Parasites and culture media

Trypanosoma cruzi epimastigotes of the Y strain (DTU II) expressing GFP (Y-GFP, resistant to G418) [31] and MJ Levin strain (DTU I) were cultivated at 28°C in BHT medium [32], supplemented with 100 U/mL penicillin, 100 μg/mL streptomycin, 10% (v/v) heat-inactivated fetal bovine serum (Natocor) and 20 μg/mL hemin (Sigma). The MJ Levin transgenic cell line was cultivated in medium supplemented with 250 μg/mL G418 (Sigma) for selection. Cell lines expressing GFP did not show differences with respect to the wild-type cells. Leishmania (Leishmania) mexicana (Costa Rica strain) promastigotes, Crithidia fasciculata (ATCC 11745) choanomastigotes and Phytomonas Jma promastigotes were cultured in BHT medium, as described for T. cruzi. Trypanosoma brucei (29–13 strain) procyclic forms were maintained in minimal media consisting on Eagle's minimum essential medium with L-glutamate (US Biological, M3859) supplemented with 0.1 mM L-alanine, 0.1 mM L-asparagine, 0.1 mM L-aspartate, 0.1 mM L-glycine, 0.1 mM L-serine, 2 mM L-glutamine and 5 mM L-proline, 30 mM Na-Hepes (pH 7.3), 26 mM NaHCO₃, 2 mM sodium citrate, 27 mM glucose, with antibiotics and serum similar to T. cruzi and 7.5 μg/mL hemin [33].

All cultures were maintained by periodically diluting (each 6–7 days) in fresh medium. When indicated, parasites were cultured in the semi-defined medium SDM-79 [34] or in a modified semi-defined medium with low riboflavin concentration (20 nM), named SDM-20, both supplemented with antibiotics, 10% serum, 7.5 μg/mL hemin and 1 mM putrescine.

Plasmid constructions

Genomic DNA was extracted from T. cruzi, T. brucei or L. (L.) mexicana with UltraPure Phenol (Invitrogen) according to the manufacturer instructions.

A fragment corresponding to TcRibJ (TcCLB.509885.70) was amplified using T. cruzi genomic DNA as template, primers F-EcoRI-TcRibJ/R-HindIII-TcRibJ and platinum Taq DNA polymerase (Invitrogen). The PCR product was purified, digested with EcoRI and HindIII (New England Biolabs) and ligated using T4 DNA ligase (Invitrogen) into the T. cruzi expression vector pRIBOTEX [35] to yield pRIBOTEX-TcRibJ.

For heterologous complementation assays in E. coli, the recombinant expression vectors pET24a-TcRibJ, pET24a-TbRibJ and pET24a-LmiRibJ were constructed harbouring TcRibJ, TbRibJ (Tb927.5.470) and LmiRibJ (LmxM.08_29.2550) fragments from T. cruzi, T. brucei and L. (L.) mexicana, respectively. A control vector was constructed using a non-related T. cruzi permease, TcPAT12 [36,37]. All fragments were amplified from their respective genomic DNA using high-fidelity PCR platinum Pfx DNA polymerase (Invitrogen) and specific primers: F-TcRibJ-NdeI-6xHis/R-TcRibJ-BamHI for T. cruzi, F-TbRibJ-NdeI-6xHis/R-TbRibJ-BamHI for T. brucei and F-LmiRibJ-NdeI-6xHis/R-LmiRibJ-BamHI for L. (L.) mexicana. The products were cloned by PCR cloning technique using Q5 high-fidelity DNA polymerase (New England Biolabs) and pET24a plasmid (Novagen) as template, according to the manufacturer instructions.

All constructions were corroborated by sequencing. Primers used in this study are listed in S1 Table.

Parasite transfection

T. cruzi MJ Levin strain was transfected with pRIBOTEX-TcRibJ and pRIBOTEX-GFP as follows: 10⁸ epimastigote cells grown in BHT medium were harvested by centrifugation, washed with PBS, and resuspended in 0.35 mL of electroporation buffer (PBS containing 0.5 mM MgCl₂ and 0.1 mM CaCl₂). The cell suspension was mixed with 50 μg of plasmid DNA in 0.2 cm gap cuvettes (Bio-Rad Laboratories). The parasites were electroporated using a single pulse of (400 V, 500 μF), showing a time constant of ~5 ms. Transfected parasites were cultured in fresh BHT for 24 h and later G418 was added at 250 μg/mL. The MJ Levin strain has been selected for these experiments because of its higher success rate in obtaining transformed clones compared to the Y strain, in the conditions tested during this work.

