cruzi (Ulrich et al.,2011). In humans,myosinVC is an effector ofRab32, and is involvedinmelanosomebiogenesis and in the trafficking of integralmembraneproteins to the melanosome (Bultema et al., 2014).Our results are also in agreement with a role for acidocalcisomes in

osmoregulation. Fusion of acidocalcisomes with the contractilevacuole, and the concomitant hydrolysis of polyP has beenpostulatedto lead to an increase in phosphate and cations in the bladder,resulting in water accumulation (Docampo et al., 2013; Rohloff andDocampo, 2008). These changes are likely to be accompanied by thetransfer of a phosphate transporter and cation exchangers (Ulrichet al., 2011) from the acidocalcisome membranes, together withTcAQP1 (Rohloff et al., 2004), providing the means for wateraccumulation (through TcAQP1), as well as for the return of cations(cation exchangers) and phosphate (Pi transporter) to the cytosol afterwater elimination (Docampo et al., 2013). Our results are consistentwith several possible interactions, including unloading of theacidocalcisome content into the CVC, moving proteins fromacidocalcisomes to the CVC and remodeling of acidocalcisomesthrough the transfer of proteins from the CVC to acidocalcisomes.In conclusion, we propose that the CVC is a trafficking hub that is

not only involved in the transfer of GPI-anchored proteins to theplasma membrane (Niyogi et al., 2014) but is also a specializedendosomal system that can be used to deliver membrane proteinsthat are important for the biogenesis of acidocalcisomes.

MATERIALS AND METHODSCell cultureEpimastigotes from T. cruzi were cultured in liver infusion tryptose (LIT)medium containing 10% fetal calf serum at 28°C. T. cruzi epimastigotesthat had been transfected with GFP–TcRab32, dominant-negative GFP–TcRab32 or GFP-TcRab32C241A/C243A, as well as GFP–TcVAMP7 andTcVAMP7–GFP, were maintained in the presence of 250 µg/ml geneticin(G418). Human foreskin fibroblasts (HFF) were grown in low glucoseDulbecco’s modified Eagle’s medium (DMEM) supplemented with 10%Cosmic Calf Serum (HyClone) and 0.1% L-glutamine. Vero cells were grownin RPMI supplemented with 10% fetal bovine serum. L6E9 myoblasts weregrown in high glucose DMEM supplemented with 10% fetal bovine serum.Host cells were maintained at 37°C with 5% CO2. Tissue-culture-derivedtrypomastigotes were obtained from Vero cells that had been infected withmetacyclic trypomastigotes fromstationary cultures of parasites that expressedTcGFP, GFP–TcRab32 and dominant-negative GFP–TcRab32. T. cruziamastigote and trypomastigote formswere collected from the culture mediumof infected myoblasts using a modification of the method of Schmatz andMurray (Schmatz and Murray, 1982), as described previously (Moreno et al.,1994).Wedetermined the growth of epimastigotes bymeasuring the change inoptical density at 600 nm in a Gilford spectrophotometer with a startingculture of 4.5×106 epimastigotes.

Chemicals and reagentsFetal bovine serum, Dulbecco’s PBS and Hank’s solution, 4′,6-diamidino-2-phenylindole (DAPI), DMEM and RPMI media, paraformaldehyde,bovine serum albumin and protease inhibitors were purchased from Sigma(St Louis, MO). Restriction enzymes were purchased from New EnglandBioLabs (Ipswich, MA). pCR2.1-TOPO cloning kit, 1 kb plus DNA ladder,rabbit antibodies against GFP and Gene Tailor Site-Directed MutagenesisSystem were from Invitrogen (Life Technologies, Grand Island, NY).

Hybond-N nylon membranes were obtained from PerkinElmer (Waltham,MA). Pierce ECL western blotting substrate and BCA protein assay reagentwas from Pierce (Thermo Fisher Scientific, Rockford, IL). All other reagentswere analytical grade. The oligonucleotides were ordered from Sigma orIntegrated DNA Technologies (Coralville, IA). Vector pET32 Ek/LIC,Benzonase® Nuclease and anti-histidine-tag antibodies were from Novagen(EMD Millipore, Billerica, MA). [1-3H(N)]-farnesyl pyrophosphate,triammonium salt (23.0 Ci/mmol), [1-3H(N)]-geranylgeranyl pyrophosphate,triammonium salt (22.4 Ci/mmol) and EN3HANCE were from Perkin Elmer.

In vitro infection assayHFF or irradiated myoblasts (6×105 cells per well) were equally distributedin a 12-well plate on a sterile coverslip in their respective growth medium(as mentioned above) and were incubated for 24 h at 37°C under a 5% CO2

atmosphere. The following day, the cells were washed once withDulbecco’s Hanks’ solution, and 6×106 wild-type trypomastigotes ortrypomastigotes expressing GFP, GFP–TcRab32 or dominant-negativeGFP–TcRab32 were added to each well (10 trypomastigotes per myoblastor HFF), and they were incubated for 4 h at 37°C under a 5% CO2

atmosphere. To decrease the chances of contamination of cell-derived-trypomastigotes with extracellular amastigotes, collections of parasiteswere centrifuged and incubated at 37°C for 2 h to allow trypomastigotes toswim to the surface. The supernatant was collected and used forsubsequent invasion assays. Next, the parasites were removed from theplate, and the infected cells were washed extensively with Dulbecco’sHank’s solution and fixed for immunofluorescence assays. For attachmentand internalization assays, recently internalized parasites, and parasitesthat were caught in the process of invasion, were included and manuallycounted in at least 200 DAPI-stained cells in three independentexperiments. The percentage of infected cells and the average number ofparasites per infected cell were determined.

