In Trypanosoma cruzi, the etiological agent of Chagas disease, the amino acid proline participates in processes related to T. cruzi survival and infection, such as ATP production, cell differentiation, host-cell invasion, and in protection against osmotic, nutritional, and thermal stresses and oxidative imbalance. However, little is known about proline biosynthesis in this parasite. Δ1-Pyrroline-5-carboxylate reductase (P5CR, EC 220.127.116.11) catalyzes the biosynthesis of proline from Δ1-pyrroline-5-carboxylate (P5C) with concomitant NADPH oxidation. Herein, we show that unlike other eukaryotes, T. cruzi biosynthesizes proline from P5C, which is produced exclusively from glutamate. We found that TcP5CR is an NADPH-dependent cytosolic enzyme with a for P5C of 27.7 μM and with a higher expression in the insect-resident form of the parasite. High concentrations of the co-substrate NADPH partially inhibited TcP5CR activity, prompting us to analyze multiple kinetic inhibition models. The model that best explained the obtained data included a non-competitive substrate inhibition mechanism (). Therefore, TcP5CR is a candidate as a regulatory factor of this pathway. Finally, we show that P5C can exit trypanosomatid mitochondria in conditions that do not compromise organelle integrity. These observations, together with previously reported results, lead us to propose that in T. cruzi TcP5CR participates in a redox shuttle between the mitochondria and the cytoplasm. In this model, cytoplasmic redox equivalents from NADPH pools are transferred to the mitochondria using proline as a reduced metabolite, and shuttling to fuel electrons to the respiratory chain through proline oxidation by its cognate dehydrogenase.
Proline (Pro) plays important roles beyond the biosynthesis of proteins. Pro is involved in cellular energy supply, redox homeostasis and stress protection in bacteria [1–3], yeast [4,5], plants  and mammalian cells . Notably, pathogenic trypanosomes, such as Trypanosoma cruzi (the etiological agent of Chagas disease), are not an exception . In T. cruzi, Pro is a relevant energy and carbon source [9–11] and participates in several processes, such as cell differentiation [12–14], cellular invasion [9,15], osmotic control  and resistance to different stresses, particularly those caused by oxidative imbalance [10,17,18]. These different functions of Pro contribute to the ability of T. cruzi to face the challenges encountered during its life cycle.
The life cycle of T. cruzi is complex and involves the colonization of several different environments, such as the bloodstream, the mammalian cell cytoplasm and different portions of the digestive tract of triatomine insects. To inhabit these environments, the parasite presents several stages that are adapted to each environmental condition it experiences during the infection of its hosts. Insect hosts become infected when they take their bloodmeal on an infected mammal having infective non-replicative trypomastigotes circulating in the bloodstream. Once in the digestive tube of the triatomine, these parasites differentiate into the proliferative stage epimastigote, which colonizes the gut. In the terminal portion of the digestive tube, epimastigotes differentiate into non-replicative metacyclic trypomastigotes, which are infective to mammals. During a new bloodmeal, the infected insect expels metacyclic forms of the parasite with the feces, which are able to invade host cells once contact is made and differentiate into replicative intracellular forms termed amastigotes. After several cell divisions, amastigotes differentiate again into infective trypomastigotes, passing through a transient stage called intracellular epimastigotes due to their similarities with the epimastigotes occurring in the insect gut. Finally, trypomastigotes can invade neighbor cells or reach the bloodstream, where the trypomastigotes are able to infect a new vector during its bloodmeal [19,20].
Cells can uptake Pro from the environment or produce it through two different pathways: one involving ornithine or another involving glutamate (Glu). Most organisms using the ornithine pathway can use ornithine to produce Δ1-pyrroline-5-carboxylate (P5C) through a reaction catalyzed by an ornithine aminotransferase (EC 18.104.22.168) . However, some bacteria can produce Pro directly from ornithine in a reaction catalyzed by ornithine cyclodeaminase (EC 22.214.171.124) [22–24]. The pathway involving Glu is considered the main pathway in most organisms and usually involves three enzymatically catalyzed steps : (i) the phosphorylation of Glu to form γ-glutamyl phosphate (catalyzed by a γ-glutamyl kinase — GK, EC 126.96.36.199) using ATP; (ii) the reduction of γ-glutamyl phosphate into glutamate γ-semialdehyde (γGSA) (catalyzed by a γ-glutamyl phosphate reductase — GPR, EC 188.8.131.52) with the concomitant oxidation of NADPH ; and (iii) γGSA undergoes a non-enzymatic cyclization to form P5C , which is reduced to Pro (reaction catalyzed by a P5C reductase — P5CR, EC 184.108.40.206) using NAD(P)H as an electron donor . Notably, in some organisms, such as plants and animals, these first two steps are catalyzed by a single bifunctional enzyme called P5C synthase (P5CS, EC not assigned) [29,30].
We currently know that T. cruzi can uptake Pro from the extracellular medium through two active transporters [18,31] and can oxidize it into Glu through reactions catalyzed by a Pro dehydrogenase and a P5C dehydrogenase [9–11]. In addition, several isoforms of Pro racemases (EC 220.127.116.11) are able to interconvert l- and d-Pro [32,33].
Despite the recognized relevance of Pro in T. cruzi biology and the presence of genes hypothetically encoding both enzymes involved in synthesizing Pro from Glu (P5CS and P5CR) in the T. cruzi genome , there is no information about Pro biosynthesis and if Pro biosynthesis occurs, how Pro is biosynthesized in this parasite. In this work, we show that T. cruzi produces Pro from P5C through a P5CR and show evidence indicating a kinetic regulation of the activity of this enzyme by NADPH. In addition, we show data supporting the possible participation of P5CR in an electron shuttle mechanism connecting the cytosol and the mitochondria with the redox pair Pro/P5C as electron carriers between these compartments.