Trypanosomatid proliferation assay

The effect of flavins or their analogs on parasites proliferation was evaluated through growth curves. Stationary phase parasites were inoculated in fresh SDM-79 or SDM-20 media (basal 20 nM riboflavin) supplemented to a defined riboflavin concentration. Initial density for T. cruzi Y-GFP, MJ Levin-GFP (referenced as wild-type) and MJ Levin-TcRibJ (referenced as TcRibJ) was 10⁷ parasite/mL, while initial density for T. brucei, L. (L.) mexicana, C. fasciculata and Phytomonas Jma was 10⁶ parasites/mL. Riboflavin, flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD) were dissolved in water and the analogs lumiflavin, lumichrome and roseoflavin (all from Sigma-Adrich) dissolved in DMSO. Then, flavins and analogs were added to the culture media at the indicated concentrations. Parasites were daily counted using a hemocytometer chamber. Proliferation was calculated as the percentage of parasite counts relative to control condition values (flavin at 20 nM or without analogs) on the fifth day. Day 5 was chosen because it is when all parasite cultures tested reach stationary phase in control conditions (S1 Fig).

Riboflavin transport assay in trypanosomatids

The riboflavin uptake measurements were performed using a radiolabeled [³H]-riboflavin (6.2 Ci/mmol) (Movarek Biochemicals Inc.) tracer, adapted from arginine transport assays performed by Canepa et al. [38] with slight modifications. Briefly, epimastigotes of T. cruzi, choanomastigotes of C. fasciculata and promastigotes of L. mexicana and of Phytomonas Jma were cultured in BHT, while T. brucei procyclic parasites were cultured in SDM-79, to late logarithmic phase. Then, they were harvested and resuspended in fresh media. The cultures were maintained at 28°C until mid-exponential growth, then parasites were harvested, washed three times with PBS-2% glucose, and resuspended in the same buffer for starvation at 28°C with shaking for 3 h. Then, parasites were collected and resuspended in PBS-2% glucose, at a cell density of 300–400 x 10⁶ parasites/mL (30–40 x 10⁶ parasites/tube), and kept at 37°C for 15 min. The assay started after the addition of the radiolabeled riboflavin solution. Riboflavin final concentrations and time points are indicated in each case. To stop the uptake assay, aliquots (0.1 mL) corresponding to each measured point, were placed in 1 mL of stop solution (ice-cold 500 μM unlabeled riboflavin in PBS). Parasites were collected and washed three times with stop solution. Cell pellets were counted for radioactivity in UltimaGold XR liquid scintillation cocktail (Packard Instrument Co., Meridien CT, USA). The kinetic parameters (V_max and apparent K_m) were determined as described in “Kinetic parameters calculation and statistical analysis” section.

Transport displacement assays were performed at 0.3 μM [³H]-riboflavin (concentration near to the apparent K_m value for riboflavin transport estimated in this work) adding unlabeled flavins or analogs at 10- or 100-fold concentration excess.

To compare the transport activity between MJ Levin-TcRibJ and MJ Levin-GFP strains, parasites were cultivated in SDM-20 fresh medium.

Bioinformatics analysis

The sequence of the riboflavin transporter Mch5/YOR306C of Saccharomyces cerevisiae [21] was obtained from the Saccharomyces Genome Database (http://www.yeastgenome.org/) and employed in the TriTryp database server (http://tritrypdb.org/tritrypdb/) to find similar proteins from trypanosomatid genomes.

The Mch5p, TcRibJ, TbRibJ and LmiRibj multiple sequence alignment (MSA) was carried out using the MUSCLE program from the EMBL-EBI server (http://www.ebi.ac.uk/Tools/msa/muscle/). Putative N-glycosylation sites of Mch5p, TcRibJ, TbRibJ and LmiRibJ were found using NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/).

