CRISPR/Cas9-mediated endogenous C-terminal Tagging of Trypanosoma cruzi Genes Reveals the Acidocalcisome Localization of the Inositol 1,4,5-Trisphosphate Receptor*

Methods for genetic manipulation of Trypanosoma cruzi, the etiologic agent of Chagas disease, have been highly inefficient, and no endogenous tagging of genes has been reported to date. We report here the use of the CRISPR (clustered regularly interspaced short palindromic repeats)/Cas9 (CRISPR-associated gene 9) system for endogenously tagging genes in this parasite. The utility of the method was established by tagging genes encoding proteins of known localization such as TcFCaBP (flagellar calcium binding protein) and TcVP1 (vacuolar proton pyrophosphatase), and two proteins of undefined or disputed localization, the TcMCU (mitochondrial calcium uniporter) and TcIP3R (inositol 1,4,5-trisphosphate receptor). We confirmed the flagellar and acidocalcisome localization of TcFCaBP and TcVP1 by co-localization with antibodies to the flagellum and acidocalcisomes, respectively. As expected, TcMCU was co-localized with the voltage-dependent anion channel to the mitochondria. However, in contrast to previous reports and our own results using overexpressed TcIP3R, endogenously tagged TcIP3R showed co-localization with antibodies against VP1 to acidocalcisomes. These results are also in agreement with our previous reports on the localization of this channel to acidocalcisomes of Trypanosoma brucei and suggest that caution should be exercised when overexpression of tagged genes is done to localize proteins in T. cruzi.

The application of the CRISPR/Cas9 technology to the study of protist parasites has dramatically increased the tools available for their genetic manipulation (1). Trypanosoma cruzi, the etiologic agent of Chagas disease, which is a significant cause of morbidity and mortality from the South of the United States to the South of Argentina and Chile, has been particularly refractory to genetic manipulation. However, the recent use of the CRISPR/Cas9 technology to knock down or knock out genes (2,3) has revolutionized their study.
The localization of proteins is important to determine their cellular function, and previous studies in T. cruzi have used either antibodies or gene tagging methods with vectors that overexpressed the proteins (4). Although specific antibodies are useful to detect the endogenous proteins, it is not always possible to obtain them because either the proteins have low antigenicity or the antibodies cross-react with other proteins. Plasmids that enable the tagging of genes at their endogenous loci are not available for T. cruzi, and a major drawback of the overexpression of tagged proteins is that the proteins of interest sometimes are retained in the endoplasmic reticulum (ER) 4 or localize to other compartments.
Here we have adapted the CRISPR/Cas9 system to tag genes of T. cruzi at their endogenous loci and tested this system with two genes encoding proteins of well recognized localization (flagellar calcium binding protein or TcFCaBP and the acidocalcisome vacuolar proton pyrophosphatase or TcVP1) and two encoding proteins of undefined or disputed localization (mitochondrial calcium uniporter or TcMCU and inositol 1,4,5-trisphosphate receptor or TcIP 3 R). TcMCU was functionally characterized more than 27 years ago as the calcium channel that transports Ca 2ϩ into the mitochondria of the parasites (5,6), and this finding was fundamental for the recent discovery of the gene encoding this channel in mammalian cells (7)(8)(9). The channel is localized to the inner mitochondrial membrane of a variety of cells, including Trypanosoma brucei (10). TcIP 3 R was reported to have ER localization in T. cruzi (11). However, the immunofluorescence evidence reported was disputed (12), because there was no clear reticular pattern or co-localization with a T. brucei ER marker, TbBiP, in the figures published (11). In addition, the T. brucei IP 3 R localized to the acidocalcisomes as demonstrated using antibodies against the endogenous tagged protein (13) and specific antibodies against the protein (14), as well as proteomic and functional analyses (13,14). In this work, we report the acidocalcisome localization of TcIP 3 R.
The use of the CRISPR/Cas9 system for C-terminal tagging of genes was recently reported for three parasites: Toxoplasma gondii (15), Plasmodium yoelii (16), and Leishmania donovani (17), but has not been previously used in T. cruzi. The availability of this technique for T. cruzi has great potential for the functional analysis of proteins in this parasite.

