Female BALB/c mice aged 6–8 weeks were obtained from the IGM/FIOCRUZ animal facility where they were maintained under pathogen-free conditions. All animal experimentation was conducted in accordance with the Guidelines for Animal Experimentation established by the Brazilian Council on Animal Experimentation (CONCEA). The local institutional review board (CEUA) approved all procedures involving animals (CEUA-015/2019-IGM/FIOCRUZ).

L. braziliensis promastigotes (MHOM/BR/01/BA788) (De Moura et al., 2005) were maintained in Medium 199 (Sigma-Aldrich) supplemented with 20% heat-inactivated fetal calf serum (FCS), HEPES (40 mM), adenine (0.1 mM), Hemin (5 μg/ml), biotin (1 μg/ml) and antibiotics (penicillin 100 IU/ml and streptomycin 100 μg/ml) (all from Thermo Scientific) at 26°C. L. braziliensis transfectants expressing Cas9 and T7 RNA polymerase (LbCas9T7) or centrin^−/− mutants (LbCen^−/−) were maintained in medium supplemented with hygromycin (100 µg/ml), or neomycin (50 µg/ml) and puromycin (10 µg/ml), respectively. Prior to in vitro and in vivo infection assays, parasites were incubated in Schneider`s insect medium (Sigma-Aldrich) supplemented with 10% FCS, 2 mM L-glutamine, penicillin (100 U/ml) and streptomycin (100 μl/ml) (Thermo Scientific). In LbCas9T7, LbCen^−/− and wild-type L. braziliensis (LbWT) promastigotes, in vitro parasite growth was determined by inoculating 5 × 10⁵ parasites/ml in supplemented Schneider’s medium at 26°C for 6 days. Parasite numbers were counted daily using a hemocytometer. LbCas9T7, LbCen^−/− and wild-type L. braziliensis (LbWT) axenic amastigotes were prepared transferring late promastigotes to a modified culture medium (20% FBS; pH 5.5) at 34°C, according to Teixeira et al. (2002). In vitro parasite growth was determined by inoculating 1 × 10⁶ parasites/ml in supplemented Schneider´s medium at 34°C for 5 days. Amastigote aggregates were disrupted by passing through a 25-gauge needle before counting in a hemocytometer.

The sequence of the putative L. braziliensis centrin gene (LbrM.22.1290) (L. braziliensis MHOM/BR/75/M2903) including flanking regions (3’ and 5’ FRs) was retrieved from TriTrypDB (http://tritrypdb.org) and this sequence was chosen because it presented the highest homology with previously characterized L. donovani centrin gene (GenBank—AF406767). The derived amino acid sequences from L. donovani and L. braziliensis centrin were aligned using the ClustalW program in the BioEdit software package (Version 7.0.4.1) and Geneious version 9.0 (https://www.geneious.com). The three-dimensional structure of putative centrin was predicted by the Phyre2 server using all default parameters. This server predicts 3D models using a single submitted protein sequence by gathering homolog sequences from a large non-redundant database, followed by secondary protein structure prediction [Hidden Markov Model (HMM)] (Kelley et al., 2015). The predicted 3D structure was visualized using the UCSF Chimera program (Pettersen et al., 2004). The quality of the predicted model was validated by Ramachandran Plot, generated using the PROCHECK server (Laskowski et al., 1996).

Wild-type L. braziliensis promastigotes (LbWT) in mid-log phase (10⁸ cells/ml) were placed in Tb-BSF electroporation buffer (Schumann Burkard et al., 2011) and mixed with plasmid pTB007 (Beneke et al., 2017) (~5 µg) carrying the humanized Streptococcus pyogenes Cas9 nuclease gene (hSPCas9) and T7 RNA Polymerase (T7 RNAP) in a pre-chilled 2 mm electroporation cuvette (SIGMA). Electroporation was performed using a Bio-Rad Gene Pulser Xcell System followed by subsequent clonal selection in M199 medium supplemented with hygromycin (50 μg/ml) in 96-well cell culture plates. Cultures were monitored for 2–4 weeks and emerging transfectants were further expanded in selective medium at 100 μg/ml hygromycin.