Materials and methods
Animals were used to obtain the polyclonal serum against the recombinant TcP5CR protein used in this work. All the manipulations were performed in the animal facility of the Department of Parasitology, at the Institute of Biomedicla Sciences — University of São Paulo. The immunization protocol followed herein was approved by the Ethical Committee of the Institute of Biomedical Sciences, University of São Paulo. All procedures followed Brazilian regulations and were approved under protocol number 129/2012.
Microorganisms and culture
Epimastigotes of T. cruzi (strain CL14, clone 14) were maintained at 28°C in liver infusion tryptose (LIT) medium supplemented with 10% (v/v) heat-inactivated fetal calf serum (FCS) . To obtain metacyclic trypomastigotes, epimastigotes in the stationary growth phase  were incubated in triatomine artificial urine (TAU) medium  for 2 h; then, epimastigotes were grown in TAU supplemented with 2 mM aspartate, 50 mM glutamate, 10 mM Pro and 10 mM glucose for 7 days at 28°C for complete in vitro differentiation. The metacyclic forms were purified using ion-exchange chromatography in diethylaminoethylcellulose columns as described previously. Trypomastigotes, amastigotes and intracellular epimastigotes were obtained by infecting the Chinese hamster ovary cell line (CHO-K1) as described by Tonelli et al. . However, the intracellular forms were isolated using the protocol described by our group . For the TcP5CDH induction experiments, the procyclic forms of T. b brucei Lister 427 29-13 (T7-RNAp+ NEO+ TET+ HYG+) (PCFs) co-expressing T7 RNA polymerase and the tetracycline repressor, designed for the conditional expression of P5CDH, were cultured and induced (or not, as controls) with tetracycline as previously described .
Measurement of intracellular free Pro levels
Epimastigote forms of T. cruzi were incubated in PBS (137 mM NaCl, 2.7 mM KCl, 8 mM Na2HPO4, 1.5 mM KH2PO4, adjusted to pH 7.3) for 2 h to deplete the intracellular pool of free Pro. Then, parasites were incubated for 1 h in the presence of different carbon sources and cofactors (Appendix S1) to determine which form of T. cruzi was able to restore the intracellular Pro levels. In addition, 5 mM l-Pro (positive control, by uptake) and non-supplemented PBS (negative control) were used. Then, parasites were washed with cold PBS and centrifuged (3000 g for 5 min at 4°C). Pellets were re-suspended in 100 µl of lysis buffer (100 mM Tris–HCl pH 8.1, 0.25 M sorbitol, 1 mM EDTA, 1% (v/v) Triton X-100, 1 mM phenylmethanesulfonylfluoride (PMSF), 4 µg/ml aprotinin, 10 µg/ml tosyl-l-lysyl-chloromethane hydrochloride (TLCK) and 10 µM E-64) and submitted to two cycles of snap freezing in liquid nitrogen and thawing. Crude extracts were clarified by centrifugation (15 000 g for 15 min at 4°C), and 100 µl of supernatant was mixed (in a separate reaction) with one volume of 20% (w/v) trichloroacetic acid for deproteinization. Samples were precipitated by centrifugation (20 000 g for 30 min at 4°C), and 200 µl of the resultant supernatant was used for proline quantification using ninhydrin assay . Briefly, 200 μl of each sample was incubated with 200 μl of glacial acetic acid and 200 μl of a freshly prepared ninhydrin solution (0.25 g ninhydrin diluted in a mixture of 6 ml of glacial acetic acid and 4 ml of 6 M phosphoric acid) for 1 hour at 100°C. Then, the samples were transferred to ice for at least 1 min. The organic solutes (including those that reacted with ninhydrin) were extracted by partitioning with 400 μl of toluene. The organic phase in each sample was separated by decantation (10 min). The organic phase (100 μl) was transferred to a 96-well polypropylene plate. The proline concentration was obtained by a spectrofotometrical absorbance measurement at 515 nm and compared with a calibration curve in which known proline concentrations (in a range between 100 and 600 μM) were used as standards.
Trypanosomes were suspended in lysis buffer (20 mM Tris–HCl pH 7.9, 0.25 M sucrose, 1 mM EDTA, 0.1% (v/v) Triton X-100, 1 mM PMSF, 4 µg/ml aprotinin, 10 µg/ml TLCK and 10 µM E-64) and submitted to five cycles of sonic disintegration, 5 s each, at 4°C and maximum power in a Branson Digital Sonifier®. The samples were centrifuged at 18 000 g for 30 min at 4°C. The protein concentration was determined by the Bradford assay using bovine serum albumin as a standard .
The full-length ORF encoding Δ1-pyrroline-5-carboxylate reductase (P5CR) was identified (TritrypDB accession number: TcCLB.509207.90) from the T. cruzi genome project database . Based on this sequence, gene-specific primers were designed to add restriction sites for the enzymes BamHI and EcoRI at the ends of the amplified target (underlined sequence) for TcP5CR-sense and TcP5CR-antisense, respectively (Appendix S2). With these gene-specific primers, the TcP5CR coding region was amplified by PCR using genomic DNA of T. cruzi CL14 as a template. A fragment of the expected size (≈0.8 kb) was amplified and cloned into the pGEM-T Easy vector (Promega®) according to the manufacturer's instructions. To express recombinant TcP5CR in the bacteria, the gene encoding the putative P5CR enzyme was further sub-cloned into the pET28a (+) expression vector (Novagen®). This construct allows for the recombinant expression of a C-terminal His6-tagged fusion protein. All selected clones were sequenced, and the expected identity of the cloned DNA fragments was confirmed by using BLAST software (http://blast.ncbi.nlm.nih.gov/).