Results
We first evaluated the endogenous C-terminal tagging method by introducing the epitope tag sequence into two different genes: the TcFCaBP gene and the TcVP1 gene. The proteins encoded by these genes are localized in well defined organelles in trypanosomes: flagellum (18) and acidocalcisomes (19), respectively. Monoclonal and polyclonal antibodies recognizing these proteins are available, as well as genetic information about the proteins. For 3ϫHA C-terminal tagging, we cotransfected a specific 3Ј end-sgRNA/Cas9/pTREX construct with a specific DNA donor molecule for each gene amplified from the pMOTag-HX1-4H vector (Fig. 1A), whereas for 3ϫc-Myc C-terminal tagging, we co-transfected the same 3Ј end-sgRNA/Cas9/pTREX constructs with a specific DNA donor molecule for each gene amplified from the pMOTag23M vector (Fig. 1B), as described under "Experimental Procedures." We obtained G418/hygromycin-or G418/puromycin-resistant cell lines after 5 weeks under selective pressure. Transfectants were analyzed by PCR, using gDNA isolated from each one, and specific primer sets to distinguish between the wild type and the tagged cell lines. Fig. 2A shows that TcVP1-3ϫHA transfectants were efficiently tagged at the endogenous locus, because the corre-sponding band amplified with a reverse primer annealing on the hygromycin marker is only present in the resistant parasites (lane HA) but absent in the WT cells, which is the negative control of the reaction. We analyzed the TcVP1-3ϫHA transfectants by Western blotting, using commercial antibodies anti-HA tag, and a band of ϳ85 kDa was clearly detected on the transfectant but absent in the wild type cells (Fig. 2B). A similar band appears in both wild type cells and TcVP1-3ϫHA transfectants when anti-TbVP1 antibodies were used (Fig. 2B). Immunofluorescence analysis (IFA) of the mutants verified the subcellular localization of the protein to the acidocalcisomes as it co-localizes with antibodies against TbVP1 (Fig. 2C), as expected (19). Similar results were obtained by 3ϫc-Myc C-terminal tagging using specific DNA donor molecules amplified from the pMOTag23M vector (Fig. 2, D-F). Site-specific insertion of DNA donor cassettes at the 3Ј end of TcVP1 gene was verified by cloning and sequencing PCR products amplified from gDNA extracted from TcVP1-3ϫHA and TcVP1-3ϫc-Myc cell lines (Fig. 3A), confirming that the mechanism of homologous-directed DNA repair (HDR) took placed in almost the entire population, and a tagging efficiency of Ͼ95% was observed in both cell lines by IFA (data shown for TcVP1-3ϫc-Myc; Fig. 3B). These results also indicate that it is feasible to use the intergenic tubulin region of T. brucei as trans-splicing signal for T. cruzi.
Using similar procedures we found that the TcFCaBP was efficiently tagged at the endogenous locus, as detected by PCR (Fig. 4, A and D), Western blotting analyses (Fig. 4, B and E) and IFA of epimastigotes generated using specific DNA donor molecules amplified from either the pMOTag-HX1-4H (Fig. 4C) or the pMOTag23M vector (Fig. 4F). TcFCaBP-3ϫHA and TcFCaBP-3ϫc-Myc exclusively localize in the flagellum, the expected localization of this protein (18), as shown by their The T. cruzi HX1 trans-splicing signal is located between the 3ϫHA tag sequence and the gene that confers resistance to hygromycin (Hygro (R)). HR1 Fw and HR2 Rv ultramers indicate oligonucleotides used to amplify DNA donor cassette. The annealing regions for ultramers to pMOTag-HX1-4H and genomic DNA (gDNA) are indicated in black and blue, respectively. Panel ii, a double-stranded gDNA break was produced by Cas9 targeted by the sgRNA both expressed from 3Ј end-sgRNA/Cas9/pTREX plasmid downstream the STOP codon of the gene of interest (GOI) in the endogenous locus. Homologous directed repair was induced co-transfecting epimastigotes with the DNA donor cassette, containing homologous regions to the GOI 3Ј end (blue) and to the GOI 3ЈUTR (light blue). Panel iii, integration of 3ϫHA and antibiotic resistance gene at 3Ј end of GOI by homologous recombination. Arrows indicate primers used for checking integration of donor DNA. B, panel i, pMOTag23M vector map. The 3ϫc-Myc tag sequence and the puromycin resistance gene (Puro (R)) are separated by the T. brucei tubulin intergenic region (Tigr). The rest of the description of panels i-iii is similar to that for A. Hygro, hygromycin resistance gene; Puro, puromycin resistance gene; UTR, untranslated region; ATG, start codon.