Putative riboflavin transporters similar to TcRibJ from parasites with complete or partially assembled genome sequence were identified using a BLAST search in NCBI database (http://www.ncbi.nlm.nih.gov/assembly) with the exception of Bodo saltans and Trypanoplasma borreli draft genomes that have been deposited in the Sanger Institute database (http://www.sanger.ac.uk/resources/downloads/protozoa/). The similarity between TcRibJ and other RibJ was determined using ClustalW program with Geneious version 4.8.4 software (http://www.geneious.com/). This program was also used to predict the putative transmembrane regions in RibJ.

The maximum-likelihood tree was constructed with the MEGA6 free software, using a ClustalW MSA of RibJ from kinetoplastids, using bootstrap support (500 pseudoreplicates) and the Le and Gascuel model [39] without Gamma Distribution as the best evolutionary model selected by the software. The accession numbers of all sequences are listed in S2 and S3 Tables.

The neighbor-joining tree was performed with MEGA6 free software (http://en.bio-soft.net/tree/MEGA.html) according to the following pipeline: MSA of riboflavin transporters with the ClustalW program; protein distance calculation with the JTT matrix and neighbor-joining consensus tree construction with a bootstrap support (1000 pseudoreplicates); all bioinformatics tools were used with default parameters. The tree was constructed using the amino acidic sequences from mammalian (RFVT/SLC52), fungi (Mch5), nematodes (rft) and kinetoplastid (RibJ) riboflavin transporter families. The results were visualized using iTOL server (http://itol.embl.de/). The accession numbers of all sequences are provided in S4 Table.The pairwise global alignment between TcRibJ and human RFVTs sequences, were performed with the software EMBOSS Needle (http://www.ebi.ac.uk/Tools/psa/emboss_needle/). The accession numbers of all sequences are given in S5 Table.

Complementation growth assay in ΔribB E. coli

The ΔribB::cat E. coli strain (ΔribB), kindly provided by García Angulo et al. [19], was used to functionally characterize RibJ riboflavin transporters. ΔribB is unable to import [17] and synthesize [19] riboflavin, and can only be cultured at high riboflavin concentrations. This strain, also resistant to chloramphenicol, was cultured in LB medium at 37°C in a shaker (250 rpm) in the presence of the antibiotic and with excess of riboflavin (750 μM).

The strain ΔribB was transformed using pET24a-TcRibJ, pET24a-TbRibJ, pET24a-LmiRibJ, pET24d-TcPAT12 or an empty vector (pET24a). As a positive control the pET24a-RibM plasmid was included, which codes for the RibM riboflavin transporter from Streptomyces davawensis and was kindly provided to us by Dr. Matthias Mack [18]. Selection was carried out on LB agar plates supplemented with kanamycin (50 μg/mL) and excess of riboflavin.

For heterologous expression assays (in solid and liquid media), the ∆ribB + pET24a, ∆ribB + pET24a-RibM, ∆ribB + pET24a-TcRibJ, ∆ribB + pET24a-TbRibJ, ∆ribB + pET24a-LmiRibJ or ∆ribB + pET24a-TcPAT12 strains were cultured overnight in LB broth with kanamycin (50 μg/mL) and excess of riboflavin at 37°C with shaking until the stationary phase. Two mL from each culture were centrifuged at 5,000 rpm, pellets were washed twice with PBS and resuspended in fresh LB with kanamycin added, but no riboflavin. Ten μL aliquots (OD₆₀₀ = 0.01) were used to inoculate LB-agar plates or 2 mL of liquid LB, in both cases supplemented with kanamycin (50 μg/mL), IPTG (0.1 mM) and the indicated riboflavin concentrations for each assay. Cultures were incubated at 37°C for 24 h. Plate images were taken and liquid LB medium cultures OD₆₀₀ values were recorded.