Expression and purification of the recombinant TcP5CR protein
The plasmid construction containing TcP5CR (pET28-TcP5CR) was used to transform Escherichia coli BL21-CodonPlus (DE3) cells (Stratagene®). The DE3 cells were grown in Lysogeny broth (LB) medium containing 30 µg/ml kanamycin and 5 µg/ml tetracycline at 37°C until an OD600 of 0.6 was reached. Expression of rTcP5CR was induced by the addition of 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG), and cells were maintained at 20°C overnight under constant agitation (250 rpm). The rTcP5CR protein was purified by metal-chelate affinity chromatography using a Ni2+-nitriloacetic acid (NTA) column (Qiagen®) according to the manufacturer's instructions. After rTcP5CR was eluted, it was dialyzed against dialysis buffer (100 mM Tris–HCl, 10 mM NaCl, 5 mM MgCl2, 10 mM 2-mercaptoethanol, 10% glycerol (v/v), pH 7,0). Protein samples were assessed by SDS–PAGE , and protein concentrations were determined using the method described by the Bradford method .
Polyclonal serum production and Western blot analysis
Polyclonal mouse antisera produced against tyrosine aminotransferase (TcTAT), aspartate aminotransferase (TcmASAT), and rabbit polyclonal antibodies against the cytosolic isoform of malate dehydrogenase (TcMDHc) were kindly provided by Dr. Cristina Nowicki (Faculty of Pharmacy and Biochemistry, University of Buenos Aires, Argentina) [43–46]. Likewise, rabbit polyclonal antibodies raised against T. brucei enolase (TbEnolase) and against TbASCT were kindly provided by Dr. Frédéric Bringaud (University of Bordeaux, France). Recombinant glyceraldehyde-3-phosphate dehydrogenase (TcGAPDH) was obtained as reported previously . This protein, as well as recombinant TcP5CR, was used as immunogens to produce polyclonal antibodies in mice according to a standard protocol . Anesthetics (100 mg/kg ketamine + 10 mg/kg xylazine) were used for all procedures involving animal manipulations.
For Western blot analysis, trypanosome cell-free extracts were submitted to protein electrophoresis (SDS–PAGE), and an equal amount of protein (30 µg) was loaded per lane. Proteins were transferred to 45 µm Hybond-C extra nitrocellulose membranes (GE Heathcare) and blocked with PBS buffer plus 0.1% (v/v) Tween-20 (PBS-T) supplemented with 5% (w/v) skim milk powder. Next, antibody β-tubulin (1 : 500) purchased from Invitrogen (#32-2600) and polyclonal mouse antisera produced against TcP5CR (1 : 30 000), TcGAPDH (1 : 2000), TcTAT (1 : 1000) or TcASATm (1 : 3000) or polyclonal rabbit antisera produced against TbEnolase (1 : 10 000) or TbASCT (1 : 32 000) were diluted in PBS containing 0.1% Tween-20 (v/v) and used to probe for the blotted proteins (16 h at 4°C). The blots were washed with PBS-T and incubated with goat anti-mouse IgG horseradish peroxidase (Sigma®) diluted in PBS-T (1 : 2500). Protein signals were developed using SuperSignal West Pico Chemiluminescent ECL substrate (Thermo Scientific) following the manufacturer's instructions.
Subcellular localization of TcP5CR
T. cruzi epimastigotes were submitted to selective digitonin permeabilization as described previously . Briefly, parasites (6.4 × 108 cells) were re-suspended in 1 ml of TSB buffer (25 mM Tris–HCl, pH 7.6, 0.25 M sucrose, 1 mM EDTA, 1 mM PMSF, 4 µg/ml aprotinin, 10 µg/ml TLCK and 10 µM E-64) in the presence of increasing concentrations of digitonin (ranging from 0 to 2.5 mg/ml) (Sigma®). The cells were incubated at 25°C for 5 min and then centrifuged at 18 000 g for 2 min at room temperature. The supernatants were gently separated immediately while the pellets were re-suspended in TSB and disrupted by sonication. The activity of pyruvate kinase (PK), hexokinase and P5CDH (which are cytosolic, glycosomal and mitochondrial markers, respectively) and TcP5CR was measured in all supernatant and pellet fractions as described elsewhere. Supernatants corresponding to solubilized fractions were mixed with 1× SDS Laemmli buffer and analyzed by Western blotting.
