Interplay between a cytosolic and a cell surface carbonic anhydrase in pH homeostasis and acid tolerance of Leishmania

Leishmania is a digenetic protozoan that alternates between sandfly vector and mammalian hosts. There are more than twenty species of Leishmania, causing a wide spectrum of clinical symptoms, collectively known as leishmaniasis. With over one million new cases reported per year, this neglected tropical disease continues to be a menace to impoverished communities across the globe (Okwor and Uzonna, 2016). The flagellated promastigote form of the parasite replicates in sandfly midgut and is transmitted to humans when a fly takes blood meal. Once inside the host, Leishmania promastigotes are rapidly phagocytosed by macrophages, either directly or by engulfing the parasite-harbouring apoptotic neutrophils (Peters et al., 2008). Intracellular parasites are then transformed into non-flagellated amastigotes, which continue to multiply within the phagolysosomes until the cell bursts, causing spread of the infection. Adaptive mechanisms that allow Leishmania to survive and replicate within an acidic phagolysosomal environment still remain poorly understood.

In 1988, Glaser et al. reported that Leishmania donovani amastigotes as well as promastigotes have an amazing ability to maintain their intracellular pH close to neutral, even when the extracellular pH is ∼5.0 (Glaser et al., 1988). Proton extrusion from the cytosol, mediated by a membrane-bound H⁺-ATPase, has been implicated in regulating the intracellular pH of the parasite (Marchesini and Docampo, 2002; Jiang et al., 1994). However, the role of this ATPase is yet to be established in Leishmania cells growing within host macrophages or under in vitro conditions when the extracellular pH is close to the lysosomal pH (∼5.0). Studies using transport inhibitors have also suggested an involvement of a ‘yet to be identified’ Cl⁻ channel in cytosolic buffering of the parasite (Vieira et al., 1995).

Important insight towards understanding the acid-survival strategy of Leishmania came from our recent study in which we showed pharmacological inhibition of carbonic anhydrase activity in Leishmania major causes intracellular acidosis, leading to parasite cell death by apoptosis and necrosis. The parasites growing in acidic medium were much more vulnerable to carbonic anhydrase inhibitors as opposed to those grown at neutral pH, suggesting carbonic anhydrase activity might be vital for Leishmania for successful colonization of the acidic phagolysosome (Pal et al., 2015). Since two carbonic anhydrases are constitutively expressed in Leishmania major (LmCA1 and LmCA2), genetic dissection of their function is crucial for understanding their role in parasite physiology (Pal et al., 2015).

Carbonic anhydrases are a group of ubiquitous metalloenzymes that catalyze reversible hydration of carbon dioxide to bicarbonate: CO₂+H₂O↔H⁺+HCO₃⁻. Based on sequence and structural features, they are classified into five distinct families, of which α-, β- and γ-carbonic anhydrases are most common (Sly and Hu, 1995; Smith and Ferry, 2000; Rowlett, 2010; Ferry, 2010). The function of α-carbonic anhydrases is extensively studied in mammals, where they play a central role in respiration, ion transport and in multiple metabolic pathways, such as gluconeogenesis, lipogenesis and ureagenesis (Sly and Hu, 1995). Comparatively, much less is known about the function of β-carbonic anhydrases that are prevalent in plants and bacteria, or of γ-carbonic anhydrases, found mainly in archaea (Rowlett, 2010; Ferry, 2010). Owing to their functional indispensability, carbonic anhydrases from several pathogenic microorganisms have been validated as potential drug targets in recent years (Supuran, 2011; Joseph et al., 2011; Monti et al., 2012; Maresca et al., 2013; Nishimori et al., 2014).

Interestingly, of the two carbonic anhydrases in L. major, LmCA1 is a β-carbonic anhydrase whereas LmCA2 belongs to the α-family. Both of them are highly conserved across Leishmania species, suggesting their functional importance (Logan-Klumpler et al., 2012; Ivens et al., 2005). The homologue of LmCA1 in Leishmania donovani chagasi has already been cloned and partially characterized (Syrjänen et al., 2013). However, the precise role of carbonic anhydrases in Leishmania remains completely unknown. In this work, we combined immunolocalization studies with targeted gene disruption and functional complementation experiments to unequivocally understand how LmCA1 and LmCA2 contribute towards parasite physiology. Our results reveal that the functional interplay between cytosolic LmCA1 and membrane-bound LmCA2 in maintaining intracellular pH is a determining factor for in vitro growth and proliferation of L. major under acidic conditions. The parasites that were deficient for LmCA1 and/or LmCA2 were also found to be susceptible to acidic phagolysosomal pH within host macrophages. Our work unravels a new pH regulatory mechanism in Leishmania that is governed by the concerted actions of LmCA1 and LmCA2, thereby establishing these two carbonic anhydrases as therapeutically exploitable molecular targets.