Riboflavin transport assay in ΔribB E. coli

The ∆ribB + pET24a, ∆ribB + pET24a-RibM, ∆ribB + pET24a-TcRibJ and ∆ribB + pET24a-TbRibJ strains were cultured overnight in LB medium with kanamycin (50 μg/mL) and excess of riboflavin. Fresh LB medium (kanamycin 50 μg/mL, IPTG 0.4 mM) was inoculated with these bacteria at an initial OD₆₀₀ = 0.01, and incubated at 28°C in the absence of riboflavin, to deplete the intracellular vitamin, until they reached the mid-exponential growth phase (OD₆₀₀ = 0.8–1.2). Cells were harvested, washed three times with PBS- 2% glucose (transport buffer), and resuspended at a final OD₆₀₀ = 5 in the same buffer. The cell suspension was pre-incubated in transport buffer for 15 min at 37°C and the uptake assay started when the [³H]-riboflavin solution was added (final concentration 2 μM). Aliquots of 0.2 mL were taken at indicated times and placed in 1 mL of stop solution (ice-cold 500 μM nonradioactive riboflavin in PBS) [19]. Cells were centrifuged at 13,000 rpm for 1 min and washed three times with stop solution. Bacterial pellets were counted for radioactivity in UltimaGold XR liquid scintillation cocktail.

The displacement assay was performed using the same strains in the presence of 0.3 μM radiolabeled riboflavin and the nonradioactive competitors at 3 μM and 30 μM.

In all cases, the value obtained with ∆ribB + pET24a, which corresponds to unspecific binding, was subtracted from the other measurements.

Antibiotics susceptibility assay

This assay was performed to corroborate that the permeability of ∆ribB E. coli was not affected by the expression of heterologous riboflavin transporters. Transformed E. coli ∆ribB strains were cultivated in liquid LB with excess of riboflavin (750 μM) and bactericidal compounds (0–40 μg/mL nalidixic acid or 0–250 μg/mL acriflavine), then growth IC₅₀ values were determined.

T. cruzi in vitro metacyclogenesis assay

The in vitro differentiation assay was performed as previously described [40]. Briefly, 7-day-old T. cruzi epimastigotes (Y-GFP strain) cultured in SDM-20 were harvested and incubated for 2 h at 37°C in triatomine artificial urine (TAU) medium. Next, parasites were diluted in TAU3AAG medium (TAU supplemented with 10 mM L-proline, 50 mM L-sodium glutamate, 2 mM L- sodium aspartate and 10 mM D-glucose) with the addition of 0–300 nM riboflavin, FMN or FAD, or 10 μM analogs. Parasites were cultured at 28°C for 48 h. The epimastigotes and differentiated metacyclic trypomastigotes (MT) mix was harvested by centrifugation at 600 × g for 15 min and resuspended in 0.5 mL of fresh human serum, which selectively lyses epimastigotes [41]. MT, easily seen by light microscopy, were quantified using a hemocytometer chamber.

T. cruzi in vitro infection assay

The in vitro infection assay was performed following protocols previously described [40]. Briefly, T. cruzi tissue culture trypomastigotes (TCT) (Y-GFP strain), obtained from VERO cells, were pre-treated with 10 μM riboflavin analogs at 37°C for 2 h. Subsequently, monolayers of H9C2 cardiomyoblasts, which had been grown in DMEM-10% fetal bovine serum (FBS) in 24-well plates containing glass coverslips, were infected with the TCT using 10 parasites per cell (MOI 10:1). The co-cultures were maintained in the presence of 10 μM riboflavin analogs at 37°C for 12 h allowing the cardiomyoblast infection. Unbound TCTs were removed by washing with fresh DMEM, and infected mammalian monolayers were incubated in DMEM-3% FBS, 10 μM riboflavin analogs, at 37°C for 48 h. For these assays, riboflavin analogs were dissolved in a DMSO/DMEM solution (1:100 v/v). Then, samples were fixed with 10% paraformaldehyde; the actin cytoskeleton of the H9C2 cells was stained with TRITC-phalloidin (Invitrogen) and parasites were directly visualized due to the stable expression of GFP [40]. Finally, cell invasion (expressed as percentage of infected host cell) and amastigotes proliferation (number of amastigotes per cell) were quantified using confocal microscopy (FV1000 Confocal Olympus microscope).

Kinetic parameters calculation and statistical analysis

Standard procedures were used to determine kinetic parameters. The apparent K_m and V_max values were obtained by nonlinear regression fit of the Michaelis-Menten equation to the data. Statistics, curve fitting, V_max and apparent K_m were calculated using the GraphPad Prism 6 software.