Different life stages of T. cruzi (epimastigotes, metacyclic trypomastigotes, amastigotes, intracellular epimastigotes and trypomastigotes derived from cell infection) were used in immunofluorescence microscopy by assessing the co-localization of TcP5CR with cytosolic TcMDH. For this purpose, the polyclonal antibodies against TcP5CR were affinity purified (anti-TcP5CRpurif.) by coupling with 100 µg of rTcP5CR previously transferred onto 45 µm Hybond-C extra nitrocellulose membranes (GE Healthcare). The areas of the nitrocellulose membranes corresponding to the recombinant protein were cut off and separated from the remainder. These fragments were blocked for 1 h in PBS-T supplemented with 5% (w/v) skim milk powder and incubated with 500 µl of anti-TcP5CR serum equilibrated in PBS for 3 h at room temperature. After the fragments were washed three times with PBS-T, they were incubated in 1 ml of a solution of 100 mM glycine pH 2 for 10 min. The glycine solution was collected and gently mixed in 100 μl of 1.5 mM Tris–HCl buffer pH 8.5. This remaining solution with TcP5CRpurif. was subjected to dialysis in PBS buffer supplemented with 10% glycerol, 1 mM PMSF, 4 µg/ml aprotinin, 10 µg/ml TLCK and 10 µM E-64 at 4°C. Finally, 1 mg/ml BSA was added to the purified antibody and stored at 20°C in 20 μl aliquots. For the immunofluorescence assay, the cells were washed with PBS, fixed with 3% (v/v) paraformaldehyde in PBS for 20 min at room temperature, permeabilized by the addition of 0.1% Triton X-100 (for 5 min), and blocked with 1% BSA dissolved in PBS (PBS–BSA) (for 30 min). For antibody staining, polyclonal antisera produced against TcMDH (1 : 400) and TcP5CRpurif. (1 : 8) were dissolved in PBS–BSA and incubated for 16 h at 4°C. After five washes, the preparations were incubated with AlexaFluor488-coupled goat anti-mouse immunoglobulin G (Invitrogen®) and with AlexaFluor546-coupled goat anti-rabbit immunoglobulin G (Invitrogen®) secondary antibodies (1 : 400) for 30 min under light-protected conditions. DNA was stained by further incubating with Hoechst 33258 probe (1 : 2000 in PBS–BSA) (Invitrogen®) for 20 min under light-protected conditions. Next, 2 µl of Fluoromount-G (GE, Healthcare) was added, and a cover slip was mounted. Fluorescence images were obtained using an Olympus IX81 confocal microscope with a 100X 1.35NA objective and processed using Autoquant X2.1 deconvolution software (Media Cybernetics, Inc.).
The reactions catalyzed by TcP5CR were monitored by spectrophotometry using 1 cm quartz cuvettes at 28°C. The progression of the reaction was followed by measuring the consumption of NADPH (). The substrate P5C used was a racemic mixture of DL-Δ1-pyrroline-5-carboxylate (DL-P5C), which was synthesized via meta-peroxidation of DL-5-hydroxylysine (Sigma®) and purified by ion-exchange chromatography using resin DOWEX 50WX8-400 (Sigma®) . Assuming that the DL-P5C racemic mixture contained equal concentrations of each isomer and that TcP5CR catalyzes only the reduction of the isomer L-P5C to produce Pro, the kinetic parameters for this substrate were determined by non-linear regression analysis of the initial reaction velocity versus half the DL-P5C concentration (5–750 µM) using the Michaelis–Menten equation. The reaction mixture contained 100 mM Tris–HCl buffer pH 7.0, P5C (freshly prepared at various concentrations), and 35 µM NADPH and was started by adding 1.5 µg of recombinant rTcP5CR or 100 µg of cell-free total extract in a final volume of 1 ml. The measured rates were corrected with the rates obtained with the same reaction mixture but without P5C. Regarding the study of the effect of NADPH concentration on P5CR activity, reactions with varying concentrations of this cofactor (4–200 µM) with a fixed concentration of 0.25 mM P5C were monitored. For this case, the P5CR activity was evaluated both by progress curve analysis and initial rates versus NADPH concentration. The progress curves were analyzed using the software DYNAFIT . DYNAFIT is freely available for academic institutions. Before the progress curve analysis, a Selwyn test was performed to test whether the enzyme was stable during the progression of such reactions  (Supplementary Figure S3). Ten different models of substrate inhibition, together with the simple Michaelis–Menten model, were compared (the Dynafit scripts and data are available as Appendix S3). Additional information about global fitting and model discrimination using Dynafit is also provided as Supplementary Material (Appendix S3). To describe the degree of uncertainty of the best fit parameters, confidence intervals (α = 0.05) were informed.
The optimum pH for recombinant TcP5CR activity was determined between pH 4.0 and 10.0 in different buffers at 100 mM: citrate phosphate (pH 4.0, 5.0), 2-(N-morpholino)ethanesulfonic acid (MES; pH 5.0, 6.0), 3-(N-morpholino)propanesulfonic acid (MOPS; pH 6.6, 7.0), Tris–HCl (pH 7.0, 7.5, 8.0, 9.0) and 2-(cyclohexylamino)ethanesulfonic acid (CHES; pH 9.0, 10.0). The effect of temperature on the specific activity of TcP5CDH was assayed by increasing the reaction temperature within a range of 20–80°C, and this behavior correlated with the activation energy determined by the Arrhenius plot.
Verification of P5C exit from mitochondria to cytoplasm
For this assay, we used T. b. brucei PFC (wt). Additionally, for P5C accumulation in the mitochondria, we used PFC with TbP5CDH silenced by RNAi (RNAiTbP5CDH) . Both wt and RNAiTbP5CDH were grown for 72 h in the absence (tet−) or presence (tet+) of 0.5 µg/ml tetracycline. Then, these parasites were washed with buffer A containing glucose (BAG: 16 mM NaCl, 5.4 mM KCl, 0.8 mM MgSO4, 50 mM HEPES-KOH, pH 7.2 and 5.5 mM d-glucose) . A total of 1 × 109 cells were re-suspended in 0.5 ml of BAG buffer with 1 mM Pro, 1 mM PMSF, 4 µg/ml aprotinin, 10 µg/ml TLCK, 10 µM E-64 and 1.2 mM digitonin and incubated for 1 h at 28°C. After this incubation, 20 µl of each permeabilized parasite was used in enzymatic assays of ProDH and PK as mitochondrial and cytosolic markers, respectively, to confirm the mitochondrial integrity. Thus, once this integrity was checked, an additional 20 µl of each permeabilized parasite was used in enzymatic assays with recombinant TcP5CR in 100 mM Tris–HCl pH 7.0, 35 µM NADPH and 1.5 µg of TcP5CR, with the permeabilized parasite as the substrate for P5C. With this P5CR activity, the P5C concentration in each permeabilized parasite was determined using the previously determined for P5C and the measured in this experiment with 0.5 mM DL-P5C as substrate.