Immunofluorescence and western blot analysesFor immunofluorescence microscopy, parasites were fixed in PBS,pH 7.4, with 4% paraformaldehyde, adhered to polylysine coverslips andpermeabilized for 3 min with PBS, pH 7.4, containing 0.3% Triton X-100.Permeabilized cells were quenched for 30 min at room temperature with50 mM NH4Cl and blocked overnight with 3% BSA in PBS, pH 8.0.Samples were incubated with both primary and secondary antibodies for 1 hat room temperature. Coverslips were mounted by using mounting mediumcontaining DAPI at 5 µg/ml for staining DNA-containing organelles.For imaging of intracellular parasites, mammalian cells were seeded ontosterile coverslips in 12-well culture plates and allowed to grow for 24 h. Tosemi-synchronize the infection, we added the parasites at a ratio of 10:1(parasite:host cell) for 4 h, washed the cells to eliminate extracellularparasites and then fixed them in cold methanol for 30 min. The dilutionsused for primary antibodies were as follows: mouse anti-Rab32 (1:200);rabbit polyclonal anti-GFP (1:500); polyclonal rabbit against T. brucei VP1(1:250) (Lemercier et al., 2002) and monoclonal rabbit against T. cruzi VP1(1:250). Differential interference contrast (DIC) and direct fluorescenceimages were obtained by using an Olympus IX-71 inverted fluorescencemicroscopewith a Photometrix CoolSnapHQ charge-coupled device (CCD)camera driven by Delta Vision softWoRx3.5.1 (Applied Precision,Issaquah, WA). Images were deconvolved for ten cycles using the samesoftware and applying the ‘noise filter’ at ‘medium’mode. This is automaticdeconvolution software and was applied to all channels; brightness andcontrast were the same in all channels. The figures were composed by usingAdobe Photoshop 13.0.5×64 (Adobe System, Inc., San Jose, CA).

BODIPY-ceramide labeling and live cell-imagingEpimastigotes overexpressing TcVAMP7–GFP were collected bycentrifugation in the exponential phase of growth and resuspended ata density of 1×107 parasites/ml in LIT medium containing 5 µM ofBODIPY-TR C5 ceramide (Invitrogen). After 1 h of incubation at 28°C, thecells were collected by centrifugation at 1600g and washed twice in PBS, pH7.4. Labeled epimastigotes were then placed onto glass-bottomed petridishes that had been previously treated with polylysine. Cells were allowed

2370

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2363-2373 doi:10.1242/jcs.169466

Journal

ofCe

llScience

Page 9

to settle in the dish for 2 min, and loose parasites were washed away withisosmotic buffer (Jimenez and Docampo, 2012). Live-cell imaging wasstarted in isosmotic conditions at 1 frame/s, and after 30 s hyposmotic bufferwas added to the cells to reach a final osmolarity of 117 mosm/l. Cellswelling and recovery was captured for 5 min. Imaging was performed usingan Olympus IX-71 inverted fluorescence microscope, as described above.

Generation of dominant-negative and prenylation-motif TcRab32mutants, and transfectionDominant-negative (GFP–TcRab32T24N) and prenylation-motif mutant(GFP–TcRab32C241A/C243A) forms of TcRab32 were constructed usingsite-directed mutagenesis by the use of the Gene Tailor Site-DirectedMutagenesis System. This method involved methylating the TOPO blunt-end vector containing the coding sequence for TcRab32 with DNAmethylase at 37°C for 1 h, followed by amplification of the plasmid in amutagenesis reaction with two overlapping primers, forward,TcRab32T24N Forw: 5′-GTGAGGGAGGCACGGGGAAAAACTG-3′and reverse, TcRab32T24N Rev: 5′-TTTCCCCGTGCCTCCCTCACC-AATGA-3′ (for dominant-negative mutants); and forward, TcRab32C241A/C243A Forw: 5′-AAGAAAAGTCGGGCGCCTCCGCTTAA-3′and reverse, TcRab32 C241A/C243A Rev 5′-GGAGCAGCCCGA-CTTTTCTTCCCGTC-3′ (for prenylation mutants) of which the forwardprimer had the target mutation, resulting in the mutation of amino acidthreonine to asparagine (dominant negative), or cysteine to alanine(prenylation-motif mutant). Mutations were confirmed by sequencing(Yale DNA Analysis Facility, Yale University, New Haven, Connecticut).After transformation, the resulting plasmids in TOPO were digested withrestriction enzymes BamHI and HindIII. The circular pTEX-N-GFP vectorwas linearized by the corresponding restriction enzymes. Finally,TcRab32T24N, and TcRab32C241A/C243A inserts were ligated to pTEX-N-GFP followed by transformation. The plasmids pTEX-GFPTcRab32T24N/C241A/C243A were sequenced to confirm that the correct reading frameswere used. T. cruzi Y strain epimastigotes were transfected in cytomix(120 mM KCl, 0.15 mM CaCl2, 10 mM K2HPO4, 2 mM EDTA, 5 mMMgCl2, pH 7.6) containing 50 μg of the plasmid construct in a 4-mm cuvette.The cuvette was cooled on ice for 10 min and pulsed three times (1.5 kV,25 μF) with a Gene Pulser Xcell™ (Bio-Rad), and expression of GFP-fusionproteins was verified by western blot analyses. Stable cell lineswere established under drug selection with G418 at 250 μg/ml. Enrichmentof GFP-fluorescent parasites was performed with a high-speed cell sorterwhen needed (MoFlo Legacy; Beckman-Coulter, Hialeah, FL).