Each experiment was carried out at least three times. Groups were analyzed using one-way ANOVA test followed by a post-hoc Tukey's multiple comparison test (significance cut-off value P = 0.05). The infectivity and amastigote replication assays were analyzed using a nonparametric test (Kruskal-Wallis) and compared against the control condition through Dunn’s test (significance cut-off value P = 0.05).

The correlation analysis was performed between the maximum density values reached with riboflavin-supplemented media from each parasite and the apparent K_m values for riboflavin uptake. The correlation was estimated using the Pearson coefficient (significance cut-off value P = 0.05, two-tailed). A linear regression was represented with confidence intervals at 95%.

Ethics statement

All parasites used during this work are laboratory strains. T. cruzi Y-GFP, Phytomonas Jma and C. fasciculata (ATCC 11745) strains were previously reported by our group [40,42,43]. T. cruzi MJ Levin strain was provided by Dr. Claudio Pereira [44]. T. brucei (29–13 strain) and L. (L.) mexicana (Costa Rica strain) were kindly provided by Dr. Guillermo Alonso [45] and Dr. Carlos Labriola [32], respectively.

Effects of flavins on proliferation of trypanosomatids

As trypanosomatids lack biosynthetic enzymes for B-vitamins, we tested the effects that extracellular flavins exert on proliferation of trypanosomatids. T. cruzi epimastigotes moderately proliferated at low concentration (20 nM) of riboflavin or its derivatives FMN and FAD. Higher concentrations of flavins (< 300 nM) significantly increased the T. cruzi proliferation (74.3 ± 3.6% for riboflavin, 32 ± 3.2% for FMN and 32.8 ± 7.7% for FAD, compared to the control); concentrations higher than 300 nM of riboflavin and FMN had a negative proliferation effect (Fig 1A). Flavins also showed a stimulatory effect on proliferation of T. brucei procyclic parasites, C. fasciculata choanomastigotes and promastigotes of L. (L.) mexicana and Phytomonas Jma, exhibiting slight differences in doses-response profiles (S2 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flavins and chemical analogs are incorporated into T. cruzi epimastigotes and affect their proliferation with opposite effects.

Chemical structures of riboflavin, FMN and FAD, and their analogs roseoflavin, lumiflavin and lumichrome are shown in A and B top panels. A and B bottom panels: T. cruzi Y strain epimastigotes were maintained at 28°C until stationary phase, then washed and incubated in fresh medium with the indicated compound concentrations: (A) flavins and (B) chemical analogs plus 300 nM riboflavin. Parasites were counted daily. T. cruzi proliferation (%) was calculated at the indicated round using fifth day-counts and control conditions -(A) 20 nM flavins or (B) 0 μM analogs- as references (100%). Log-phase Y strain epimastigotes grown in BHT-10% FBS were harvested, washed, resuspended in PBS-2% glucose and incubated at 37°C. (C) Riboflavin uptake velocity was calculated at 0–5 μM final substrate concentration. Aliquots were sampled at 0 and 5 min after the addition of radioactive material. Displacement assays were performed at 0.3 μM radioactive riboflavin mix (Ctrl: control, 100%) and 3–30 μM of (D) unlabeled flavins (RF, FMN or FAD) or (E) unlabeled analogs (RoF: roseoflavin, n = 3; LF: lumiflavin, n = 4; or LC: lumichrome; n = 4). Values are expressed as mean ± SD. Statistical analysis was performed by one way ANOVA test followed by a post-hoc Tukey's multiple comparison test (*P < 0.05, **P < 0.01, ***P < 0.005).

https://doi.org/10.1371/journal.pntd.0005513.g001

On the other hand, the riboflavin analogs roseoflavin [46], lumiflavin and lumichrome [47] affected trypanosomatid proliferation rates. While lumiflavin and lumichrome (at 10 μM) showed effects on T. cruzi in the third round of culture (32.8 ± 6.3% and 28.5 ± 2.5% lower than the control, respectively), roseoflavin treatment resulted in a marked reduction of parasite proliferation in the first round of culture (21.4 ± 0.8% lower than control) (Fig 1B and S3 Fig).

These results suggest that trypanosomatids incorporate extracellular flavins and that one could interfere with flavin uptake and/or the downstream metabolism to affect proliferation of the parasites.