co-localization with monoclonal antibodies against TcFCaBP (Fig. 4, C and F), which recognizes both the tagged and endogenous proteins by Western blotting analyses (Fig. 4, B and E). In both cases, detection of the endogenous non-tagged TcFCaBP was much stronger than the tagged version of the protein. We attribute this result to the fact that TcFCaBP is encoded by three identical copies of the gene arranged in tandem in the T. cruzi genome, and probably not all of them were tagged. The localization of C-terminal tagged TcVP1 and TcFCaBP at the expected compartments indicates that the method used is appropriate to detect the native localization of proteins in T. cruzi and that the two vectors employed, one of them designed for endogenous tagging of genes in T. brucei, are adequate for this purpose.
We next investigated the localization of two proteins for which either no previous localization studies have been reported (TcMCU) (20) or for which its localization has been disputed (TcIP 3 R) (12). TcMCU is the T. cruzi orthologue of the recently discovered MCU from vertebrate cells (8,9) and of TbMCU (10). MCU localizes to the inner membrane of mitochondria in both vertebrate cells (8,9) and T. brucei (10) and is solely responsible for mitochondrial Ca 2ϩ uptake in T. brucei (10). Functional studies done in T. cruzi clearly established the presence of MCU in these cells (5,6) and were important for the identification of the molecular nature of this channel in vertebrate cells (7). Using the same technique that we used to localize TcVP1 and TcFCaBP (see above), we found the TcMCU was tagged at the endogenous locus, as detected by PCR (Fig. 5, A and D), Western blotting analyses (Fig. 5, B and E), and IFA of cells obtained using specific DNA donor molecules amplified from either the pMOTag-HX1-4H (Fig. 5C) or the pMOTag23M vector (Fig. 5F). TcMCU co-localized with the mitochondrial voltage-dependent anion channel (VDAC; Fig. 5, C and F), as expected.
Before doing endogenous tagging of TcIP 3 R, we overexpressed the gene with an HA epitope tag (TcIP 3 R-HA-OE) to investigate the localization of the overexpressed protein. Fig.  6A shows the Western blotting analysis of lysates from WT and TcIP 3 R-HA-OE epimastigotes (IP3R-HA) incubated with anti-HA antibodies showing that transfected cells express the tagged protein of the expected size (ϳ340 kDa). Fig. 6B shows that the overexpressed protein does not co-localize with the acidocalcisome marker TcVP1, as detected with antibodies against HA and TbVP1, respectively. However, TcIP 3 R-HA-OE cells shows the same perinuclear and reticular localization pattern as BiP, an ER marker (21), as detected with antibodies against HA and TbBiP (Fig. 6C). Note that although the same distribution pattern is observed for both proteins, their localization is in general not superimposable. This is probably due to the membrane localization of TcIP 3 R-HA-OE, the intra-ER localization of the soluble BiP, and the fact that these images were deconvolved to eliminate background fluorescence. Fig. 7 shows the efficient tagging of TcIP 3 R at the endogenous locus, as detected by PCR (Fig. 7, A and E), Western blotting analyses (Fig. 7, B and F), and IFA of tagged cell lines generated  DECEMBER 2, 2016 • VOLUME 291 • NUMBER 49 using specific DNA donor molecules amplified from either the pMOTag-HX1-4H (Fig. 7, C and D) or the pMOTag23M vector (Fig. 7, G and H). TcIP 3 R-3ϫHA and TcIP 3 R-3ϫc-Myc localize to the acidocalcisomes, as previously described in T. brucei (13) and shown by the co-localization of anti-HA and anti-c-Myc antibodies with VP1 ( Fig. 7, C and G). The anti-HA and anti-cMyc antibodies recognize the tagged proteins but not the endogenous IP 3 R in the WT by Western blotting analyses (Fig. 7, B and F). Fig. 7 (D and H) shows that there is no significant co-localization with the reticular distribution of TbBiP antibodies in the ER. Some co-localization with TbBiP antibodies, especially using TcIP 3 R-3ϫc-Mycs was also detected (Fig.   7H) and could correspond to the site of synthesis of the TcIP 3 R in the ER.