Generation of TcVAMP7-overexpressing parasitesTcVAMP7 tagged at the N-terminal with GFP constructs were generated aspreviously described (Ulrich et al., 2011). For tagging of the C-terminalwith GFP, the TcVAMP7 open reading frame (ORF) was amplified fromgenomic DNA from CL-strain parasites with primers forward 5′-GAATT-CATGGCCATTATATCATCTTTTGTT-3′ and reverse 5′-AAGCTTTTT-TTTGCACTTTTTAAAATC-3′. The ORF, flanked by EcoRI and HindIIIrestriction sites, was cloned into TOPO blunt vector, sequenced andsubcloned into pTREX-GFP vector. Transfection of epimastigotes withlinearized pTREX-VAMP7-GFP was performed as described above.Parasites were selected with G418 at 250 μg/ml to obtain a stablepopulation overexpressing TcVAMP7–GFP.

Cell volume measurementsT. cruzi epimastigotes (wild type and those expressing GFP–TcRab32 ordominant-negative GFP–TcRab32) at log phase of growth (3 days) werecollected at 1600 g for 10 min (at a density of 1×108/ml), and volumemeasurement experiments after stress were performed exactly as describedpreviously (Li et al., 2011).

Recombinant protein expression, purification and antibodygenerationThe DNA sequence corresponding to the entire ORF of TcRab32 wasPCR-amplified fromT. cruziYstraingDNA(forwardprimer: 5′-GACGACGA-CAAGATGTCATACTCGAA-3′, reverse primer: 5′-GAGGAGAAGCC-

CGGTTTAACAGGAGCAGCCCGAC-3′) and inserted into vectorpET32 Ek/LIC using ligation-independent cloning for heterologousexpression in bacteria. The sequence of several recombinant clones wasverified, and they were transformed by heat shock into E. coli BL21 CodonPlus (DE3)-RIPL chemically competent cells. Expression of recombinantprotein was obtained by induction in 0.5 mM isopropyl-β-D-thiogalactopyranoside (IPTG) in Luria Bertani (LB) broth overnight at37°C. His-tagged recombinant protein was purified under denaturingconditions with His-Bind cartridges (Novagen). Recombinant TcRab32was used as the immunogen for production of polyclonal antibody in mice.This antibody was generated at the Monoclonal Antibody Facility of theCollege of Veterinary Medicine, University of Georgia (Athens, GA).

In vitro prenylationIn vitro prenylation reactions were performed as described previously(Cuevas et al., 2005) with minor modifications. A total of 2 µCi of [3H]-FPPor [3H]-GGPP was used as the isoprenoid donor. The assay reaction wascarried out at 30°C for 30 min, 1 h and 3 h. The optimum reaction time forthis assay was 30 min. Products were resolved by using 10% SDS-PAGE.The gel was incubated in EN3HANCE, dried, and exposed to film at −80°Cfor 2 weeks.

H+ and Ca2+ transportCa2+ uptake by digitonin-permeabilized epimastigotes was measured usingthe fluorescence indicator Calcium Green-5N, as described previously(Huang et al., 2013). Acidification of digitonin-permeabilized cells wasfollowed by measuring the changes in the fluorescence of Acridine Orangein a fluorometer, as described previously (Docampo et al., 1995).Differences in Ca2+ and H+ uptake were evaluated by measuring the ratesof Ca2+ or Acridine Orange uptake, respectively, during the first 20 s afteruptake initiation (linear rate), and rates are expressed as changes in arbitraryfluorescence units.

Fura-2 measurementsFura-2 determinations were performed, essentially, as described previously(Docampo et al., 1995). Excitation was at 340 nm and 380 nm, andemission was at 510 nm. The Fura-2 fluorescence response to intracellularCa2+ concentration was calibrated from the ratio of 340/380 nmfluorescence values after subtraction of 340/380 nm fluorescence of thecells at 340 and 380 nm, as described previously (Grynkiewicz et al.,1985).

Quantification of PPi, and short-chain and long-chain polyPCells (2×108) in log phase were harvested and washed twice with buffer A(116 mM NaCl, 5.4 mM KCl, 0.8 mMMgSO4, 50 mM Hepes, pH 7.2, and5.5 mMglucose). The PPi and short-chain polyP were extracted using 0.5 Mperchloric acid (HClO4) (Ruiz et al., 2001), and the long chain polyP wasextracted using glass milk (Molecular Probes) as described previously(Ault-Riche et al., 1998). PPi levels were determined by the amount of Pi

released upon treatment with an excess of Saccharomyces cerevisiaeinorganic pyrophosphatase (catalog no. I-1891, Sigma). The free Pi

(released) amount was determined by using a standard curve. Briefly, theenzymatic reaction was performed on 96-well plates with 50 mM Tris-HCl(pH 7.4), 6 mM MgCl2, inorganic pyrophosphatase and extracted PPisamples at a final volume of 100 µl. After incubation at 30°C for 10 min, thereaction was immediately stopped by the addition of an equal amount ofthe fresh mixture of three parts of 0.045% malachite green with one partof 4.2% ammonium molybdate (Sigma), which was filtered before use.The absorbance at 660 nm was read using a SpectraMax M2e platereader (Molecular Devices, Sunnyvale, CA).