To confirm Cas9 expression, cell lysates were prepared with 1 × 10⁷ mid-log parasites using NuPAGE sample reducing agent and LDS Buffer (Invitrogen) following manufacturer’s protocol. LbWT and LbCas9T7 cell lysates were separated on NuPAGE mini protein gel (4–12%) (Invitrogen) and subsequently transferred onto a nitrocellulose membrane using the iBLOT system (Invitrogen). Membranes were blocked in 10% non-fat milk in PBS-T (Tween 20—0.1%) at 4°C, overnight. Following PBS-T washes, membranes were probed with mouse anti-Cas9 antibody (Clone 7A9, BioLegend); washed and probed with goat anti-mouse IgG-HRP (Clone Poly4053, BioLegend). Bound conjugates were detected by ECL western blotting reagents (Thermo Scientific) visualized under ImageQuant LAS 4000 (GE Healthcare).

Gene deletion was performed as previously described (Beneke et al., 2017; Beneke and Gluenz, 2019). Briefly, commercially synthesized oligonucleotides were used to generate the DNA templates for target-specific single guide (sg) RNA (3`and 5`) containing the T7 promoter, the 20nt sgRNA target sequence and a sequence complementary to the sgRNA scaffold. Primers for the amplification of donor DNA containing the neomycin phosphotransferase II (NEO) and puromycin N-acetyltransferase (PAC) genes with target-specific 30nt homology flanks and G00 primer (sgRNA scaffold) were designed using LeishGEdit (www.leishgedit.net) (Beneke et al., 2017) (Supplemental Table 1) and commercially synthesized. To amplify the sgRNA template DNA, a mixture containing G00 primer (100 μM), high fidelity polymerase (Invitrogen), dNTPs and MgCl₂ was prepared, then added to the diluted sgRNA primers (4 µM) and placed in separate tubes containing 5’ sg RNA and 3’ sg RNA primers (pre-frozen at −80°C for 10 min). PCR was performed on a preheated block under the following cycling conditions: 98°C for 30 s (1 cycle), 98°C for 10 s, 60°C for 30 s, 72°C for 15 s (35 cycle), 72°C for 10 min (1 cycle).

PAC and NEO donor DNA were amplified using the pTPuro or Neo plasmid templates, respectively (Beneke et al., 2017). Briefly, 0.5 μl (30 ng/μl stock) of each template was mixed with 8 μl of diluted forward and reverse primers (10 μM) and frozen at −80°C for at least 10 min. A master mix containing high-fidelity polymerase (Invitrogen), DMSO, dNTPs, MgCl_2, and nuclease-free water was prepared and added to the primer template mix. PCR was performed on a preheated block under the following cycle conditions: 98°C for 5 min (1 cycle), 98°C for 30 s, 65°C for 30 s, 72°C for 2 min (40 cycles), 72°C for 7 min (1 cycle). The amplification products of sgRNA templates and donor DNA were further confirmed by running samples on 2 and 1% agarose gel, respectively.

To generate LbCen^−/−, transfection of Cas9 and T7 RNAP-expressing L. braziliensis (LbCas9T7) was performed as previously described (Beneke and Gluenz, 2019), with minor modifications. A mixture containing sgRNA templates and donor DNA was heated at 94°C for 5 min prior to transfection, mixed with 150 μl mid-log Cas9 T7-expressing L. braziliensis resuspended in transfection buffer (Tb-BSF; 10⁸ cells/ml) and transferred to a prechilled 2 mm cuvette. Transfection was performed using two pulses and the X-001 program on a Nucleofector™ 2b Device (Lonza Biosciences, USA). Transfectants were transferred to 5 ml M199 medium in a 25 cm² flask and incubated for 24 h at 26°C. LbCen^−/− clones were selected from plates containing M199 supplemented with 50 μg/ml of neomycin (PROMEGA) and 10 μg/ml of puromycin (SIGMA).

Following transfection and clonal selection, total genomic DNA from LbWT, LbCas9T7 and LbCen^−/− was isolated using PureLink Genomic DNA Mini Kit, according to manufacturer’s instructions (Invitrogen). Centrin deletion and integration of antibiotic resistance markers was confirmed by PCR using specific primers, designed using Geneious software (Biomatters) (Supplemental Table 1). PCR conditions employed were: (i) Centrin ORF detection: 94°C for 3 min (1 cycle), 94°C for 30 s, 53°C for 30 s, 72°C for 1 min (40 cycle), 72°C for 7 min (1 cycle). (ii) Neomycin and (iii) puromycin resistance cassettes: 94°C for 3 min (1 cycle), 94°C for 30 s, 55°C for 30 s, 72°C for 1 min (40 cycles), 72°C for 7 min (1 cycle). Amplicons were separated on 2% agarose gel and visualized using ImageQuant LAS 4000 (GE Healthcare).