As per the annotated Leishmania genome database, L. major is predicted to encode two carbonic anhydrases, LmjF.06.010 and LmjF.28.0480 (Logan-Klumpler et al., 2012; Ivens et al., 2005). They have been referred to as LmCA1 and LmCA2, respectively, and both these genes were shown to be constitutively expressed in L. major (Pal et al., 2015). These two carbonic anhydrases are also conserved in the genomes of other species of Leishmania (Logan-Klumpler et al., 2012; Ivens et al., 2005). In fact, the homologue of LmCA1 in Leishmania donovani chagasi has recently been cloned and partially characterized, and has been established as a new member of the β-carbonic anhydrase family (Syrjänen et al., 2013). The most striking feature that distinguishes the β-carbonic anhydrases from other family members of carbonic anhydrases is the Cys₂-His-X-Zn²⁺ coordination sphere (where X is an exchangeable ligand, such as water or acetate) (Rowlett, 2010). Clustal Omega alignment of the LmCA1 primary sequence with sequences of other well-known β-carbonic anhydrases revealed that all these Zn²⁺-coordinating amino acids are absolutely conserved in LmCA1, namely Cys110, His166 and Cys169 (Sievers et al., 2011). Apart from these, other important active site residues of β-carbonic anhydrases, as indicated in Fig. 1A, have also been found to be conserved or replaced with similar amino acids in LmCA1 (Rowlett, 2010). LmCA2, on the contrary, resembles the α-family of carbonic anhydrases, the hallmark of which is an active site Zn atom tetrahedrally coordinated with three His residues and a water molecule (Sly and Hu, 1995). All three Zn-binding His residues are conserved in LmCA2, namely His155, His157 and His174. The gatekeeper Thr residue, which accepts a hydrogen bond from the Zn-bound water, is also conserved in LmCA2 (Thr253). This hydrogen bond is crucial for correctly orienting the Zn-bound water for nucleophilic attack on the CO₂ (Sly and Hu, 1995). Also, most of the amino acid residues that form the hydrophobic CO₂-binding pocket of α-carbonic anhydrases are either conserved or replaced with similar amino acids in LmCA2 (Fig. 1B) (Sly and Hu, 1995). As per our bioinformatic analysis, LmCA1 is devoid of any signal sequence and hence is predicted to be localized in the cytosol. Interestingly, LmCA2 has a predicted secretory signal peptide and a transmembrane domain, and thus appears to be a membrane protein. To experimentally validate their subcellular localization, we generated two L. major strains stably expressing either LmCA1 or LmCA2 as C-terminal GFP-tagged proteins. As illustrated in Fig. 1C, LmCA1 and LmCA2 colocalized, respectively, with cytosolic and plasma membrane marker proteins – adenosine kinase and Na⁺/K⁺-ATPase (Sen et al., 2007; Bose et al., 2012). This observation is consistent with the results of a cell fractionation experiment. Upon immunoblotting of the cell fractions with an antibody against GFP, LmCA1–GFP was almost exclusively detected in the cytosolic fraction, whereas the majority of LmCA2–GFP was found in the membrane fraction. The authenticity of the cell fractionation procedure was checked by immunoblotting with antibodies against cytosolic adenosine kinase and plasma membrane protein Na⁺/K⁺-ATPase (Fig. 1D). Our results thus confirmed that LmCA1 is localized in the cytosol and LmCA2 is a plasma-membrane-bound protein.

We next attempted to create LmCA1^−/− and LmCA2^−/− null mutants of L. major in order to study function of the respective genes in Leishmania physiology. Both LmCA1 and LmCA2 are single-copy genes located in chromosomes 6 and 28, respectively, of the L. major genome (Logan-Klumpler et al., 2012; Ivens et al., 2005). Since Leishmania is an asexual diploid organism, we planned to knockout both the alleles of LmCA1 and LmCA2 using homologous recombination, sequentially replacing them with neomycin phosphotransferase (NEO) and hygromycin phosphotransferase (HYG) gene cassettes (Fig. 2A) (Zhang et al., 2004). As described in Materials and Methods, we initially generated LmCA1^+/− and LmCA2^+/− heterozygous strains and selected them under appropriate antibiotic pressure. The authenticity of the heterozygous strains was verified by genomic DNA PCR using primers flanking the LmCA1 or LmCA2 locus (Fig. 2A). Two bands, of sizes 1.58 kb (corresponding to the LmCA1 allele) and 3.82 kb (corresponding to the NEO cassette), were amplified from LmCA1^+/−. Similarly, two bands, of sizes 2.35 kb (corresponding to the LmCA2 allele) and 3.83 kb (corresponding to the HYG cassette) were amplified from LmCA2^+/−. In wild-type L. major, as expected, only one band (corresponding to either LmCA1 or LmCA2 allele, depending on the primers used) was amplified in the respective PCR reactions (Fig. 2B). The correct integration of NEO or HYG cassettes into the heterozygous strains was further verified by performing PCR with a NEO or HYG internal primer as appropriate, and an external primer recognizing the 5′ or 3′ region of the LmCA1 or LmCA2 locus as appropriate (Fig. S1A–C).

The LmCA1^+/− and LmCA2^+/− strains were then used for a second round of electroporation to create the respective null mutants. The double-targeted LmCA1 mutant strain was selected under G418 and hygromycin B. Genomic PCR analysis of the double antibiotic-resistant strain revealed that one allele of LmCA1 was still intact in the genome, whereas both the NEO and HYG cassettes also got integrated in the same locus (Fig. S1D–G). Repeated attempts to generate an LmCA1^−/− null mutant failed, suggesting a change in chromosomal ploidy in the double-targeted LmCA1 mutant strain. Such a phenomenon has been reported previously when knockout of essential genes in Leishmania has been attempted (Cruz et al., 1993; Dumas et al., 1997; Mittra et al., 2016). Knockout of both the alleles of LmCA2 was lethal as the double-targeted LmCA2 mutant strain failed to grow in G418- and hygromycin-B-containing medium. Failure to generate LmCA1^−/− or LmCA2^−/− null mutants suggested that these two genes are essential.