Short-chain and long-chain polyP levels were determined by the amountof Pi released upon treatment with an excess of the purified recombinantexopolyphosphatase of S. cerevisiae (rScPPX1) freshly purified in ourlaboratory (Ruiz et al., 2001)

Short-chain polyP that had been extracted from 5×108 cells was mixedwith 6× dye (0.01% Orange G, 30% glycerol, 10 mM Tris-HCl, pH 7.4,1 mM EDTA) and resolved by using 20% urea-PAGE. Samples were run at

2371

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2363-2373 doi:10.1242/jcs.169466

Journal

ofCe

llScience

Page 10

600 V, 6 mA overnight at 4°C until the Orange G had run through two-thirdsof the gel. Gels were stained with 0.1% Toluidine Blue.

Transmission electron microscopyFor imaging whole epimastigote forms, cells were washed with filteredbuffer A twice, and directly applied to Formvar-coated copper grids,allowed to adhere for 10 min, carefully blotted dry and then observed usinga JEM-1210 electron microscope operating at 80 kV. Whole unfixed wild-type epimastigotes and epimastigotes expressing dominant-negativeTcRab32 were randomly selected, and the number of acidocalcisome percell was counted in 70 cells from two different preparations.

T. cruzi epimastigotes expressing GFP–TcRab32, GFP–TcVAMP7 orTcVAMP7–GFP were washed twice in 0.1 M sodium cacodylate buffer,pH 7.4, and fixed for 1 h on ice with 0.1% glutaraldehyde, 4%paraformaldehyde and 0.1 M sodium cacodylate buffer, pH 7.4. Sampleswere processed for cryo-immunoelectron microscopy at the MolecularMicrobiology Imaging Facility, Washington University School ofMedicine. GFP-fusion protein localization was detected with a polyclonalantibody against GFP (Invitrogen) and anti-rabbit IgG conjugated to gold asa secondary antibody.

Electron tomography 3D reconstructionCells were prepared as previously described (Girard-Dias et al., 2012).Briefly, a pellet of cells was sandwiched between 3×0.5 mm aluminiumcarries (Bal-Tec, Liechtenstein) or inserted into 2-mm pieces of 200-µm Øcellulose capillaries and frozen using a Bal-Tec HPM 010 high-pressurefreezing machine (Bal-Tec, Liechtenstein). Freeze substitution wasperformed at −80°C for 72 h in a medium comprising acetone with 2%osmium tetroxide, 0.1% glutaraldehyde and 1% of water, using a LeicaEMP apparatus. After substitution, samples were embedded in Epon. Forelectron tomography analyses, 200-nm sections or ribbons of serial sectionswere collected in 200 mesh copper grids or onto Formvar-coated slot coppergrids and stained. Finally, the sections were coated with 5 nm of carbon andobserved under a FEI G2 transmission electron microscope (Tecnai G2, FEICompany, Eindhoven) that was equipped with a 4 k×4 k CCD camera(Eagle, FEI Company, Eindhoven). Tilt series from −65° to +65° with anangular increment of 1° were acquired. Alignments were applied usingfiducial markers, and weighted back projections were performed usingIMOD software package (University of Colorado). For segmentation anddata display, AMIRA (Visage Imaging) and IMOD were used.

AcknowledgementsWe thank Melina Galizzi (University of Georgia, GA) for help in the preparation ofT. cruzi infective stages; Melissa Storey andMelinaGalizzi for help in the preparationof the antibody against Rab32; Mary Ard (University of Georgia, GA) for help withtransmission electron microscopy; and Wandy L. Beatty (Washington University,St Louis, MO) for the cryo-immunoelectron microscopy.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsS.N., V.J., W.d.S., K.M. and R.D. designed experiments. S.N., V.J., W.G.-D. andK.M. performed experiments; S.N. and R.D. wrote the manuscript. V.J. and K.M.reviewed the manuscript.

FundingThis work was supported by the US National Institutes of Health (NIH) [grantAI107663 to R.D.]. V.J. was supported by the NIH [grant AI101167]. W.G.-D., W.d.S.and K.M. were supported by Conselho Nacional de Desenvolvimento Cientıfico eTecnologico (CNPq-Universal grants 480184/2012-7, 449256/2014-6, INCT emBiologia Estrutural e Bioimagem); Fundaça o Carlos Chagas Filho de Amparo a Pesquisa do Estado do Rio de Janeiro (FAPERJ-Programa Nucleos Emergentesgrant E-26/111.185/2011); Coordenaça o de Aperfeiçoamento do Pessoal de NıvelSuperior (CAPES); and Financiadora de Estudos e Projetos (FINEP). Deposited inPMC for release after 12 months.

Supplementary materialSupplementary material available online athttp://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.169466/-/DC1

ReferencesAlto, N. M., Soderling, J. and Scott, J. D. (2002). Rab32 is an A-kinase anchoring

protein and participates in mitochondrial dynamics. J. Cell Biol. 158, 659-668.Ault-Riche, D., Fraley, C. D., Tzeng, C. M. and Kornberg, A. (1998). Novel assay

reveals multiple pathways regulating stress-induced accumulations of inorganicpolyphosphate in Escherichia coli. J. Bacteriol. 180, 1841-1847.