To confirm Cas9-mediated centrin deletion, whole genome sequencing (WGS) was performed in the LbCas9T7 and LbCen^−/− clones using an Illumina Hiseq2000 sequencer, with ~60× coverage, paired-end read libraries with a 100 bp read size and 350 bp insert size. FastQC (Andrews, 2010) and Trimmomatic (Bolger et al., 2014) were used for read quality control and adapter sequence removal, respectively. BWA-mem (Li and Durbin, 2009) was used to map the processed reads to the L. braziliensis MHOM/BR/75/M2904 v45 reference genome, obtained from TriTrypDB (Aslett et al., 2010). Reads with a mapping quality score <30 were discarded using SAMtools (Li et al., 2009). Confirmation of centrin deletion was performed based on read depth estimations and visualized via the Integrative Genomics Viewer (IGV) tool (Robinson et al., 2011) using bam files containing the aligned reads and the L. braziliensis MHOMBR75M2904 v45 General Feature Format (GFF), available at TriTrypDB (Aslett et al., 2010). The raw Illumina sequencing reads were deposited under the Bioproject accession number PRJNA763382.

To further confirm LbrM.22.1290 deletion and estimate the potential for off-target genomic alterations, the read depth of each L. braziliensis gene was estimated and normalized in the LbCas9T7 and LbCen^−/− read libraries by the genome coverage using BedTools (Quinlan and Hall, 2010), and in-house Perl scripts. The absolute LbCas9T7 and LbCen^−/− read depth difference of each gene was estimated using the R program (https://www.R-project.org/) and visualized using ggplot2 (Wickham, 2016). All genes with an absolute difference higher than 0.5 genome coverage, which corresponds to approximately one copy per haploid genome, were reported.

Axenic LbWT, LbCas9T7 and LbCen^−/− amastigotes were generated as described elsewhere (Teixeira et al., 2002). After 96 h in culture, axenic amastigote aggregates were disrupted and the cells then washed in PBS, fixed in 4% paraformaldehyde, washed again in PBS and resuspended in DAPI solution (1 µg/ml, Molecular Probes) for 30 min. Parasites were washed twice in PBS and resuspended in 100 µl of PBS. Parasites were centrifuged, stained and visualized under a TCS SP8 scanning laser confocal system (Leica). Alternatively, axenic amastigotes (48 h of culture) were fixed in Karnovsky fixative (2% glutaraldehyde and 2% paraformaldehyde in 0.1 M sodium cacodylate buffer, pH 7.4), washed in cacodylate buffer and then post-fixed with 1% osmium tetroxide (OsO₄) in 0.1 M sodium cacodylate buffer. Samples were mounted on poly-Lysine-coated coverslips, and then dehydrated in an ascending ethanol series (30, 50, 70, 90, and 100%). Finally, samples were critical-point dried in CO_2, mounted on stubs, coated with gold (20–30 nm) and visualized by scanning electron microscopy (SEM). Axenic amastigotes were also evaluated by Transmission Electron Microscopy (TEM). After 48 h, axenic amastigotes (LbWT, LbCas9T7, and LbCen^−/−) were washed in PBS, fixed and then post-fixed as described above. Samples were dehydrated in an acetone series (30, 50, 70, 90, and 100%) and embedded in Poly/Bed^® 812 resin. Ultrathin sections were obtained and mounted on 300-mesh grids, then stained in 5% uranyl acetate and lead citrate. Cells were analyzed at Jeol JEM 1230 transmission electron microscope and Jeol JSM 6390LV scanning electron microscope.

Primary macrophages obtained from mouse bone marrow were suspended in RPMI 1640 medium (Sigma-Aldrich), supplemented with 10% FBS, 2.5% HEPES, 100 IU/ml penicillin and 100 µg/ml streptomycin (all from Invitrogen) and seeded at 3 × 10⁵ cells/500 μl/coverslip per well on 24-well plates. Monolayers of adherent macrophages formed on the coverslips were washed to remove any non-adherent cells and then exposed to stationary phase LbWT and LbCen^−/− at a 10:1 parasite/host cell ratio (clone #4), i.e., 3 × 10⁶/500 μl in RPMI 1640 + 10% FBS/well. The plates were incubated at 35°C under 5% CO₂ for 24 h, and then coverslips were extensively washed to remove any non-internalized Leishmania. At different time periods after parasite exposure, coverslips were methanol-fixed and stained with hematoxylin and eosin (H&E). Leishmania loading or macrophage infection was assessed by scanning 200 macrophages in each sample for the enumeration of cells with and without Leishmania; the total number of intracellular parasites was counted by optical microscopy.