However, we were able to use the LmCA1^+/− strain to generate an LmCA1^+/−:LmCA2^+/− double heterozygous strain in which one allele each of LmCA1 and LmCA2 were replaced with NEO and HYG cassettes, respectively. The double heterozygous strain was verified by genomic PCR using primers flanking the LmCA1 or LmCA2 locus. The PCR results confirmed the presence of the NEO cassette in the LmCA1 locus and HYG cassette in the LmCA2 locus, along with one copy of LmCA1 and LmCA2 alleles at their respective loci (Fig. 2A,C).

Semi-quantitative reverse-transcriptase (RT)-PCR revealed that LmCA1 transcript was reduced by ∼45% in the LmCA1^+/− and LmCA1^+/−:LmCA2^+/− mutant strains compared to that in wild-type L. major. Reduction in the amount of LmCA2 transcript in the LmCA2^+/− and LmCA1^+/−:LmCA2^+/− mutant strains was ∼40% compared to the wild-type strain (Fig. 2D,E). Total carbonic anhydrase activity in LmCA1^+/−, LmCA2^+/− and LmCA1^+/−:LmCA2^+/− mutant strains was decreased by ∼35%, ∼30% and ∼55%, respectively, as compared to their wild-type counterpart (Fig. 2F).

At normal pH, the growth rates of LmCA1^+/−, LmCA2^+/− and LmCA1^+/−:LmCA2^+/− mutant L. major strains were identical to that of the wild-type counterpart (Fig. 3A). However, when the cells were subjected to low pH stress (pH of medium maintained at 5.5) to mimic phagolysosomal conditions, all the heterozygous mutant strains grew much more slowly as compared to the wild-type L. major (Fig. 3B). After 72 h of growth at this low pH, a 48% and 24% reduction in the total number of cells was observed, respectively, for the LmCA1^+/− and LmCA2^+/− mutants in comparison to the wild type. For the double heterozygous strain, LmCA1^+/−:LmCA2^+/−, the growth defect was even more pronounced, resulting in a 56% reduction in the total number of cells (Fig. 3C). Complementation of the heterozygous mutants with episomal LmCA1 or LmCA2 constructs restored expression of the corresponding mRNAs and carbonic anhydrase activity to varying extents, leading to significant reversal of their growth defects (Fig. S2; Fig. 3C). Exogenous supplementation of the carbonic anhydrase reaction product, bicarbonate, also resulted in a similar growth reversal in the LmCA1^+/−, LmCA2^+/− and LmCA1^+/−:LmCA2^+/− mutants (Fig. 3C). Our results indicate that the functional interplay of cytosolic LmCA1 and membrane-bound LmCA2 is crucial for in vitro propagation of Leishmania parasites at low pH. Furthermore, under acidic growth conditions, all mutant parasites exhibited rounded morphology with more than a 50% reduction in cell length as compared to their wild-type counterpart. The cell length of the corresponding complementation strains showed signs of significant recovery, as expected (Fig. 3D,E). Although such crippled morphology is an indicator of metabolic stress, the majority of the mutant parasites were viable, as confirmed by MTT assay (Fig. S3). These data suggest that the observed growth defects in the mutant parasites are not due to cell death.

We next investigated whether haploinsufficiency of LmCA1 or LmCA2 affected cytosolic buffering capacities of the mutant parasites. For this, we measured the intracellular pH of the wild-type and mutant parasites growing in acidic (pH 5.5) medium (Table 1). Consistent with an earlier report, we too observed that the wild-type parasites could maintain near-neutral intracellular pH (6.82), even when the external pH was acidic (Glaser et al., 1988). However, the intracellular pH of the LmCA1^+/−, LmCA2^+/− and LmCA1^+/−:LmCA2^+/− strains was reduced to 6.62, 6.71 and 6.54, respectively, indicating that their cytosolic buffering capacities were compromised. This drop in intracellular pH was restored to different extents upon genetic complementation using appropriate episomal construct(s) or through exogenous supplementation of bicarbonate. Our results confirmed that LmCA1 and LmCA2 play an important role in maintaining cytosolic pH in Leishmania cells, which might be crucial for its growth and/or survival in the low pH environment of the phagolysosome.

We next performed fluorescence-activated cell sorting (FACS) analysis to test whether intracellular acidosis results in a slowing of the cell cycle in the mutant parasites. Following serum starvation and hydroxyurea-mediated synchronization, the wild-type and the mutant parasites were incubated in pH 5.5 medium to allow them to progress through their cell cycles. At 0-h post synchronization, most of the wild-type and mutant parasites were in the G₀/G₁ phase of the cell cycle, indicating the efficiency of the synchronization process (Fig. 4A,B). Thereafter, the wild-type cells, as expected, progressed normally through the cell cycle with the majority of the cells clearing the G₀/G₁ phase, and reaching either the S phase (∼ 25%) or G₂/M phase (∼50%) at 8 h. The LmCA1^+/−, LmCA2^+/− and LmCA1^+/−:LmCA2^+/− mutant L. major cells on the other hand showed abnormal distribution in the cell cycle and were predominantly stuck in the G₀/G₁ phase (>50%), even after 8 h. Genetic complementation using appropriate episomal construct(s) restored normality in their cell cycle distributions to a large extent (Fig. 4A,B). The above results indicate that slow cell cycle progression resulting from intracellular acidosis is responsible for the growth defect of LmCA1^+/−, LmCA2^+/− and LmCA1^+/−:LmCA2^+/− mutant parasites in a low pH environment.