Riboflavin uptake in trypanosomatids

A [³H]-riboflavin transport assay was performed to measure its uptake in trypanosomatids. T. cruzi, as well as T. brucei, L. (L.) mexicana, C. fasciculata and Phytomonas Jma showed riboflavin uptake following Michaelis-Menten kinetics, with a maximal velocity at 1–2 μM of riboflavin and apparent K_m values in the submicromolar range, indicative for the involvement of a high-affinity transporter (Fig 1C, S4 Fig and Table 1). Riboflavin derivatives were efficient competitors of the [³H]-riboflavin uptake in T. cruzi epimastigotes, with FMN being more effective than FAD (Fig 1D). On the other hand, lumichrome showed mild competitive effects, while roseoflavin and lumiflavin significantly reduced 70 to 90% and 35 to 65% the [³H]-riboflavin uptake, respectively (Fig 1E).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Riboflavin apparent K_m and V_max uptake parameters measured in trypanosomatids.

https://doi.org/10.1371/journal.pntd.0005513.t001

RibJ: A novel family of riboflavin transporters in trypanosomatids

As a first approach for in silico studies, sequences of previously characterized transporters from bacteria, Saccharomyces cerevisiae, Caenorhabditis elegans and mammals [12–14,16–21] were used as queries in BLAST searches against the T. cruzi genome. Only the riboflavin transporter Mch5p from S. cerevisiae showed a hit with low identity (21%) and similarity (37%) values. The gene alleles, TcCLB.509885.70 and TcCLB.508397.70, are encoded in chromosome 28-S and 28-P, respectively. This putative riboflavin transporter, which we named here TcRibJ, is 472 amino acids long and contains three possible N-glycosylation sites (N108, N236 and N431) and 12 hydrophobic regions, probably corresponding to 12 transmembrane segments, which are also present in Mch5p (Fig 2A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. RibJ is a novel family of riboflavin transporters characteristic of trypanosomatids.

(A) Multiple sequence alignment between Mch5p (S. cerevisiae), TcRibJ, TbRibJ and LmiRibJ using MUSCLE program. Red boxes: putative transmembrane domains; light blue box: MFS domain; green boxes: the Asn in N-glycosylation context. Black, gray and violet backgrounds represent 100%, 80%, and 60% conservation within similarity groups, respectively. (B) A Neighbor-joining tree (1000 pseudoreplicates) was constructed using the amino acid sequences from riboflavin transporters: 22 mammalian (RFVT/SLC52, light green), 15 fungal (Mch5p, yellow), 5 nematodal (rft, dark green) and 16 RibJ belonging to subclass Metakinetoplastids (light blue, and inside, trypanosomatids group in violet). The blue arrows indicate RibJ from T. cruzi, T. brucei and L. (L.) mexicana, and the human and S. cerevisiae riboflavin transporters.

https://doi.org/10.1371/journal.pntd.0005513.g002

The TcRibJ sequence was used as query to find homologs in T. brucei and L. (L.) mexicana, finding ortholog genes that code for proteins similar to TcRibJ: TbRibJ (62.2% identity) and LmiRibJ (48.9% identity). The three RibJ members showed at the N-terminal region a Major Facilitator Superfamily domain (MFS domains), found in small solute transporters [48] (Fig 2A and S2 Table).

RibJ orthologs, exhibiting an MFS domain and 12 predicted transmembrane segments, were retrieved from totally or partially assembled genomes from representatives of the Metakinetoplastina subclass (S2 and S3 Tables, respectively). RibJ members were even found in phylogenetically distant taxa from T. cruzi, as the free-living kinetoplastid Bodo saltans (order Eubodonida) and the fish endoparasite Trypanoplasma borreli (order Parabodonida), while the Perkinsela sp. genome (subclass Prokinetoplastina) did not show any recognizable homolog. Using all these identified sequences, a phylogenetic maximum-likelihood tree was constructed (S5 Fig) with a topology in accordance with the currently accepted Kinetoplastea phylogeny [49]. Therefore, it is likely that the RibJ transporter family has a common origin in Metakinetoplastids, which diversified reminiscently of the speciation process, with Phytomonas spp. RibJ being the most distant transporter in the order.