Baqui, M. M. A., De Moraes, N., Milder, R. V. and Pudles, J. (2000). A giantphosphoprotein localized at the spongiome region of Crithidia luciliaethermophila. J. Eukaryot. Microbiol. 47, 532-537.

Beranger, F., Paterson, H., Powers, S., de Gunzburg, J. and Hancock, J. F.(1994). The effector domain of Rab6, plus a highly hydrophobic C terminus, isrequired for Golgi apparatus localization. Mol. Cell. Biol. 14, 744-758.

Berriman, M., Ghedin, E., Hertz-Fowler, C., Blandin, G., Renauld, H.,Bartholomeu, D. C., Lennard, N. J., Caler, E., Hamlin, N. E., Haas, B. et al.(2005). The genome of the African trypanosome Trypanosoma brucei. Science309, 416-422.

Besteiro, S., Tonn, D., Tetley, L., Coombs, G. H. and Mottram, J. C. (2008). TheAP3 adaptor is involved in the transport of membrane proteins to acidocalcisomesof Leishmania. J. Cell Sci. 121, 561-570.

Bright, L. J., Kambesis, N., Nelson, S. B., Jeong, B. and Turkewitz, A. P. (2010).Comprehensive analysis reveals dynamic and evolutionary plasticity of RabGTPases and membrane traffic in Tetrahymena thermophila. PLoS Genet. 6,e1001155.

Bultema, J. J. and Di Pietro, S. M. (2013). Cell type-specific Rab32 and Rab38cooperate with the ubiquitous lysosome biogenesis machinery to synthesizespecialized lysosome-related organelles. Small GTPases 4, 16-21.

Bultema, J. J., Ambrosio, A. L., Burek, C. L. and Di Pietro, S. M. (2012). BLOC-2,AP-3, and AP-1 proteins function in concert with Rab38 and Rab32 proteins tomediate protein trafficking to lysosome-related organelles. J. Biol. Chem. 287,19550-19563.

Bultema, J. J., Boyle, J. A., Malenke, P. B., Martin, F. E., Dell’Angelica, E. C.,Cheney, R. E. and Di Pietro, S. M. (2014). Myosin Vc interacts with Rab32 andRab38 proteins and works in the biogenesis and secretion of melanosomes.J. Biol. Chem. 289, 33513-33528.

Clark, T. B. (1959). Comparative morphology of four genera of trypanosomatidae.J. Protozool. 6, 227-232.

Cuevas, I. C., Rohloff, P., Sanchez, D. O. and Docampo, R. (2005).Characterization of farnesylated protein tyrosine phosphatase TcPRL-1 fromTrypanosoma cruzi. Eukaryot. Cell 4, 1550-1561.

Delevoye, C., Hurbain, I., Tenza, D., Sibarita, J.-B., Uzan-Gafsou, S., Ohno, H.,Geerts, W. J. C., Verkleij, A. J., Salamero, J., Marks, M. S. et al. (2009). AP-1and KIF13A coordinate endosomal sorting and positioning during melanosomebiogenesis. J. Cell Biol. 187, 247-264.

Docampo, R. and Moreno, S. N. J. (2011). Acidocalcisomes. Cell Calcium 50,113-119.

Docampo, R., Moreno, S. N. J. and Vercesi, A. E. (1993). Effect of thapsigargin oncalcium homeostasis in Trypanosoma cruzi trypomastigotes and epimastigotes.Mol. Biochem. Parasitol. 59, 305-313.

Docampo, R., Scott, D. A., Vercesi, A. E. and Moreno, S. N. (1995). IntracellularCa2+ storage in acidocalcisomes of Trypanosoma cruzi. Biochem. J. 310,1005-1012.

Docampo, R., de Souza, W., Miranda, K., Rohloff, P. and Moreno, S. N. (2005).Acidocalcisomes - conserved from bacteria to man. Nat. Rev. Microbiol. 3,251-261.

Docampo, R., Jimenez, V., Lander, N., Li, Z.-H. and Niyogi, S. (2013). Newinsights into roles of acidocalcisomes and contractile vacuole complex inosmoregulation in protists. Int. Rev. Cell Mol. Biol. 305, 69-113.

Figarella, K., Uzcategui, N. L., Zhou, Y., LeFurgey, A., Ouellette, M.,Bhattacharjee, H. and Mukhopadhyay, R. (2007). Biochemical characterizationofLeishmaniamajoraquaglyceroporinLmAQP1:possible role in volume regulationand osmotaxis. Mol. Microbiol. 65, 1006-1017.

Girard-Dias,W., Alcantara, C. L., Cunha-e-Silva, N., deSouza,W. andMiranda,K.(2012). On the ultrastructural organization of Trypanosoma cruzi usingcryopreparation methods and electron tomography. Histochem. Cell Biol. 138,821-831.

Grynkiewicz, G., Poenie, M. and Tsien, R. Y. (1985). A new generation of Ca2+

indicators with greatly improved fluorescence properties. J. Biol. Chem. 260,3440-3450.

Harris, E., Yoshida, K., Cardelli, J. and Bush, J. (2001). Rab11-like GTPaseassociates with and regulates the structure and function of the contractile vacuolesystem in dictyostelium. J. Cell Sci. 114, 3035-3045.

Hasne, M.-P., Coppens, I., Soysa, R. and Ullman, B. (2010). A high-affinityputrescine-cadaverine transporter from Trypanosoma cruzi. Mol. Microbiol. 76,78-91.