Having confirmed the role of LmCA1 and LmCA2 for in vitro propagation of L. major in acidic medium, we next investigated whether single allele deletion of LmCA1 and/or LmCA2 had any effect on the parasites residing within the host phagolysosomes. For this, we infected J774A.1 macrophages with either wild-type or mutant L. major strains, and analyzed the intracellular parasite burden for each set of infection. As shown in Fig. 5A, infection with LmCA1^+/−, LmCA2^+/− or LmCA1^+/−:LmCA2^+/− L. major strains exhibited 45%, 36% and 47% lowered intracellular parasite burden, respectively, in comparison to the infection with the wild-type parasite. This trend was also reflected in the percentage of infected macrophages with the wild-type or mutant parasites (Table S1). The intracellular parasite burdens as well as the percentages of infected macrophages for the corresponding complementation strains were significantly higher (almost similar to the wild-type strain), thus corroborating the role of LmCA1 and LmCA2 for intracellular parasite growth and/or survival (Fig. 5A; Table S1).

To test whether the acidic pH of lysosomes is actually responsible for this lower number of mutant parasites within the host cells, we treated the macrophages with chloroquine to raise intra-phagolysosomal pH (Poole and Ohkuma, 1981; Levitz et al., 1999). Diminished fluorescence of the pH-sensitive probe, LysoTracker Red, confirmed the lysosomes were more alkaline upon chloroquine treatment (Fig. 5B). The chloroquine treatment did not interfere with the replication of macrophages or Leishmania parasites, as confirmed by live cell counting (Fig. S4). When these alkalized macrophages were infected with the wild-type or mutant parasites, intracellular parasite burdens as well as the percentage of infected macrophages were almost the same for each set of infection, whether with the wild-type or mutant strains (Fig. 5C; Table S1). These data confirmed that both LmCA1 and LmCA2 are vital for the Leishmania parasite to counteract phagolysosomal acid stress and proliferate within host macrophages.

Unlike most macrophage-dwelling pathogens, which either escape from early phagosomes to the cytosol or block phagosome–lysosome fusion, Leishmania have adopted the adverse phagolysosomal environment as their replication niche (Thi et al., 2012; Ray et al., 2009; Ernst et al., 1999). How they thrive in such acidic condition remains an intriguing question that has an important bearing in the understanding of pathogenesis of leishmaniasis. We report here that the parasite relies on the combined action of two carbonic anhydrases, one located in its cytosol (LmCA1) and the other a transmembrane protein (LmCA2), for proliferation in acidic environments, both in vitro as well as within the phagolysosomes of host macrophages. Our study points towards a new pH homeostatic mechanism in Leishmania parasite.

The functional importance of LmCA1 and LmCA2 is suggested by the fact that, even with only a ∼50% reduction in their expression levels, the corresponding heterozygous mutant strains exhibited significant growth retardation and morphological abnormalities at low pH. Interestingly, the growth retardation in the LmCA1^+/− strain was more pronounced compared to the LmCA2^+/− strain. This gave rise to the notion that, of the two enzymes, LmCA1 might play a more dominant role in promoting Leishmania cell growth under acidic conditions. Although the maximum growth defect was observed in the LmCA1^+/−:LmCA2^+/ strain, the double heterozygosity of this mutant did not have an additive effect. This suggests that LmCA1 and LmCA2 work in tandem in maintaining optimum conditions for parasite growth in a low pH environment. The intracellular pH in LmCA1^+/−, LmCA2^+/− and LmCA1^+/−:LmCA2^+/− mutant strains was lowered by about 0.2, 0.1 and 0.3 pH units, respectively, as compared to the wild type. This finding is consistent with the role of carbonic anhydrases in cytosolic buffering that are seen in various cell types across species (Thomas, 1976; Geers and Gros, 1990; Leniger et al., 2004; Marcus et al., 2005; Schewe et al., 2008). In fact, a similar drop in intracellular pH has been observed by us previously when carbonic anhydrase activity in L. major was pharmacologically inhibited with metal dithiocarbamates (Pal et al., 2015). It is worth noting that although treatment with metal dithiocarbamates led to Leishmania cell death, the slow-growing heterozygous mutants were all viable at acidic pH (Pal et al., 2015). Thus, besides their role in intracellular pH maintenance, LmCA1 and/or LmCA2 appear to perform other crucial functions in Leishmania physiology, and total inhibition of carbonic anhydrase activity might be detrimental for parasite survival. Repeated failure to generate LmCA1^−/− or LmCA2^−/− knockout strains further substantiated this idea. Even though the heterozygous mutants did not undergo cell death in acidic medium, they experienced significant delay at the G₀/G₁ phase of the cell cycle, which eventually resulted in their growth retardation. This cell cycle delay could be a direct consequence of the lowering of intracellular pH, as has been reported previously (Aerts et al., 1985; Karagiannis and Young, 2001; Sellier et al., 2006). The LmCA1^+/−, LmCA2^+/− and LmCA1^+/−:LmCA2^+/− mutant parasites were also susceptible to the acidic phagolysosomal pH of the host macrophages, a phenotype that could be reversed by lysosomal alkalization with chloroquine. Thus, by combining gene disruption with episomal complementation, we propose the physiological role of LmCA1 and LmCA2 in acid acclimatization of Leishmania parasites both in vitro and within host macrophages. It is worth noting that, like Leishmania, Trypanosoma cruzi (another member of the trypanosomatid family) also encodes two carbonic anhydrases in its genome (Logan-Klumpler et al., 2012; Pan et al., 2013). Since T. cruzi also resides transiently within the host cell phagolysosomes, it will be interesting to see if such a dual carbonic anhydrase-mediated acid acclimatization mechanism is operational in this parasite (Andrade and Andrews, 2004).