To complete the phylogenetic analysis of eukaryotic riboflavin transporters, the amino acid sequences from the human riboflavin transporter (RFVT1/SLC52A1) [12], C. elegans rft-1 [13] and yeast Mch5p [21] were used as queries in a BLAST search in the NCBI database. Representative sequences comprising the mammalian, nematodal and fungal datasets (S4 Table) and, including the kinetoplastid RibJ transporters, were used to construct a neighbor-joining tree (Fig 2B). Riboflavin transporter members cluster in three distinct major groups, revealing that RibJ constitutes a novel family, the first one identified in protists and the third in eukaryotes, distant from the RFVT/SLC52 family and more related to the Mch5p family (97% bootstrap support).

Functional characterization of RibJ members

To evaluate the functionality of RibJ as a riboflavin transporter we designed an experiment of heterologous expression to restore growth of a riboflavin auxotrophic bacterium. The ribB E. coli null mutant strain (∆ribB) [19] was transformed with expression vectors carrying either the TcRibJ, TbRibJ or LmiRibJ gene. An empty vector and one harbouring the TcPAT12 gene (the polyamine transporter of T. cruzi) [36,37] were included as negative controls, and a plasmid encoding the RibM riboflavin transporter from Streptomyces davawensis [18] as a positive control. All transformants grew when plated on LB medium plus an excess (750 μM) of riboflavin, while only ΔribB expressing RibM could grow when no riboflavin was supplemented (Fig 3A), since RibM incorporates riboflavin traces present in the LB medium [19]. When riboflavin was added in a restrictive concentration (5 μM), TcRibJ and TbRibJ also supported ∆ribB growth (Fig 3A). Similar results were obtained in liquid LB medium assays, where TcRibJ and TbRibJ restored the growth capacity in the presence of riboflavin, and also FMN and FAD, although at higher concentrations (Fig 3B). These findings confirm that RibJ proteins possess flavin transporter activity in vivo. In all conditions, negative controls (TcPAT12 or empty vector) failed in restoring ΔribB growth. Strikingly, despite the similarities with the other RibJ, LmiRibJ was unable to transport riboflavin in this heterologous system (Fig 3A and 3B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. TcRibJ and TbRibJ display flavin transport activity in E. coli.

The E. coli ∆ribB strain was transformed with expression plasmids coding for either RibM, TcRibJ, TbRibJ, LmiRibJ or TcPAT12, or with an empty vector. Strains were (A) plated on LB agar (left: strain plating scheme) with the addition of 0, 5 or 750 μM riboflavin (RF) or (B) cultured at 37°C for 24 h in liquid LB supplemented with 0.02–100 μM FMN or FAD, or 0.02–20 μM riboflavin. (C) [³H]-riboflavin uptake (2 μM) was measured from 0 to 180 min in bacteria expressing RibM, TcRibJ or TbRibJ or containing an empty vector. (D) Displacement assays performed with 0.3 μM [³H]-riboflavin in the absence of competitors (Ctrl: control) or in the presence of 3–30 μM of unlabelled riboflavin; aliquots were sampled at 0 and 120 min after the addition of the radioactive material. Values are expressed as mean ± SD. Statistical analysis was performed by one way ANOVA test followed by a post-hoc Tukey's multiple comparison test (*P < 0.05, **P < 0.01, ***P < 0.005).

https://doi.org/10.1371/journal.pntd.0005513.g003

TcRibJ and TbRibJ heterologous expression allowed ΔribB strains to incorporate [³H]-riboflavin with a linear time-dependent velocity for the first 60 min at 4.1 and 7.3 fmol/OD₆₀₀.min, respectively (Fig 3C). The transport specificity was confirmed by displacement assays performed at 10- or 100-fold unlabeled riboflavin excess (Fig 3D). To rule out membrane permeability alterations or unspecific transport due to the heterologous expression of membrane proteins, an antibiotics susceptibility assay was performed. IC₅₀ values for nalidixic acid and acriflavine in E. coli ∆ribB strains (0.47–0.76 μg/mL and 21.3–29.1 μg/mL, respectively) (S6 Fig) were similar to those reported in the literature [50,51], indicating that the flavin uptake measured in these strains was the result of specific flavin transporter activity.