BALB/c mice (n = 10) were inoculated with 2 × 10⁵ LbWT or LbCen^−/−. Parasites were inoculated into the left ear dermis in 10 μl PBS using a 27G needle. Parasite loads were determined at different time points by limiting dilution analysis in samples obtained from mouse ears, draining lymph nodes and spleen as previously described (Novais et al., 2013). Lesion development was monitored weekly by measuring the thickness of the ear using a digital caliper (Thermo).

Comparisons between two groups were performed using Mann–Whitney (non-parametric t-test), and among three of more groups using the Kruskal–Wallis test. Analyses were conducted using Prism software (V.8.0, GraphPad) and p-values ≤0.05 were considered significant. Data are presented as mean± standard deviation.

A BLASTP search in the TriTrypDB database revealed the presence of a putative centrin gene in L. braziliensis (LbrM.22.1290) showing homology to previously characterized L. donovani centrin [GenBank—AF406767; (Selvapandiyan et al., 2001)]. The putative L. braziliensis centrin is 450 nucleotides (nt) long (including stop codon; orf −447 bp) and encodes a protein containing 149 amino acids (AAs) (Supplemental Figure 1A). Pairwise amino acid alignment between LbrM.22.1290 and L. donovani centrin indicated high identity (94.63%) and high degree of conservation among the calcium binding sites (EF-hand 1 and 4; alpha helical stretch) (Supplemental Figure 1B).

The amino acid sequence of LbrM.22.1290 was used to predict the in silico three-dimensional structure using the solved structure of Trypanosoma brucei centrin 4 (Tb927.7.3410). Again, the amino acid sequences in LbrM.22.1290 and Tb927.7.3410 showed high sequence homology (Figure 1A). Molecular modeling using 100% of LbrM.22.1290 (149 AAs) allowed prediction of the protein structure with 99.9% confidence (Figure 1B). The stereochemical quality of the modeled structure was validated by Ramachandran plot analysis, revealing that 93.1 and 6.9% of the residues were in the most favored and additionally allowed regions, with no residues in disallowed regions (Figure 1C). These in silico analyses suggest that the putative centrin gene in L. braziliensis is a structural homolog of T. brucei centrin 4 and centrin 4 is highly conserved across Leishmania and Trypanosoma spp.

Our strategy consisted in first generating an L. braziliensis line to co-express Cas9 and T7 RNAP (LbCas9T7). Following transfection of mid-log LbWT promastigotes with pTB007, carrying genes for Cas9 and T7 RNAP, transfectants were selected under antibiotic pressure (HYG) and resulting clones were grown in selective medium. Western blot analysis confirmed successful expression of the Cas9 protein (162 kDa) (Figure 2A). One resulting clone (LbCas9T7) was selected for deletion of centrin ORF. Next, L. braziliensis centrin alleles were targeted with PAC and NEO donor DNA cassettes, each flanked by 30 nt homologous sequences. Primers were designed for the identification of centrin^−/− L. braziliensis (LbCen^−/−) line by PCR (Figure 2B and Supplemental Table 1). LbCas9T7 parasites were first transfected with the two donor DNA constructs (PAC and NEO) plus two sgRNAs (DNA) templates targeting the 5’ and 3’ FRs of the centrin ORF. Following transfection and serial limiting dilution, clonal selection under antibiotic pressure (puromycin and neomycin) yielded six clones, which were further expanded. Centrin deletion was confirmed by PCR as shown by the absence of the centrin amplicon (344 bp orf fragment) paralleled by presence of amplicons specific for PAC (788 bp) and NEO (886 bp), indicating the successful integration of both donor cassettes (Figure 2C). Additionally, PCR products spanning the integration sites were Sanger-sequenced, confirming integration of both donor cassettes (Supplemental Figure 2). These data confirm the successful use of the LeishGEdit for deletion of the putative centrin gene (LbrM.22.1290) in L. braziliensis.