Heuser, J., Zhu, Q. and Clarke, M. (1993). Proton pumps populate the contractilevacuoles of Dictyostelium amoebae. J. Cell Biol. 121, 1311-1327.

Hirota, Y. and Tanaka, Y. (2009). A small GTPase, humanRab32, is required for theformation of autophagic vacuoles under basal conditions. Cell. Mol. Life Sci. 66,2913-2932.

2372

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2363-2373 doi:10.1242/jcs.169466

Journal

ofCe

llScience

Page 11

Huang, G., Fang, J., Sant’Anna, C., Li, Z.-H., Wellems, D. L., Rohloff, P. andDocampo, R. (2011). Adaptor protein-3 (AP-3) complex mediates the biogenesisof acidocalcisomes and is essential for growth and virulence of Trypanosomabrucei. J. Biol. Chem. 286, 36619-36630.

Huang, G., Vercesi, A. E. and Docampo, R. (2013). Essential regulation of cellbioenergetics in Trypanosoma brucei by the mitochondrial calcium uniporter. Nat.Commun. 4, 2865.

Jean, S. and Kiger, A. A. (2012). Coordination between RAB GTPase andphosphoinositide regulation and functions. Nat. Rev. Mol. Cell Biol. 13, 463-470.

Jeffries, T. R., Morgan, G. W. and Field, M. C. (2001). A developmentally regulatedRab11 homologue in Trypanosoma brucei is involved in recycling processes.J. Cell Sci. 114, 2617-2626.

Jimenez, V. and Docampo, R. (2012). Molecular and electrophysiologicalcharacterization of a novel cation channel of Trypanosoma cruzi. PLoS Pathog.8, e1002750.

Lander, N., Ulrich, P. N. and Docampo, R. (2013). Trypanosoma brucei vacuolartransporter chaperone 4 (TbVtc4) is an acidocalcisome polyphosphate kinaserequired for in vivo infection. J. Biol. Chem. 288, 34205-34216.

Lemercier, G., Dutoya, S., Luo, S., Ruiz, F. A., Rodrigues, C. O., Baltz, T.,Docampo, R. and Bakalara, N. (2002). A vacuolar-type H+-pyrophosphatasegoverns maintenance of functional acidocalcisomes and growth of the insect andmammalian forms of Trypanosoma brucei. J. Biol. Chem. 277, 37369-37376.

Li, Z.-H., Alvarez, V. E., De Gaudenzi, J. G., Sant’Anna, C., Frasch, A. C. C.,Cazzulo, J. J. andDocampo, R. (2011). Hyperosmotic stress induces aquaporin-dependent cell shrinkage, polyphosphate synthesis, amino acid accumulation,and global gene expression changes in Trypanosoma cruzi. J. Biol. Chem. 286,43959-43971.

Linder, J. C. and Staehelin, L. A. (1979). A novel model for fluid secretion by thetrypanosomatid contractile vacuole apparatus. J. Cell Biol. 83, 371-382.

Lu, H.G., Zhong,L., deSouza,W., Benchimol,M.,Moreno,S.N. andDocampo,R.(1998). Ca2+ content and expression of an acidocalcisomal calcium pump areelevated in intracellular forms of Trypanosoma cruzi. Mol. Cell. Biol. 18,2309-2323.

Marchesini, N., Ruiz, F. A., Vieira, M. and Docampo, R. (2002). Acidocalcisomesare functionally linked to the contractile vacuole of Dictyostelium discoideum.J. Biol. Chem. 277, 8146-8153.

Martinez-Arca, S., Rudge, R., Vacca, M., Raposo, G., Camonis, J., Proux-Gillardeaux, V., Daviet, L., Formstecher, E., Hamburger, A., Filippini, F. et al.(2003). A dual mechanism controlling the localization and function of exocyticv-SNAREs. Proc. Natl. Acad. Sci. USA 100, 9011-9016.

Mendoza, M., Mijares, A., Rojas, H., Rodrı´guez, J. P., Urbina, J. A. and DiPolo, R.(2002).Physiological andmorphological evidences for thepresenceacidocalcisomesin Trypanosoma evansi: single cell fluorescence and 31P NMR studies. Mol.Biochem. Parasitol. 125, 23-33.

Moniakis, J., Coukell, M. B. and Janiec, A. (1999). Involvement of theCa2+-ATPase PAT1 and the contractile vacuole in calcium regulation inDictyostelium discoideum. J. Cell Sci. 112, 405-414.

Montalvetti, A., Rohloff, P. and Docampo, R. (2004). A functional aquaporinco-localizes with the vacuolar proton pyrophosphatase to acidocalcisomes andthe contractile vacuole complex of Trypanosoma cruzi. J. Biol. Chem. 279,38673-38682.

Moreno, S. N. J., Vercesi, A. E., Pignataro, O. P. and Docampo, R. (1992).Calcium homeostasis in Trypanosoma cruzi amastigotes: presence of inositolphosphates and lack of an inositol 1,4,5-trisphosphate-sensitive calcium pool.Mol. Biochem. Parasitol. 52, 251-261.

Moreno, S. N., Silva, J., Vercesi, A. E. and Docampo, R. (1994). Cytosolic-freecalcium elevation in Trypanosoma cruzi is required for cell invasion. J. Exp. Med.180, 1535-1540.

Moreno-Sanchez, D., Hernandez-Ruiz, L., Ruiz, F. A. and Docampo, R. (2012).Polyphosphate is a novel pro-inflammatory regulator of mast cells and is located inacidocalcisomes. J. Biol. Chem. 287, 28435-28444.