Although Leishmania, by virtue of its surface lipophosphoglycans, is protected from the hydrolytic action of lysosomal enzymes, the acidic environment within the phagolysosomes poses a serious threat for their intracellular growth and/or survival (Chang and McGwire, 2002; Spath et al., 2003). Life sustenance in such a condition becomes much more challenging since Leishmania utilizes the proton motive force for H⁺-driven glucose and amino acid uptake (Zilberstein and Dwyer, 1985; Bonay and Cohén, 1983; Glaser and Mukkada, 1992). Therefore, the parasite must possess robust acid extrusion machinery to resist a drop in intracellular pH resulting from rapid influx of protons and also from generation of metabolic acids. A dicyclohexylcarbodiimide-sensitive plasma membrane H⁺-ATPase has been implicated in regulating the intracellular pH of Leishmania (Marchesini and Docampo, 2002). Jiang et al. have reported that ATPase inhibitor-sensitive proton extrusion in Leishmania is coupled with K⁺ movement and, hence, they predicted that the H⁺-ATPase might in fact be a K⁺/H⁺-ATPase, acting as an antiporter (Jiang et al., 1994). However, direct involvement of this proton-translocating ATPase in counteracting acid stress in Leishmania remains to be established. Also, sole dependence on the ATPase-mediated acid extrusion against such a steep proton gradient (∼ 2 pH units) would be energetically costly for the intracellular parasite. The work presented here reveals that Leishmania is endowed with an alternative mechanism for intracellular pH regulation that is mediated by the coupled action of LmCA1 and LmCA2. Based on our subcellular localization data, we propose that the cytosolic LmCA1 catalyzes bicarbonate buffering of intracellular acids by converting protons into water and carbon dioxide, which can easily diffuse out of the cell. The pericellular carbon dioxide is then hydrated to bicarbonate ions and protons by the membrane-bound LmCA2. This H⁺-ATPase-independent pH regulatory circuit is completed once the bicarbonate ions are transported back to Leishmania cells, thereby replenishing the cytosolic bicarbonate pool. It is likely that LmCA2 is spatially close to a bicarbonate transporter, such that locally produced bicarbonate ions can be directly delivered to the cell (Fig. 6). Although the protein involved in bicarbonate transport in Leishmania has not yet been identified, its existence is almost certain since exogenously added bicarbonate ions have been shown to regulate the intracellular pH of the parasite (Vieira and Cabantchik, 1995). This was further validated by our experiment in which supplementation with bicarbonate rescued the heterozygous mutant parasites from growth defects and intracellular acidosis in low pH medium. Physical and functional interaction of the Na⁺/HCO₃⁻ cotransporter or Cl⁻/HCO₃⁻ anion exchanger with transmembrane carbonic anhydrase 9 has been shown to control cytosolic pH in cancer cells and cardiomyocytes (Ditte et al., 2011; Svastova et al., 2012; Swietach et al., 2007; Orlowski et al., 2012). It will be fascinating to verify if such a carbonic anhydrase–bicarbonate transport metabolon is operational in Leishmania parasites. Molecular characterization of the bicarbonate transport system of Leishmania will therefore provide a comprehensive understanding of its acid acclimatization strategy inside the host. Nevertheless, conclusive evidence has been provided here to establish LmCA1 and LmCA2 as important players in the pH regulatory network of the Leishmania parasite and also as potential drug targets. Since the anti-leishmanial properties of many carbonic anhydrase inhibitors have recently been demonstrated (Pal et al., 2015; Syrjänen et al., 2013), LmCA1- and LmCA2-targeted therapies for leishmaniasis appears to be on the horizon.

Unless otherwise mentioned, all reagents were purchased from Sigma-Aldrich. Primers were procured from Integrated DNA Technologies and their sequence details are provided in Table S2. Leishmania expression and gene targeting vectors (pXG-GFP+, pXG-NEO, pXG-HYG, pXG-SAT and pXG-PHLEO) were generous gifts from Dr Stephen M. Beverley (Washington University Medical School, St. Louis).

Wild-type L. major promastigotes (strain: 5ASKH; kindly provided by Dr Subrata Adak, IICB, Kolkata) and J774A.1 murine macrophage cell line (obtained from National Centre for Cell Science, Pune) were cultured as described previously (Pal et al., 2015).