Next, we performed whole genome sequencing of LbCas9T7 and LbCen^−/− to verify the targeted deletion of L. braziliensis centrin in LbCen^−/−. Initially, the read depth across the genome of LbrM.22.1290 in both LbCas9T7 and LbCen^−/− was estimated and the only relevant alteration was the expected absence of coverage in the centrin gene region in LbCen^−/− (Figure 3). These data show targeted centrin deletion and do not indicate translocation to another genomic locus. Also, allele frequency analysis of heterozygous SNPs revealed that, different from the L. braziliensis M2904 reference genome which essentially triploid (Rogers et al., 2011), the L. braziliensis BA788 genome is overall diploid, with chromosome 22 that harbors centrin gene having two copies (data not shown). Therefore, the successful generation of centrin knockout was achieved with two resistance markers. To evaluate the potential occurrence of Cas9-induced deletions or duplications in other genomic regions that could have an impact on phenotype, the read depth of all L. braziliensis MHOMBR75M2904 genes normalized by the genome coverage was estimated and compared, for both LbCas9T7 and LbCen^−/− (Figure 4). Most differences in read depth ranged from “0” to “0.3”, suggesting minimal fluctuations in the read libraries that are unlikely to result in a mutant-specific phenotype (Figures 4A, B). Read depth differences larger than 0.5 were only observed in 12 genes, all of which were multicopy genes. These 12 occurrences corresponded to five multicopy genes with a basal coverage higher than “3”, such as amastins, zinc-transporters and metallopeptidases. In addition, five were multi-copy structural RNAs, e.g., snoRNA and snRNA. One was a 40S ribosomal protein (LbrM.30.0750), and the other the expected centrin gene, LbrM.22.1290 (Table 1). Taken together, these results do not indicate the presence of off-target deletions or duplications in the genome of the newly generated LbCen^−/−.

Next, we evaluated the phenotype of LbWT, LbCas9T7, and LbCen^−/−. Comparative promastigote growth curves showed similar proliferation rates over a 7-day time course (Figure 5A). Promastigotes (96 h cultures) were stained with DAPI and examined by confocal microscopy. LbWT, LbCas9T7, and LbCen^−/− promastigotes did not exhibit any morphological changes differences and all three lines presented mostly single nuclei (Figure 5B). LbCen^−/− axenic amastigotes, however, exhibited a significantly deficient growth, starting at 24 h and persisting over a 5-day time course, when compared to LbWT and LbCas9T7 axenic amastigotes (Figure 5C). As expected, LbCen^−/− amastigotes had multiple nuclei and kinetoplasts (Figure 5D). CRISPR-mediated deletion of centrin in L. braziliensis yielded a cell line that displayed slower growth at the amastigote stage, with similar morphological changes as those described in L. donovani centrin^−/− mutants (Selvapandiyan et al., 2004).

LbCen^−/− axenic amastigotes (48 h culture) were subjected to morphological analyses by SEM and TEM. SEM results showed the presence of abnormally large cells compared to the normal morphology observed in LbWT and LbCas9T7 (Figure 6A). LbCen^−/− axenic amastigotes (48 h culture) observed under TEM showed prominently large cells with multiple nuclei and/or kinetoplasts, suggesting a defective cytokinesis whereas both LbWT and LbCas9T7 axenic amastigotes did not (Figure 6B). Thus, LbCen^−/− exhibited similar morphology to that described for centrin^−/− L. donovani (Selvapandiyan et al., 2001; Selvapandiyan et al., 2004).

Given the growth arrest (Figure 5C) and the prominent structural alterations (Figure 6) observed in axenic amastigotes, we predicted that LbCen^−/− would also exhibit reduced infectivity in the host cell. A time course analysis of macrophage infection revealed a significant reduction in the percentage of cells infected with LbCen^−/− compared to cells infected with LbWT, as early as 72 h after parasite exposure (Figure 7A). This result was paralleled by a significantly reduced number of amastigotes detected within macrophages at the same time point (Figure 7B). H&E staining of infected cells also confirmed the presence of multiple nuclei in cells infected with LbCen^−/− compared to LbWT (Figure 7C). Therefore, CRISPR-Cas9-mediated deletion of centrin impairs the in vitro growth of L. braziliensis.

Upon inoculation of LbCen-/- promastigotes in the ear dermis of BALB/c mice, parasites were detected after four days, at both the inoculation site and in dLNs (Figures 8A and B, respectively). However, LbCen-/- failed to induce lesion development, differently from mice inoculated with LbWT, in which lesions peaked at 4 weeks, then gradually subsided by 10 weeks (Figures 8C). Also, in mice infected with LbCen-/- promastigotes, parasites were not detected at the inoculation site, dLNs or in the spleen, 6 or 12 weeks post-inoculation (Figures 8D and E, respectively), indicating that LbCen-/- the impaired in vitro growth capacity (Figure 7) is recapitulated in vivo.