T. cruzi produces Pro from P5C through the glutamate pathway
To verify whether T. cruzi has the capability to biosynthesize Pro, we initially evaluated the ability of different metabolites to promote this process in T. cruzi epimastigotes. To accomplish this task, we subjected epimastigotes to starvation by incubating them in PBS for 2 h, which rendered parasites with diminished levels of intracellular free Pro (Figure S1). Then, these parasites were incubated with different possible precursors of Pro biosynthesis, and the recovery of intracellular levels of this amino acid was measured. We evaluated two groups of precursors: one group was related to the possible Pro biosynthesis from Glu (Glu itself, P5C/GSA and glutamine), and the other was related to the possible Pro biosynthesis from arginine (arginine itself and ornithine). As a positive control, we incubated the parasites with l-proline to measure the recovery of the intracellular Pro levels by its direct uptake from the extracellular medium (Appendix S1). The intracellular Pro levels in epimastigotes was restored through Pro uptake, as expected, and through Glu, P5C/GSA and glutamine (all carbon sources involved in Pro biosynthesis from Glu) but not from metabolites related to the arginine biosynthesis pathway (Figure 1A), supporting the functionality of the Glu — Pro pathway in T. cruzi.
Proline biosynthesis is operative in
To initiate the biochemical characterization of P5CR, its enzymatic activity was measured in cell-free epimastigote lysates at a single concentration of P5C (0.25 mM) and different NAD(P)H concentrations (35, 55 and 100 µM). The reduction of P5C was up to 10 times faster when NADPH rather than NADH was used at a concentration of 35 μM and six times faster when both were used at 100 μM (Figure 1B). As the initial rates of the reduction of P5C with NADPH as an electron donor in the presence of cell-free epimastigote lysates could be measured, we determined the kinetic parameters for this reaction as a function of the P5C concentration at a fixed concentration of NADPH (35 µM). The obtained data were adjusted to a Michaelis–Menten kinetics model (, mg soluble Protein)−1 (Figure 1C).
Identification of P5CR from
To identify the putative gene(s) encoding the enzyme(s) responsible for the biosynthesis of Pro in T. cruzi, we examined all the sequences annotated as P5CR in the CL-Brener genome database at TriTrypDB (http://tritrypdb.org/tritrypdb/). We found a gene present in a single copy per haplotype (two sequences) annotated as encoding a putative P5CR (systematic names TcCLB.506857.20 and TcCLB.509207.90). Both nucleotide sequences had 810 bp and were 99% similar, with no differences in the corresponding deduced amino acid sequence. For this reason, we arbitrarily chose the sequence TcCLB.509207.90 to continue our analyses (Appendix S2 in Supplementary Text S1). Importantly and according to the annotation, this sequence encodes the critical motifs for any P5CR: motif B, which is responsible for the binding of NAD(P)H, and motif E, which is characteristic of enzymes that bind P5C and Pro. In addition, motif D was also found, which is critical for the formation of a homodimer, the basic structural conformation of most of the active P5CRs  (Figure 2).
TcP5CR alignment with some orthologous.
Characterization of TcP5CR activity
Once we identified the putative gene encoding the P5C reductase of T. cruzi (TcP5CR), we were interested in characterizing the activity of the encoded protein. To achieve that purpose, we obtained an active recombinant version of P5CR from T. cruzi (rTcP5CR) and tested it under the conditions previously established for the activity measurements in extracts. When the P5CR activity was measured as a function of P5C concentration, at a fixed concentration (35 µM) of NADPH, we obtained a , which was consistent with the value obtained using the cell-free epimastigote lysates and a . With the recombinant protein, it was possible to determine the value of the catalytic constant as (Table 1).
|Substrate .||Km (µM) .||kcat (s−1) .||Ki (µM)1 .||Approach .|
|P5C||27.7 [20.9 … 34.4]||1.00 [0.95 … 1.05]||n.d.||Initial rates at fixed NADPH concentration (35 µM)|
|NADPH||12.8 [12.6 … 13.0]||1.99 [1.97 … 2.00]||45.0 [44.3 … 45.6]||Progress curves at fixed P5C concentration (250 µM)|
|Substrate .||Km (µM) .||kcat (s−1) .||Ki (µM)1 .||Approach .|
|P5C||27.7 [20.9 … 34.4]||1.00 [0.95 … 1.05]||n.d.||Initial rates at fixed NADPH concentration (35 µM)|
|NADPH||12.8 [12.6 … 13.0]||1.99 [1.97 … 2.00]||45.0 [44.3 … 45.6]||Progress curves at fixed P5C concentration (250 µM)|
Confidence intervals were calculated for α = 0.05. n.d., not determined.
Ki: non-competitive substrate inhibition constant.
Reaction conditions such as pH and temperature influence enzyme activities; therefore, we explored how these variables affected rTcP5CR. The enzyme showed an optimal pH at 7.0 (Figure 3A) and a maximum activity at 70°C (in a range between 20 and 80°C) (Figure 3B). The increase in activity in relation to the temperature was roughly linear from 20 to 70°C; thus, this range was used to calculate the activation energy by the Arrhenius equation, which resulted in a value of 19.2 ± 1.9 ([23.2 … 15.2]) kJ mol−1 (see the Arrhenius plot as inset in Figure 3B). These changes can be attributed to the intrinsic catalytic properties of the enzyme. However, the effect of pH and temperature could be explained as well as a consequence of changes of in enzyme stability or a combination of both.