Munro, S. (2002). Organelle identity and the targeting of peripheral membraneproteins. Curr. Opin. Cell Biol. 14, 506-514.

Nepomuceno-Silva, J. L., Yokoyama, K., de Mello, L. D. B., Mendonca, S. M.,Paixao, J. C., Baron, R., Faye, J.-C., Buckner, F. S., Van Voorhis, W. C., Gelb,M. H. et al. (2001). TcRho1, a farnesylated Rho family homologue fromTrypanosoma cruzi: cloning, trans-splicing, and prenylation studies. J. Biol.Chem. 276, 29711-29718.

Niyogi, S., Mucci, J., Campetella, O. and Docampo, R. (2014). Rab11 regulatestrafficking of trans-sialidase to the plasma membrane through the contractilevacuole complex of Trypanosoma cruzi. PLoS Pathog. 10, e1004224.

Park, M., Serpinskaya, A. S., Papalopulu, N. and Gelfand, V. I. (2007). Rab32regulates melanosome transport in Xenopus melanophores by protein kinase arecruitment. Curr. Biol. 17, 2030-2034.

Pfeffer, S. R. (2001). Rab GTPases: specifying and deciphering organelle identityand function. Trends Cell Biol. 11, 487-491.

Rak, A., Pylypenko, O., Durek, T., Watzke, A., Kushnir, S., Brunsveld, L.,Waldmann, H., Goody, R. S. and Alexandrov, K. (2003). Structure of Rab GDP-dissociation inhibitor in complex with prenylated YPT1 GTPase. Science 302,646-650.

Raposo, G. and Marks, M. S. (2007). Melanosomes–dark organelles enlightenendosomal membrane transport. Nat. Rev. Mol. Cell Biol. 8, 786-797.

Rohloff, P. and Docampo, R. (2008). A contractile vacuole complex is involved inosmoregulation in Trypanosoma cruzi. Exp. Parasitol. 118, 17-24.

Rohloff, P., Montalvetti, A. and Docampo, R. (2004). Acidocalcisomes and thecontractile vacuole complex are involved in osmoregulation in Trypanosoma cruzi.J Biol. Chem. 279, 52270-52281.

Ruiz, F. A., Rodrigues, C. O. and Docampo, R. (2001). Rapid changes inpolyphosphate content within acidocalcisomes in response to cell growth,differentiation, and environmental stress in Trypanosoma cruzi. J. Biol. Chem.276, 26114-26121.

Ruiz, F. A., Lea, C. R., Oldfield, E. andDocampo, R. (2004). Human platelet densegranules contain polyphosphate and are similar to acidocalcisomes of bacteriaand unicellular eukaryotes. J. Biol. Chem. 279, 44250-44257.

Schmatz, D. M. and Murray, P. K. (1982). Cultivation of Trypanosoma cruzi inirradiated muscle cells: improved synchronization and enhanced trypomastigoteproduction. Parasitology 85, 115-125.

Seabra, M. C. and Wasmeier, C. (2004). Controlling the location and activation ofRab GTPases. Curr. Opin. Cell Biol. 16, 451-457.

Sesaki, H., Wong, E. F. S. and Siu, C.-H. (1997). The cell adhesion moleculeDdCAD-1 in Dictyostelium is targeted to the cell surface by a nonclassicaltransport pathway involving contractile vacuoles. J. Cell Biol. 138, 939-951.

Smith, S. A., Mutch, N. J., Baskar, D., Rohloff, P., Docampo, R. and Morrissey,J. H. (2006). Polyphosphate modulates blood coagulation and fibrinolysis. Proc.Natl. Acad. Sci. USA 103, 903-908.

Sriskanthadevan, S., Lee, T., Lin, Z., Yang, D. and Siu, C.-H. (2009). Celladhesion molecule DdCAD-1 is imported into contractile vacuoles by membraneinvagination in a Ca2+- and conformation-dependent manner. J. Biol. Chem. 284,36377-36386.

Tamura, K., Ohbayashi, N., Maruta, Y., Kanno, E., Itoh, T. and Fukuda, M. (2009).Varp is a novel Rab32/38-binding protein that regulates Tyrp1 trafficking inmelanocytes. Mol. Biol. Cell 20, 2900-2908.

Tamura, K., Ohbayashi, N., Ishibashi, K. and Fukuda, M. (2011). Structure-function analysis of VPS9-ankyrin-repeat protein (Varp) in the trafficking oftyrosinase-related protein 1 in melanocytes. J. Biol. Chem. 286, 7507-7521.

Theos, A. C., Tenza, D., Martina, J. A., Hurbain, I., Peden, A. A., Sviderskaya,E. V., Stewart, A., Robinson, M. S., Bennett, D. C., Cutler, D. F. et al. (2005).Functions of adaptor protein (AP)-3 and AP-1 in tyrosinase sorting fromendosomes to melanosomes. Mol. Biol. Cell 16, 5356-5372.

Turkewitz, A. P. and Bright, L. J. (2011). A Rab-based view of membrane traffic inthe ciliate Tetrahymena thermophila. Small GTPases 2, 222-226.

Ulrich, P. N., Jimenez, V., Park, M., Martins, V. P., Atwood, J., III, Moles, K.,Collins, D., Rohloff, P., Tarleton, R., Moreno, S. N. J. et al. (2011). Identificationof contractile vacuole proteins in Trypanosoma cruzi. PLoS ONE 6, e18013.