Transfection of DNA into L. major was performed using electroporation as described previously (Zhang et al., 2004). Briefly, 10–30 µg of DNA construct was resuspended in electroporation buffer (21 mM HEPES, 0.7 mM NaH₂PO₄, 137 mM NaCl and 6 mM glucose; pH 7.4) along with 3.6×10⁷ L. major promastigotes. The suspension was incubated in a 0.2 cm electroporation cuvette for 10 min on ice, following which electroporation was performed on the Bio-Rad Gene Pulsar at 450 V, 550 µF capacitance. Transfected cells were selected in appropriate antibiotic-containing medium.

Forward and reverse primer sets, P1/P2 or P3/P4, were used to PCR-amplify the LmCA1 or LmCA2 open reading frames, respectively, from intronless L. major genomic DNA (Myler et al., 1999). Amplified LmCA1 and LmCA2 coding regions were each cloned into the SmaI and EcoRV sites of the pXG-GFP+ vector to generate the respective C-terminal GFP-tagged constructs. The constructs were verified by sequencing and 30 µg of LmCA1–GFP or LmCA2–GFP construct was transfected into wild-type L. major promastigotes. Transfected L. major cells were selected and maintained in medium containing 100 µg/ml G418 sulphate (Amresco, Dallas).

To sequentially replace both the alleles of LmCA1 or LmCA2 with NEO and HYG gene cassettes, we generated the respective deletion constructs in pXG-NEO and pXG-HYG vectors, as described previously (Zhang et al., 2004). Briefly, 959 bp from the 5′ flanking region (starting 60 bp upstream of the start codon) and 1014 bp from the 3′ flanking region of the LmCA1 open reading frame were PCR-amplified from L. major genomic DNA using primer sets P5/P6 and P7/P8, respectively. Similarly, 517 bp from the 5′ flanking region and 953 bp from the 3′ flanking region of the LmCA2 open reading frame were PCR-amplified using primer sets P9/P10 and P11/P12, respectively. HindIII- and SalI-digested 5′ flanking region, and SmaI- and BamHI-digested 3′ flanking region of LmCA1 or LmCA2 were ligated on either side of the NEO- or HYG-encoding open reading frame in pXG-NEO or pXG-HYG plasmids, respectively. The constructs were verified by sequencing, following which they were digested with HindIII and BamHI to generate the linearized NEO and HYG LmCA1- or LmCA2-targeting cassettes: ^5′LmCA1-NEO-LmCA1^3′, ^5′LmCA1-HYG-LmCA1^3′, ^5′LmCA2-NEO-LmCA2^3′ and ^5′LmCA2-HYG-LmCA2^3′. To generate LmCA1^+/− or LmCA2^+/− strains, wild-type L. major promastigotes were separately transfected with 10 µg of ^5′LmCA1-NEO-LmCA1^3′ or ^5′LmCA2-HYG-LmCA2^3′ linearized cassette. LmCA1^+/− or LmCA2^+/− heterozygous strains were selected and maintained in the presence of 100 µg/ml G418 sulphate or hygromycin B, respectively. Heterozygous strains were plated on semi-solid M199 plates (supplemented with appropriate antibiotics) and small white colonies appeared within 2–3 weeks. The heterozygous genotype of individual colonies was confirmed by genomic DNA PCR using forward and reverse primer pairs flanking the LmCA1 or LmCA2 locus, P13/P14 for LmCA1 and P15/P16 for LmCA2 (Fig. 2A). Correct integration of the NEO or HYG gene cassette in LmCA1^+/− or LmCA2^+/− strains was validated by genomic DNA PCR using the NEO or HYG internal primer and external primer from the 5′ or 3′ flanking region of LmCA1 or LmCA2 locus, P17/P21 or P22/P20 for LmCA1-NEO, and P25/P23 or P24/P26 for LmCA2-HYG (Fig. S1A). To generate LmCA1^−/− or LmCA2^−/− null mutants, the G418-resistant LmCA1^+/− and hygromycin resistant LmCA2^+/− promastigotes were further transfected with 10 µg of ^5′LmCA1-HYG-LmCA1^3′ or ^5′LmCA2-NEO-LmCA2^3′ linearized cassettes, respectively. A double-targeted strain for the LmCA1 gene was selected and maintained in 100 µg/ml G418 sulphate and hygromycin B. Subsequently, this strain was plated, and individual colonies were used to verify its genotypic identity by genomic DNA PCR using a NEO, HYG or LmCA1 internal primer and external primer from the 5′ or 3′ flanking region of LmCA1 [P17/P18 or P19/P20 primers for LmCA1 allele verification, P17/P21 or P22/P20 for NEO cassette verification and P17/P23 or P24/P20 for verification of the HYG cassette (Fig. S1D)]. The LmCA2 double-targeted strain failed to grow in the presence of 50 µg/ml G418 sulphate and hygromycin B. A double heterozygous strain (LmCA1^+/−:LmCA2^+/−) was generated by transfecting G418-resistant LmCA1^+/− with 10 µg of ^5′LmCA2-HYG-LmCA2^3′ linearized cassette. This strain was selected and maintained in 100 µg/ml G418 sulphate and hygromycin B. Subsequently, this strain was plated, and individual colonies were selected to verify its genotypic identity by genomic DNA PCR using forward and reverse primer pairs flanking the LmCA1 (P13/P14) or LmCA2 (P15/P16) locus (Fig. 2A).