Effects of pH and temperature variations on the activity of recombinant TcP5CR.
Inhibitory effect of NADPH on recombinant TcP5CR activity.
TcP5CR is a cytosolic enzyme
As mentioned before, the T. cruzi Pro oxidation pathway has been characterized and occurs inside the mitochondrion . To determine whether biosynthesis occurs in the same compartment, we used two complementary approaches to determine the subcellular location of TcP5CR. The first approach consisted of detecting the protein by Western blotting using a specific serum raised against TcP5CR (Figure S4A) and measuring its enzymatic activity in different subcellular fractions of T. cruzi epimastigotes obtained by selective permeabilization of the cells with increased digitonin concentrations. The second approach consisted of analyzing the subcellular localization of TcP5CR by immunofluorescence using affinity-purified polyclonal antibodies against TcP5CR (Figures S4B and S5). In this case, we could extend the analysis to the other forms of the parasite. Therefore, we were able to determine the subcellular location of the enzyme in epimastigotes, metacyclic trypomastigotes, amastigotes, intracellular epimastigotes and trypomastigotes of T. cruzi.
The Western blot and activity measurement profiles of the different digitonin-treated fractions were compared with those obtained for proteins with well-established subcellular locations. In the Western blot analysis, TcP5CR was detected as released in all digitonized fractions, following the same pattern as the cytosolic marker TcTAT (Figure 5A). As controls, glycosomal and mitochondrial markers (glyceraldehyde-3-phosphate dehydrogenase and aspartate aminotransferase, respectively) were investigated in all fractions and found to be released only at digitonin concentrations above 0.5 mg/ml (Figure 5A). These data were confirmed by measuring the specific activity of P5CR in different digitonin fractions and compared with a different panel of markers. The profile of P5CR activity overlapped with that of PK (cytosolic marker), even though the latter was detected in fractions obtained with higher (1–2.5 mg/ml) digitonin concentrations as well (Figure 5B). Notably, different profiles were obtained for hexokinase and P5CDH (glycosomal and mitochondrial markers, respectively), which showed very modest activity at low (0.1 mg/ml) digitonin concentrations (Figure 5B). We confirmed these results by immunofluorescence assays using an affinity-purified anti-P5CR polyclonal serum. As a control for co-localization, we used a polyclonal antiserum raised against a cytosolic isoform of malate dehydrogenase like (TcMDHc) which actually presents l-alpha hydroxyl acid dehydrogenase activity [46,56]. Our immunolabelling assays revealed that TcP5CR partially co-localizes with the cytosolic marker TcMDHc, and this cellular distribution is consistent in all stages of the T. cruzi life cycle (Figure 6).
Subcellular localization of P5CR enzyme by digitonin fractionation.
Subcellular localization of P5CR enzyme in all
T. cruzi stages.
T. cruzi forms occurring in the insect vector showed the highest levels of TcP5CR expression
As shown in the immunofluorescence assays, P5CR is present in all the T. cruzi life cycle stages. However, their levels of expression in the different stages during the T. cruzi life cycle have not been accurately determined. To verify the possible variations in TcP5CR and the potential for Pro biosynthesis in all stages, we compared the abundance of TcP5CR and its specific activity in all the T. cruzi life cycle stages. Western blotting analysis showed that TcP5CR had higher expression in both insect forms, epimastigotes and metacyclic trypomastigotes, and it had only detectable levels in cell culture-derived trypomastigotes when host-cell stages were analyzed (Figure 7A). Interestingly, the protein levels were below our detection threshold in amastigotes or intracellular epimastigotes. However, the TcP5CR activity in amastigotes and intracellular epimastigotes was detectable. Moreover, amastigotes did not show a significant difference in TcP5CR activity when compared with trypomastigotes, and the TcP5CR activity in amastigotes and trypomastigotes was slightly higher than that measured in intracellular epimastigotes (which had the lowest activity among the measured stages). Consistent with the Western blotting results, the highest specific activities were measured in the insect stages. Uninfected CHO-K1 was used as a control, and no significant activity was observed, confirming that the P5CR activity measured in the host-cell derived stages is specific to the parasite and not due to a contamination by a host-cell derived enzyme (Figure 7B).
Analysis of P5CR expression throughout parasite life stages.