Ulrich, P. N., Lander, N., Kurup, S. P., Reiss, L., Brewer, J., Soares Medeiros,L. C., Miranda, K. and Docampo, R. (2014). The acidocalcisome vacuolartransporter chaperone 4 catalyzes the synthesis of polyphosphate in insect-stages of Trypanosoma brucei and T. cruzi. J. Eukaryot. Microbiol. 61, 155-165.

Urbina, J. A., Moreno, B., Vierkotter, S., Oldfield, E., Payares, G., Sanoja, C.,Bailey, B. N., Yan, W., Scott, D. A., Moreno, S. N. J. et al. (1999). Trypanosomacruzi contains major pyrophosphate stores, and its growth in vitro and in vivo isblocked by pyrophosphate analogs. J. Biol. Chem. 274, 33609-33615.

Yokoyama, K., Gillespie, J. R., Van Voorhis, W. C., Buckner, F. S. and Gelb,M. H. (2008). Protein geranylgeranyltransferase-I of Trypanosoma cruzi. Mol.Biochem. Parasitol. 157, 32-43.

Zerial, M. and McBride, H. (2001). Rab proteins as membrane organizers. Nat.Rev. Mol. Cell Biol. 2, 107-117.

2373

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2363-2373 doi:10.1242/jcs.169466

Journal

ofCe

llScience

Page 12

1

Fig. S1. TcRab32 purification and sequence information. (A) Eluted and desalted fractions (E1 to E9) obtained during recombinant TcRab32 purification from E. coli as analyzed by SDS PAGE showing a band of correct size (42 kDa) corresponding to the His-tagged protein. The 10% SDS PAGE gel was stained with Coommassie blue. (B) Comparison of the deduced amino acid sequence of T. cruzi Rab32 with human Rab32 (HsRab32). The presence of the “WDIAGQE” and C terminal “CSC” domain is boxed in red and the presence of phenylalanine “F” at position 194 is denoted in blue asterisk.

Journal of Cell Science | Supplementary Material

Page 13

2

Fig. S2. TcRab32 localization in different life stages of T. cruzi. (A, B) TcRab32 was

detected in the contractile vacuole bladder of epimastigotes (Epi), trypomastigotes

(Trypo), and amastigotes (Ama) (arrows in B) (note the circular profile) with additional

punctated staining in Epi and Trypo using specific antibodies against TcRab32 and

shown by IFA (A) and differential interference contrast microscopy merged with DAPI

(blue) and antibody (red) staining (B). Scale bars = 10 µm. (C) Western blot analysis of

lysates from different life cycle stages. Arrow shows the 26-kDa bands corresponding to

TcRab32. Amastigotes show a double band. Tubulin (Tub) antibody was used as a

loading control. Molecular markers are on the left. (D, E) Cryo-immunogold electron

microscopy analysis of GFP-TcRab32 expressing epimastigotes as investigated using

antibodies against GFP. K, kinetoplast; F, flagellum; FP, flagellar pocket (arrow); CV,

contractile vacuole. Scale bars = 200 nm.

Journal of Cell Science | Supplementary Material

Page 14

3

Fig. S3. Localization of GFP-TcVAMP7 in the contractile vacuole complex as detected by cryo-immunogold electron microscopy. (A, B) Gold particles are seen in the CV, spongiome (Sp), flagellar pocket (FP), and plasma membrane (arrows). K, kinetoplast, N, nucleus; V, vacuole. Scale bars = 2.0 µm. (C) Western blot analyses of lysates from wild type epimastigotes (WT) or epimastigotes expressing GFP-TcVAMP7, TcVAMP7-GFP, or GFP. Arrows show the bands corresponding to the tagged proteins and GFP. Molecular weight markers are on the left side.

Journal of Cell Science | Supplementary Material

Page 15

4

Fig. S4. Localization of GFP-TcRab32DN in trypomastigotes and amastigotes. Cytosolic localization of GFP-TcRab32DN in trypomastigotes and amastigotes was detected using antibodies against GFP (green). DNA was stained with DAPI (blue). Scale bars = 10 µm.

Journal of Cell Science | Supplementary Material

Page 16

Movie 1. Interaction of TcVAMP7-GFP-labeled acidocalcisomes with the contractile vacuole. Live-cell fluorescence imaging of epimastigotes expressing TcVAMP7-GFP (green) labeled with BODIPY-TR C5 Ceramide (red). It can be observed that under hypos-motic stress, the cells swell and the CVC becomes more evident (ring shape). TcVAMP7-GFP-labeled acidocalcisomes initially move towards the CVC but after 130 sec (compensatory phase of RVD) some of them can be seeing moving away from it.

Movie 2. Interaction of TcVAMP7-GFP-labeled acidocalcisomes with the contractile vacuole. Same as movie 1. Overlay of the fluorescence and DIC images.

Journal of Cell Science | Supplementary Material

Page 17

Movie 3. Virtual slices along a tomogram obtained from the anterior region of the parasite. Inspection of the virtual slices along the sample volume shows the structure of the contractile vacuole, containing a bladder and the interconnecting tubules of the spon-giome. Slice #98 shows the contact of an acidocalcisome with the bladder of the contractile vacuole. The CVC (blue) and the fusing acidocalcisome (orange) were modeled. Sectioning of the model evidenced the contact region.

Journal of Cell Science | Supplementary Material