The LmCA1 and LmCA2 open reading frames were PCR-amplified from L. major genomic DNA using primer sets P27/P28 and P29/P30, respectively. Amplified LmCA1 and LmCA2 gene fragments were cloned into the SmaI site of pXG-SAT or the BamHI site of pXG-PHLEO vector to generate the respective overexpressing constructs. Constructs were verified by sequencing. Subsequently, 30 µg of each construct was transfected into LmCA1^+/− or LmCA2^+/− strains to generate the LmCA1^+/:CM or LmCA2^+/−:CM complementation strains to restore the missing alleles in the heterozygous mutant strains. Additionally, the LmCA1^+/−:LmCA2^+/− strain was sequentially transfected with 30 µg of both the recombinant plasmids to generate a double complementation strain (LmCA1^+/−:LmCA2^+/−:CM strain). The LmCA1^+/−:CM strain was selected and maintained in 100 µg/ml G418 sulphate and 200 µg/ml nourseothricin (Jena Bioscience, Jena), whereas 100 µg/ml hygromycin B and 8 µg/ml phleomycin (Invivogen, San Diego) were used for selection of LmCA2^+/−:CM. The LmCA1^+/−:LmCA2^+/−:CM strain was selected and maintained in the presence of all four antibiotics.

For subcellular localization studies, the cells were mounted on poly-L-lysine-coated coverslips, fixed with acetone:methanol (1:1) and permeabilized with 0.1% Triton X-100. After blocking non-specific binding with 0.2% gelatin, the cells were incubated with the desired primary antibodies for 1.5 h. The following primary antibodies were used for immunostaining: mouse anti-GFP antibody (1:100; cat. no. 2955, Cell Signaling, Danvers, MA), anti-L. donovani adenosine kinase antibody (1:600; gift from Dr Alok Datta, IICB, Kolkata), rabbit anti-GFP antibody (1:50; cat. no. 11122, Molecular Probes, Eugene), mouse anti-sheep Na⁺/K⁺-ATPase antibody (1:20, cat. no. A276; Sigma, St Louis). Cells were then washed and incubated with appropriate secondary antibodies for 1.5 h. The following fluorophore-conjugated secondary antibodies (Molecular Probes) were used at 1:800 dilution: goat anti-mouse Alexa Fluor 488, goat anti-rabbit Alexa Fluor 568, goat anti-rabbit Alexa Fluor 488, goat anti-mouse Alexa Fluor 568. Finally, the cells were embedded in anti-fade mounting medium containing DAPI (VectaShield from Vector Laboratories, Burlingame), and images were acquired with a Zeiss LSM 710 confocal microscope.

The intra-lysosomal pH of untreated or chloroquine-treated J774A.1 macrophages was estimated using LysoTracker Red DND-99 (Invitrogen) as per the manufacturer's instructions. Fluorescence intensities of the samples were visualized under a Carl Zeiss Axio Observer epifluorescence microscope (with an ApoTome attachment).

As described previously, the morphologies of L. major promastigotes, grown in acidic (pH 5.5) medium, were visualized using a Zeiss Supra 55VP scanning electron microscope (SEM), and the images were analyzed with ImageJ software (Pal et al., 2015). At least 60 cells were analyzed for each experimental condition. During the course of the experiment, selection antibiotics were removed from culture medium.

Cytosol and membrane fractions were isolated from L. major promastigote whole-cell lysates using Qproteome cell compartment kit (Qiagen), as per the manual. SDS-PAGE was performed with the cell fractions obtained from 7×10⁶ cells. LmCA1–GFP or LmCA2–GFP was detected by western blot using rabbit anti-GFP antibody (1:1000). The purity of the cytosol or membrane fraction was confirmed by western blot using a rabbit anti-L. donovani adenosine kinase antibody (1:1000) or mouse anti-sheep Na⁺/K⁺- ATPase antibody (1:250), respectively. Incubations with primary antibodies were performed overnight at 4°C, following which the blots were probed with their respective anti-mouse or anti-rabbit horseradish peroxidase (HRP)-conjugated secondary antibody, at a dilution of 1:4000, for 2 h. Finally, blots were developed using SuperSignal West Pico Chemiluminescent substrate (Thermo Scientific) and viewed in a ChemiDoc imaging system (Syngene).

Total RNA was isolated from wild-type or mutant L. major promastigotes using TRIzol reagent (Invitrogen). The RNA was subjected to DNase I (Invitrogen) treatment to remove DNA contaminants. cDNA was synthesized from 1 µg of DNase-I-treated total RNA using an oligo(dT) primer and MMLV reverse transcriptase (Epicentre, Madison). Expression of LmCA1 and LmCA2 genes was quantified by performing semi-quantitative PCR using the cDNA as template and gene-specific primer sets, P31/P32 and P33/P34, respectively. The number of cycles was optimized at 28. The densitometry value of each PCR product was measured using MacBiophotonics ImageJ software. The relative expression of LmCA1 or LmCA2 mRNA was normalized using wild-type cells as the reference sample and the rRNA45 gene as an endogenous control. For amplification of rRNA45 from L. major cDNA, the P35/P36 primer pair was used.