Mitochondrial P5C can be the source of cytoplasmic Pro biosynthesis
Until now, we showed that Pro is biosynthesized from P5C ‘through the capture of redox power’ of cytoplasmic NADPH, and previous work from our group showed that Pro can be further transported into the mitochondria . Once inside the mitochondrial matrix, Pro is able to ‘deliver’ reducing equivalents to FAD (yielding FADH2) while turning into P5C , which can be further oxidized into glutamate . Alternatively, it was proposed that Pro/P5C could constitute an electron shuttle between the cytoplasm and the mitochondria [57–59]. For such a Pro/P5C shuttle to exist and to be functional, the following conditions should be verified: (i) Pro must be able to enter the mitochondria; (ii) Pro must be oxidized inside the mitochondria delivering electrons there; (iii) the resulting P5C must be able to exit the mitochondria towards the cytosol; and (iv) P5C must be able to be reduced into Pro. Pro entry into the mitochondria and its further oxidation into P5C, with the delivery of electrons inside the mitochondria (more specifically into the respiratory chain), have been shown (for T. cruzi see [10,60]), and condition iv (P5C reduction to form Pro) is demonstrated in this work. However, so far, the exit of P5C from the mitochondria has not been demonstrated in any cell. We reasoned that upon the accumulation of mitochondrial P5C, it would be possible to demonstrate its exit from the mitochondria into the cytosol. As P5CDH seems to be essential in T. cruzi (unpublished data), and there is no genetic tools available for producing inducible knock outs or knock downs for essential genes in this parasite, we used a lineage of a related microorganism, Trypanosoma brucei brucei, in which it was possible to induce the knock down of P5CDH . Therefore, we initially induced the parasites RNAiTbP5CDH and control (wt) with tetracycline to diminish the P5CDH levels in the former lineage, or not (in the wt control). Then, we confirmed by Western blot that the amount of P5CDH was significantly decreased in the RNAi lineage, while the enzyme level remained unaltered in the wt, confirming that the knock down system was working properly. As controls of the solubilization with diginonin of the cytosolic and mitochondrial contents we used the markers enolase and Acetyl:succinate CoA-transferase (Figure 8A). Simultaneously, these parasites were permeabilized with digitonin until the solubilization of the cytosolic content kept the mitochondrial membranes intact. As a control of the solubilization of cytosolic content and the integrity of the mitochondrial membrane, we measured the enzymatic activities in the supernatant of well-established enzymatic markers: pyruvate kinase (PK — cytosolic marker) and Pro dehydrogenase (ProDH — mitochondrial marker). The presence of PK and the absence of ProDH activities confirmed the release of cytosolic enzymes but not mitochondrial enzymes, confirming the mitochondrial integrity (Figure 8B and Figure S6). After treating the permeabilized cells with 1 mM Pro for 1 h, we measured the P5C in the soluble fraction. As expected, supernatants of wt procyclics and non-induced RNAiTbP5CDH (tet−) showed no detectable P5C amounts, while supernatants of digitonized P5CDH-knocked down parasites had accumulated P5C outside the mitochondria (Figure 8C). The antibodies against TcP5CDH recognized TbP5CDH only in the wt and RNAiTbP5CDH tet-insoluble fractions, while antibodies against Acetyl:sucinnate CoA-transferase (TbASCT) recognized the enzyme in all insoluble fractions, and antibodies against TbEnolase recognized the enzyme in insoluble and soluble fractions (Figure 8A). Taken together, these data show that P5C is able to exit the mitochondria (as schematized in Figure 8D).
Verification of the release of P5C from the mitochondrial space towards the cytosol.
Proline is involved in several major cellular processes governing T. cruzi biology, as previously described. Here, we demonstrated that this parasite is able to produce Pro from Glu. It is worth mentioning that Glu in turn can be biosynthesized from several internal sources (for example, α-ketoglutarate [43,61,62] or glutamine ) or can be taken up from the extracellular medium . Among trypanosomatids, de novo biosynthesis of Pro from Glu has been described for Leishmania spp. [65–67], but it does not occur in T. brucei.
TcP5CR showed a strong preference for NADPH as a coenzyme, suggesting that this is the preferred electron donor in vivo [68,69]. The kinetic analysis showed a Km for P5C in a non-toxic range [38,70,71]. As previously described for P5CRs from other species, the enzyme of T. cruzi exhibited maximal catalytic activity at physiological pH 7.0–7.5 [28,68,72,73]. In aqueous solution, the cyclic P5C ring is in equilibrium with γGSA (ring-opened form) , and at a neutral pH, the cyclic form is predominant and thermodynamically preferred over γGSA [9,74]. TcP5CR showed thermal-resistance by displaying activity at temperatures up to 70°C. This characteristic seems to be consistent with the fact that Pro was shown to be involved in the resistance against high temperature stress . The observed increase in activity as a function of temperature could favor Pro production as a mechanism of protection against thermal stress.
An interesting point is the kinetic regulation of TcP5CR activity through substrate inhibition caused by NADPH. Substrate inhibition can be important in physiological conditions: it is estimated that 20% of enzymes can be targets of some kind of substrate inhibitory effect . In humans, for example, tyrosine hydroxylase is inhibited by its substrate tyrosine to maintain stable levels of dopamine synthesis during variations in available tyrosine between meals. Another example is given by acetylcholinesterase, which is inhibited by acetylcholine, preventing its rapid degradation before reaching its receptors at the post-synaptic membrane . Finally, a well-known example is the strong inhibition of phosphofructokinase by ATP, reducing the glycolytic flux when the energetic charge is high [76,77]. In our work, we observed that NADPH inhibits TcP5CR activity in a non-competitive mode, suggesting that NADPH could have a regulatory binding site. The experiments showing this inhibitory effect were carried out in conditions preventing the artifact known as stray light, which has been the cause of false claiming of inhibitory effects caused by NAD(P)H . In T. cruzi, NADPH is mainly produced in some reactions of the oxidative branch of the pentose phosphate pathway, in the reaction catalyzed by the cytosolic malic enzyme and in the reaction catalyzed by the isocitrate dehydrogenase. Remarkably, a cytosolic NADP+-dependent isocitrate dehydrogenase has been described for T. cruzi, which is significantly more abundant in amastigotes and cell derived and metacyclic trypomastigotes than in epimastigotes, and it was postulated that an increase in the flux through this enzyme (with a concomitant increment in NADPH production) could defend against the oxidative stress this pathogen experiences when infecting the mammalian host . Moreover, a high cytosolic NADPH concentration can occur concomitantly as well through an increase in the activities of the two cytosolic enzymes of the oxidative branch of the pentose phosphate pathway (glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase) when the availability of glucose is high [80–84]. Notably, NADPH is re-oxidized in anabolic reactions. Therefore, to achieve an increase in the NADPH cytosolic concentration, favoring the defence against oxidative stress, it could be beneficial not only to increase the production of NADPH but also to halt or decrease the rate of NADPH consumption by other reactions. In this context, the observed substrate inhibition could contribute to maintaining the transiently high NADPH concentration. The value of the inhibition constant (∼45 µM) suggests that this inhibition could have such a physiological role. This high NADPH concentration would lead to TcP5CR inhibition; then, NADPH can be used to fuel another biosynthetic pathway, for example, the activity of trypanothione reductase, maintaining redox homeostasis [85–87]. In contrast, if NADPH pools are partially decreased, TcP5CR activity would be favored, increasing the Pro content, which would be available for further oxidation in the mitochondria.