Discussion
Our work demonstrates that the use of the CRISPR/Cas9 system in T. cruzi is not limited to loss of function studies (gene deletion/disruption/mutation) (2, 3) but could be used for C-terminal gene tagging. As proof of concept of the methodology employed, we confirmed the localization of TcVP1 and TcFCaBP to the acidocalcisomes and flagellum of the parasite, respectively. To our knowledge, this is the first report of endogenous tagging of proteins in T. cruzi. We also report the mito-

CRISPR/Cas9-mediated Endogenous Tagging in T. cruzi
chondrial localization of TcMCU, the previously identified pore of the mitochondrial Ca 2ϩ uniporter complex (22). In addition, we report the acidocalcisome localization of TcIP 3 R.
We previously reported in T. cruzi the HDR mechanism for double-stranded break repair in CRISPR/Cas9-induced PFR2 knock-out cell line (3). In that work we used a DNA donor molecule with 100-nt homology regions to induce DNA repair by homologous recombination in this organism, generating a homogeneous population where 100% cells exhibited gene disruption. Now, by providing a DNA donor template for CRISPR/ Cas9-mediated gene tagging, we confirmed the high efficiency of this mechanism, because no other DNA repair mechanism was detected by sequencing in TcVP1-3ϫHA and TcVP1-3ϫc-Myc homogeneously tagged populations (Fig. 3A). High efficiency (Ͼ95%) gene tagging was observed in all tagged cell lines generated in this study, using DNA donor templates con-   DECEMBER 2, 2016 • VOLUME 291 • NUMBER 49

JOURNAL OF BIOLOGICAL CHEMISTRY 25509
taining 100-nt homology arms, which makes this methodology a promising tool for cellular localization studies and immunoprecipitation assays.
Although most vertebrate IP 3 Rs reside in ER membranes, IP 3 can stimulate Ca 2ϩ release from the Golgi complex (23), the nucleus (24), and the secretory granules (25) of mammalian cells. IP 3 Rs can also be targeted to the plasma membrane, where they are important for Ca 2ϩ entry (26). Secretory granules of a variety of cells (27)(28)(29)(30)(31) were reported to possess IP 3 Rs, although this was challenged (32). An IP 3 R localizes to the contractile vacuole in Paramecium tetraurelia, as detected with specific antibodies (33). The acidocalcisome localization of

CRISPR/Cas9-mediated Endogenous Tagging in T. cruzi
TbIP 3 R in T. brucei was demonstrated by endogenous gene tagging (13), by studies using specific antibodies (14), and by proteomic (14) and functional (13) studies. Interestingly, when TbIP 3 R (13) or the IP 3 Rs of Capsaspora owczarzaki (another protist in which biochemical characterization of the channel was done (34)) are expressed in DT40 -3KO cells (chicken lymphocytes in which the three vertebrate IP 3 R have been knocked out), the proteins localize to the ER.
It was therefore puzzling that in the related trypanosomatid T. cruzi, the IP 3 R had an ER localization, as studied in epimastigotes overexpressing the channel tagged with GFP (11). This was also against our proteomic analysis of acidocalcisomes of T. cruzi, 5 which supported the acidocalcisome localization. It has been reported that when overexpressed, membrane-targeted GFP fusion proteins have a propensity to form organelle aggregates that may lead to misinterpretations of sorting pathways of trafficked proteins (35). We found that this was indeed the case with the TcIP 3 R. Overexpression of TcIP 3 R resulted in the same pattern of perinuclear and reticular localization than that of the ER marker BiP. However, endogenous gene tagging of TcIP 3 R using CRISPR/Cas9 revealed the acidocalcisome localization of the channel. The presence of a Ca 2ϩ uptake pump (Ca 2ϩ -ATPase) (36) and a Ca 2ϩ release channel (TcIP 3 R, this work) in acidocalcisomes of T. cruzi suggests an important role of the organelles in Ca 2ϩ signaling. The results also indicate that caution should be exercised when overexpression of tagged genes is done to localize proteins in T. cruzi.