As described previously, carbonic anhydrase activity in whole-cell lysates of wild-type or mutant L. major promastigotes was measured using Maren's end-point titration assay (Pal et al., 2015; Sundaram et al., 1986). 4×10⁷ L. major promastigotes were used for each assay.

Wild-type or mutant L. major strains were seeded in neutral (pH 7.2) or acidic (pH 5.5) medium, in the absence or presence of 5.95 mM NaHCO₃, and their growth was monitored at different time points by counting the cells under a microscope with a haemocytometer. Leishmania cell viability was quantified by using an MTT assay (Invitrogen), as described previously (Pal et al., 2015; Wolday et al., 1999). Briefly, 4×10⁶ wild-type or mutant cells were incubated with 0.5 mg/ml MTT for 3 h to allow formation of insoluble purple formazan. Cells were then harvested, washed and dissolved in 0.04 N HCl in isopropanol. The absorbance of the solution was measured on a microplate reader at 595 nm wavelength. The percentage viability for each of the mutant strains was calculated using the formula (OD_mutant/OD_{wild type}×100). The viability of the wild-type cells was considered as 100%, which was also verified by live-cell counting by Trypan Blue dye exclusion. Cell viability by Trypan Blue dye exclusion was analyzed using a microscope, as described previously (Wolday et al., 1999). Selection antibiotics were removed from culture medium during the course of all these experiments.

As described previously, intracellular pH of wild-type or mutant L. major strains was measured using the pH-sensitive fluorescent sensor, BCECF-AM (Invitrogen) (Pal et al., 2015; Manzano et al., 2011). Briefly, 1×10⁷ late log phase L. major promastigotes were resuspended in 10 μM BCECF-AM-containing buffer and incubated for 30 min. Cells were then washed and resuspended in the buffer without BCECF-AM. Samples were simultaneously excited at 490 nm (proton-sensitive wavelength) and at 440 nm (isobestic point). The resultant fluorescence intensity ratio, recorded at 535 nm emission wavelength, was measured in a microplate reader. The pH value for each sample was obtained from the corresponding fluorescence intensity ratio using a calibration curve. The calibration curve was generated by recording BCECF-AM fluorescence intensity ratio as a function of pH by incubating the BCECF-loaded parasites in phosphate buffer of varying pH (pH 5–8) containing 5 μg/ml of the ionophore nigericin (Invitrogen). Selection antibiotics were removed from culture medium during the experiment.

Wild-type or mutant promastigotes were grown in medium without selection antibiotics (Thermo Fisher, Waltham, MA), following which 1×10⁷ cells were harvested and seeded in serum-depleted medium containing 200 µg/ml hydroxyurea. After 12 h of cell synchronization, hydroxyurea was removed and cells were transferred to serum-containing acidic (pH 5.5) medium and cultured for 8 h. At indicated time points, cells were harvested, washed and re-suspended in 300 µl PBS. 700 µl of chilled ethanol was added to cell suspension and left overnight at 4°C. Samples were incubated with propidium iodide solution (50 µg/ml propidium iodide, 0.05% Triton X-100 and 100 µg/ml RNase A) for 40 min, and subsequently analyzed on a FACS Calibur instrument (BD Biosciences). The proportions of G₀/G₁, S- and G₂/M-phase populations were determined using BD CellQuest Pro software (Sen Santara et al., 2013). At least 10,000 events were analyzed for each sample.

Macrophage infection was performed as described previously, with minor modifications (Pal et al., 2015; Mukherjee et al., 2009). Briefly, J774A.1 macrophages were activated with 100 ng/ml lipopolysaccharide to enhance their chemotactic and phagocytic abilities (Wu et al., 2009). L. major promastigotes were then added to the macrophages at a ratio of 30:1 (parasite:macrophage), and the infection was allowed to continue for 12 h. The non-phagocytosed parasites were removed by washing, and infected macrophages were incubated for a further 18 h. Wherever indicated, the macrophages were treated with 8 µM chloroquine for 8 h before infection (to make the phagolysosomal compartment more alkaline), and the infection was performed in the presence of chloroquine. Finally, cells were washed, fixed with acetone:methanol (1:1) and embedded in anti-fade mounting medium with DAPI. Intracellular parasite burden (number of amastigotes per 100 macrophages) for each L. major strain was quantified by counting the total number of DAPI-stained nuclei of macrophages and amastigotes in a field. The percentage of macrophages infected by the L. major strain was determined by counting the total number of DAPI-stained nuclei of uninfected and infected macrophages in a field. For each condition, at least 100 macrophages were analyzed.

All statistical analyses were performed using a Student's or paired t-test, and the results were expressed as the mean±s.d. from at least three independent experiments. P-values ≤0.05 were considered statistically significant, and levels of statistical significance are indicated as *P≤0.05, **P<0.01, ***P<0.001, ****P<0.0001.

The authors sincerely thank Dr Stephen M. Beverley (Washington University Medical School, St. Louis) for Leishmania expression and gene targeting vectors. Mr. Ritabrata Ghosh, Mr. Kashinath Sahu, Mr. Tamal Ghosh and Mr. Sujoy Bose are acknowledged for their technical assistance with confocal microscopy, SEM imaging, FACS and Leishmania culturing. The authors are thankful to Prof. Jayasri Das Sarma and Dr Mohit Prasad of IISER Kolkata for help with various reagents used in this work.