Different T. cruzi stages have different Pro requirements. However, a net Pro biosynthesis seems to occur during the trypomastigote phase and/or during the transition between trypomastigote and the intracellular amastigote since both stages show little (if any) Pro uptake activity, but concentrations of this metabolite are ∼3.5 and 8 mM, respectively [14,17]. Taken together, the TcP5CR activity levels found in the mammalian stages of the parasites are consistent with its demand. However, the fact that TcP5CR is much higher in the stages found in the insect vector strongly suggests that Pro biosynthesis and degradation work simultaneously, but compartmentalization of both routes avoids the occurrence of a futile Pro–P5C cycle.
The possible purpose of a Pro biosynthesis–degradation cycle will be further investigated; however, a hypothesis is proposed. Pro and P5C constitute a redox pair, and their interconversion depends on P5CR and ProDH. In T. cruzi, both reactions occur in different subcellular locations. Under this condition, we propose that Pro–P5C could constitute a redox shuttle between these compartments, with the function of regulating the cellular redox balance and feeding electrons in the respiratory chain (Figure 9), as already proposed for other organisms [57,59,88,89]. In plants such as tobacco and Arabidopsis thaliana, it was demonstrated that a Pro–P5C cycle between the mitochondria and the cytosol (similar to what we describe here for T. cruzi) is required to avoid P5C accumulation and to maintain a homeostatic ratio Pro/P5C . In humans, a more complex Pro–P5C cycle comprising three isoforms of HsP5CR was described: one cytosolic- and NADPH-dependent and the other two mitochondrial- and NADH-dependent [91–93]. Phang et al. proposed that this cycle should be analyzed not only in terms of the source and endpoint of this metabolism but in terms of maintaining the levels of available intermediates to connect with other metabolic and even signaling pathways. In T. cruzi, ProDH can produce P5C, ATP (through oxidative phosphorylation) and eventually ROS, while P5CR can produce the recycling of redox equivalents NADPH into NADP+, maintaining total pyridine nucleotide pools and redox balance, which could have a role in regulating glycolysis and the pentose phosphate pathway . Additionally, trypanosomatids do not have a NADPH dehydrogenase enzymatic system.
Scheme of proposed redox shuttle between cytoplasm and mitochondria involving the Pro/P5C redox pair.
The dissection and further demonstration of all possible functional roles of a Pro–P5C cycle is beyond the scope of this paper and requires technical challenges to be solved, mainly in organisms such as T. cruzi, for which a limited variety of genetic tools are available. To show the feasibility of the Pro–P5C cycle, we used procyclics of T. brucei as a proxy for T. cruzi to show for the first time (to the best of our knowledge) that P5C can exit the mitochondria and become available in the extra-mitochondrial spaces. In summary, based on data from the literature  on the presence of an active TcP5CR and on this strategy, we were able to provide information supporting the existence of all the components of a Pro–P5C cycle in T. cruzi. Based on our data, we propose the hypothesis that the Pro–P5C cycle is operative and can transfer reducing equivalents from the cytoplasmic pools of NADPH to the respiratory chain for ATP synthesis, acting as a cytoplasm/mitochondria electron shuttle.
The authors declare that there are no competing interests associated with the manuscript.
This work was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo [grant no. 2016/06034-2] (awarded to A.M.S.); Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) [grant nos 308351/2013-4 and 404769/2018-7] (awarded to A.M.S.), Research Council United Kingdom Global Challenges Research Fund under grant agreement ‘A Global Network for Neglected Tropical Diseases’ [grant no. MR/P027989/1] (awarded to AMS). Sequences used referred in this work were obtained from the TriTrypDB (http://tritrypdb.org/tritrypdb/) under the accession numbers TcCLB.506857.20 and TcCLB.509207.90. LM and ROOS were fellows of CNPq and KO, BSM, CCA and FSD were fellows of FAPESP during this work.
Conception and design, acquisition of data or analysis and interpretation of data: L.M., K.O., B.S.M., C.C.A., R.O.O.S., F.S.D., M.C.E. and A.M.S.; drafting the article or substantively contributing to revisions in intellectual content: L.M., K.O., B.S.M. and A.M.S.; final approval of the version to be published: L.M., K.O., B.S.M., C.C.A., R.O.O.S., F.S.D., M.C.E. and A.M.S.
We acknowledge Ivan Novaski Avino for technical support in the capture and processing of immunofluorescence images; Dr. Cristina Nowicki (Universidad de Buenos Aires, Argentina) for providing rabbit polyclonal antibodies against TcMDHc, Dr. Frédéric Bringaud (University of Bordeaux, France) for providing rabbit polyclonal antibodies raised against TbEnolase and against TbASCT and Dr. Alan H. Fairlamb for critical reading of the manuscript.
Current address: Department of Biosciences, Durham University, DH7 6SL Durham, U.K.