TcRibJ is a flavin transporter in T. cruzi

To confirm the RibJ functionality in trypanosomatids, we transformed T. cruzi epimastigotes with TcRibJ to generate an over-expressing strain (TcRibJ). This strain showed increased proliferation rates in the presence of low concentrations (20 nM) of any flavin (riboflavin, FMN or FAD) (Fig 4A), higher sensibility to riboflavin analogs (Fig 4B) and increased riboflavin uptake (1.5 ± 0.4-fold) (Fig 4C), compared to the wild-type strain. These results confirm that TcRibJ is a functional flavin transporter in T. cruzi, and that its activity affects proliferation of epimastigotes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. TcRibJ functions as flavin transport in vivo in T. cruzi.

Epimastigotes transfected with pRIBOTEX-GFP (wt) or pRIBOTEX-TcRibJ (TcRibJ) were incubated in fresh SDM-20-10% FBS in the presence of (A) 20 nM riboflavin (RF), FMN or FAD, or (B) roseoflavin (RoF), lumiflavin (LF), or lumichrome (LC), at the indicated concentrations. Parasites were counted daily using a hemocytometer chamber. T. cruzi proliferation (%) was calculated at the third culture round, where the control condition (20 nM flavin) was referenced as 100%. (C) [³H]-riboflavin (100 nM) uptake measurements in control and over-expressing TcRibJ epimastigotes. Values are expressed as mean ± SD. (A-B) Statistical analysis was performed by a two-tailed unpaired t test (*P < 0.05, **P < 0.01, ***P < 0.005).

https://doi.org/10.1371/journal.pntd.0005513.g004

Flavins play an important role in T. cruzi life cycle

To get a broader view of the role that flavins play in trypanosomatid physiology, we tested how they affect T. cruzi metacyclogenesis (Fig 5A). When flavins were added to the differentiation medium (TAU3AAG), a significant higher percentage of epimastigotes differentiated to metacyclic trypomastigotes (MT) compared to control conditions, while all flavin analogs produced a dramatic reduction of MT counts (Fig 5B and 5C). Strikingly, roseoflavin completely abolished MT counts (Fig 5C); however, it still remains to be determined whether this effect is a consequence of affecting (i) the viability of epimastigotes in the process of metacyclogenesis, (ii) directly metacyclogenesis and/or (iii) MT viability.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Flavins stimulate while analogs retard progression through T. cruzi life cycle.

(A) In vitro metacyclogenesis assay scheme (for more details, see ‘Material and methods‘ section). Percentage of MT obtained in differentiation media (TAU3AAG) supplemented with (B) flavins (riboflavin: RF, FMN or FAD), or (C) chemical analogs (roseoflavin: RoF, lumiflavin: LF, or lumichrome: LC), at the indicated concentrations at 48 h. (D) In vitro infection assay scheme (for more details, see ‘Material and methods‘ section). Effect of 10 μM analogs on (E) cellular invasion at 48 h, expressed as percentage of H9C2 host cells infected with T. cruzi Y-GFP (∼400 cells counted) or (F) amastigote proliferation at 48 h, expressed as number of T. cruzi Y-GFP amastigotes per infected H9C2 host cell. Values are expressed as mean ± SD. Statistical analysis was performed by a Kruskal-Wallis non-parametric test followed by a post-hoc Dunn's multiple comparison test (*P < 0.05, **P < 0.01, ***P < 0.005).

https://doi.org/10.1371/journal.pntd.0005513.g005

Once inside the mammalian host, cellular invasion and intracellular replication are essential events for a successful T. cruzi infection [40]. Hence, we evaluated the effect of flavins on these processes in an in vitro infection model (Fig 5D). Flavin analogs produced a mild reduction in infected host cell counts (Fig 5E) and reduced amastigote intracellular proliferation, with statistically significance for roseoflavin (11.4 ± 0.3 vs 2.2 ± 2.3 amastigotes per infected cell for control and roseoflavin treatment, respectively) (Fig 5F).

Taken together, these results strongly suggest that flavins are necessary for metacyclogenesis and amastigote proliferation in T. cruzi, and that interfering with riboflavin transport and/or downstream metabolism may impair these processes.