In summary, the tools developed in this work will enable rapid endogenous gene tagging in T. cruzi, allowing us to establish the localization of proteins and to gain insight into their function. The molecular tools available in T. cruzi have lagged behind those developed for T. brucei (4). Our results indicate that it is possible to use the intergenic tubulin region of T. brucei as trans-splicing signal for T. cruzi, which expands the molecular toolbox available for T. cruzi, because pMOTag vectors (37) developed for C-terminal tagging in T. brucei could be also used for T. cruzi. The method developed in this work will facilitate the functional analysis of genes in T. cruzi, as well as physiological studies, allowing the identification of targets for drugs, diagnostics, and vaccines.
Cell Culture-T. cruzi Y strain epimastigotes were cultured in liver infusion tryptose medium containing 10% heat-inactivated FBS at 28°C (40). CRISPR/Cas9 mutant cell lines were maintained in medium containing 250 g/ml G418 and 5 g/ml puromycin or 350 g/ml hygromycin. We determined the growth rate of epimastigotes by counting cells in a Neubauer chamber.
sgRNA targeting the 3Ј end of these genes were designed to induce the double-stranded break by Cas9 nuclease downstream their stop codons. Chimeric sgRNAs were obtained by PCR from plasmid pUC_sgRNA as previously described (3) using specific oligonucleotides ( Table 1, primers 1-5), which include a BamHI restriction site, the 20-nt specific protospacer region, and a 20-nt sequence that anneals to the sgRNA backbone. Subsequently these sgRNAs were cloned into Cas9/ pTREX-n vector through BamHI site. To avoid Cas9 off-targeting, protospacers were analyzed with ProtoMatch 1.0 script (15).
For the generation of a DNA donor cassette (DNA template to induce homologous-directed DNA repair) containing the tag sequence and a marker for antibiotic resistance, we used a modified version of the pMOTag4H vector (37), where the T. brucei tubulin intergenic region for trans-splicing was replaced by the

TABLE 1 Oligonucleotides used in this study
Bold uppercase, specific protospacer; italic underlined uppercase, restriction site; lowercase, sgRNA annealing region; bold underlined uppercase, gene-specific homologous region; italic lowercase, pMOTag vector annealing region, italic bold uppercase, stop codon; bold double-underlined uppercase, HA tag sequence. HX1 T. cruzi trans-splicing signal present in the pTREX vector (41). The HX1 fragment was amplified using primers 6 and 7 from Table 1 and cloned into pMOTag4H vector by SalI/Hin-dIII restriction sites. We named the resulting vector pMOTag-HX1-4H, and it allowed the generation of a DNA donor molecule containing a 3ϫHA tag and the hygromycin resistance marker (Fig. 1A). We also used the pMOTag23M vector (Fig.  1B) developed for C-terminal tagging in T. brucei (37). This vector contains a 3ϫc-Myc tag and the puromycin resistance gene. Templates for homologous recombination were amplified by PCR with 120-bp ultramers, of which 100 bp correspond to regions located right upstream of the stop codon (forward primer) and downstream of the Cas9 target site (reverse primer) of the TcVP1, TcFCaBP, TcMCU, and TcIP 3 R genes ( Table 1, primers 8 -15) and 20 bp for annealing on the pMOTag-HX1-4H and pMOTag23M vectors that were used as PCR templates (Fig. 1, A and B). PCRs were carried out using the following cycling conditions: initial denaturation for 2 min at 95°C followed by 40 cycles of 20 s at 95°C, 20 s at 63°C, and 1 min 40 s at 72°C and then a final extension for 10 min at 72°C.
Epimastigotes co-transfected with specific combinations of 3Ј end-sgRNA/Cas9/pTREX-n and DNA donor were cultured for 5 weeks with G418 and puromycin or hygromycin for selection of double resistant parasites. Endogenous gene tagging was verified by PCR from gDNA using gene-specific primer sets (Table 1, primers 16 -21) and Western blotting analysis. All constructs were verified by sequencing.
Overexpression of TcIP 3 R-The TcIP 3 R gene (9044 nt) including the C-terminal human influenza HA tag was cloned into the pTREX-n vector following a three-step design and subsequently transfected to epimastigotes. Briefly, PCR was performed with primer set Fw_TcIP 3 R-XbaI/Rv_TcIP 3 R-BglII for N-terminal region (N), and primer set Fw_TcIP 3 R-BglII/ Rv_TcIP 3 R-HA-XbaI_BglII for C-terminal region (C) ( Table 1, primers 23-26), using T. cruzi Y strain gDNA as template. The amplified fragment N was cloned into pET-32 EK/LIC vector (Novagen) (N-TcIP 3 R/pET-32) by XbaI and BglII restriction sites. Then amplified fragment C was cloned by BglII into plasmid N-TcIP 3 R/pET-32, previously treated with Antarctic phosphatase, to obtain the TcIP 3 R/pET-32 plasmid. Next, the full sequence of TcIP 3 R-HA was excised with XbaI from TcIP 3 R/pET-32 plasmid and subcloned into dephosphorylated pTREX-n vector by XbaI to generate the TcIP 3 R-HA-OE/ pTREX-n plasmid. Insert orientation was determined by PCR and sequencing.
PCR Analysis of Transfected Epimastigotes-Genomic DNA of double-resistant transfectants was used as template in PCRs to verify the integration of the DNA donor molecules into the 3Ј end of the tagged genes. In each PCR was included a genespecific forward primer (Table 1, primers 16 -19) and a reverse primer annealing at 3Ј end of the antibiotic marker present in the donor DNA (Table 1, primers 20 and 21). PCR conditions in a 25-l reaction volume using GoTaq DNA polymerase with ϳ20 ng of gDNA were as follows: 35 cycles of 95°C for 20 s, 57°C for 30 s, and 72°C for 2 min 20 s followed by a final extension 72°C for 10 min. TcVP1 site-directed tagging at nucleotide level was confirmed by sequencing several clones of PCR products obtained with primers 17 and 22 (Table 1) cloned into pGEM-T easy vector.
Immunofluorescence Analysis-Epimastigotes were washed with PBS and fixed with 4% paraformaldehyde in PBS for 1 h at room temperature. The cells were allowed to adhere to poly-Llysine-coated coverslips and then permeabilized for 5 min with 0.1% Triton X-100. Permeabilized cells were blocked with PBS containing 3% BSA, 1% fish gelatin, 50 mM NH 4 Cl, and 5% goat serum overnight at 4°C. Then cells were incubated with a primary antibody (monoclonal anti-TbFCaBP, 1:10; polyclonal rabbit anti-TbVP1, 1:250; monoclonal anti-HA tag, 1:500; rat CRISPR/Cas9-mediated Endogenous Tagging in T. cruzi DECEMBER 2, 2016 • VOLUME 291 • NUMBER 49 anti-HA tag, 1:10; monoclonal anti-c-Myc tag, 1:10; rabbit antic-Myc tag, 1:50; rabbit anti-TbBiP, 1:50; rabbit anti-TbVDAC, 1:200) diluted in 1% BSA in PBS (pH 8.0) for 1 h at room temperature. The cells were washed three times with 1% BSA in PBS (pH 8.0) and then incubated for 1 h at room temperature in the dark with Alexa Fluor 488-conjugated goat anti-mouse and Alexa Fluor 546-conjugated goat anti-rabbit or Alexa Fluor 546-conjugated goat anti-mouse and Alexa Fluor 488-conjugated goat anti-rabbit secondary antibodies (1:1,000). Following incubation with the secondary antibody, the cells were washed and mounted on slides. DAPI (5 g/ml) was included in the Fluoromount-G mounting medium to stain DNA. Controls were performed as described above but in the absence of a primary antibody. Differential interference contrast and fluorescence optical images were captured with a 100ϫ objective (1.35 aperture) under nonsaturating conditions with an Olympus IX-71 inverted fluorescence microscope with a Photometrix CoolSnapHQ charge-coupled device camera driven by Del-taVision software (Applied Precision, Issaquah, WA) and deconvolved for 15 cycles using Softwarx deconvolution software ( Fig. 6) or with a confocal microscope Leica TCS SP5 II, with a 100ϫ objective (1.44 aperture) under nonsaturating conditions, that uses photomultiplier tubes for detection of emission, and LAS AF software (Leica, Wetzlar, Germany) for acquisition and processing of digital images (Figs. 2-5 and 7).