Abstract
The Trichomonas vaginalis 120 kDa protein adhesin (AP120) is induced under iron-rich conditions and has sequence homology with pyruvate : ferredoxin oxidoreductase A (PFO A), a hydrogenosomal enzyme that is absent in humans. This homology raises the possibility that, like AP120, PFO might be localized to the parasite surface and participate in cytoadherence. Here, the cellular localization and function of PFO that was expressed under various iron concentrations was investigated using a polyclonal antibody generated against the 50 kDa recombinant C-terminal region of PFO A (anti-PFO50). In Western blot assays, this antibody recognized a 120 kDa protein band in total protein extracts, and proteins with affinity to the surface of HeLa cells from parasites grown under iron-rich conditions. In addition to localization that is typical of hydrogenosomal proteins, PFOs that were expressed under iron-rich conditions were found to localize at the surface. This localization was demonstrated using immunofluorescence and co-localization assays, as well as immunogold transmission electron microscopy. In addition to describing its enzyme activity, we describe a novel function in trichomonal host interaction for the PFO localized on the parasite surface. The anti-PFO50 antibody reduced the levels of T. vaginalis adherence to HeLa cell monolayers in a concentration-dependent manner. Thus, T. vaginalis PFO is an example of a surface-associated cell-binding protein that lacks enzyme activity and that is involved in cytoadherence. Additionally, PFO behaves like AP120 in parasites grown under iron-rich conditions. Therefore, these data suggest that AP120 and PFO A are encoded by the same gene, namely pfo a.
-
The GenBank/EMBL/DDBJ accession numbers for the T. vaginalis sequences discussed in this paper are: AY661465.1 GI: 56122435 (complete pfo a gene); AY652962 (pfo a amplicon); HQ657199 (pfo bII amplicon); HQ657200 (pfo c amplicon); HQ657201 (pfo d amplicon) and HQ657202 (pfo f amplicon).
-
A supplementary figure, showing cellular fractionation of T. vaginalis grown in 250 µM iron, and a supplementary table, showing primers used for PCR and RT-PCR assays, are available with the online version of this paper.
-
Edited by: L. Knoll
Introduction
Trichomonas vaginalis is the protozoan parasite that is responsible for human trichomoniasis (Schwebke & Burgess, 2004). Adherence of this mucosal parasite to human vaginal epithelial cells (VECs) is essential for the initiation and maintenance of infection (Alderete & Garza, 1985). Five surface proteins (AP120, AP65, AP51, AP33 and AP23) that interact with VECs and cervical cells have been characterized as trichomonad adhesins (Alderete & Garza; 1988; Arroyo et al., 1992; Moreno-Brito et al., 2005). Iron concentration and cellular contact positively regulate the levels of adherence and adhesin synthesis (Arroyo et al., 1993; Engbring & Alderete, 1998a, b; Garcia et al., 2003; Lehker et al., 1991). In addition, four of the five adhesins (AP120, AP65, AP51 and AP33) have sequence homology with metabolic enzymes, and the majority are positively regulated by iron at the levels of transcription and translation (Alderete et al., 1995, 1998, 2001; Arroyo et al., 1995; Engbring et al., 1996; Engbring & Alderete, 1998a, b; Lehker et al., 1991; Moreno-Brito et al., 2005). The expression of pyruvate : ferredoxin oxidoreductase (PFO) and other hydrogenosomal proteins is likewise positively regulated by iron (Gorrell, 1985; Vanácová et al., 2001).
A growing number of proteins have been found to perform divergent functions as a consequence of their localization and environmental conditions (Jeffery, 1999, 2003, 2005, 2009). This feature is typical of a group of functionally diverse metabolic enzymes that localize to the cell surface (Alderete et al., 2001; Collingridge et al., 2010). In Plasmodium falciparum and T. vaginalis, the glycolytic enzyme enolase, which is found in mammalian cells and parasites, additionally serves as a plasminogen receptor when it is localized at the cell surface (Mundodi et al., 2008; Pal-Bhowmick et al., 2007; Redlitz et al., 1995). In Plasmodium sp., enolase localization is a function of the parasite’s life cycle, whereas in trichomonads, its localization depends on iron concentration (Mundodi et al., 2008; Pal-Bhowmick et al., 2007). Similarly, the T. vaginalis glutaraldehyde-3-phosphate dehydrogenase (GAPDH) has been identified as a fibronectin-binding protein that is localized at the parasite surface. The expression and surface localization of GAPDH are positively regulated by iron (Alderete et al., 2001; Lama et al., 2009). Thus, T. vaginalis is part of a growing list of microbial pathogens that contain surface-associated enzymes that have alternate, non-enzymic functions. However, the mechanisms by which these enzymes are localized at the trichomonad surface and the pathways in which they act are poorly understood.
Another example of this type of enzyme is the T. vaginalis PFO A, which was originally identified as a hydrogenosomal enzyme that has sequence homology with the adhesin AP120 (Moreno-Brito et al., 2005). AP120 is expressed on the surface of parasites when they are grown under iron-rich conditions (Moreno-Brito et al., 2005), whereas PFO is a metabolic enzyme that is located on the membranes of hydrogenosomes and is responsible for the decarboxylation of pyruvate to acetyl-CoA (Hrdý & Müller, 1995). Sequence homology between these two proteins was previously demonstrated by peptide analysis using MS: 22 of the peptide masses from AP120 corresponded to peptide masses from PFO A. Homology between AP120 and PFO A was further verified by an immunodetection assay using a heterologous antibody to the Entamoeba histolytica PFO (EhPFO) (Moreno-Brito et al., 2005; Rodríguez et al., 1998). At the time of that study, the complete genome sequence of T. vaginalis was not available, and this work was not pursued further. It therefore remained unclear whether the AP120 adhesin is encoded by the pfo a gene. Interestingly, the draft of the T. vaginalis genome revealed other pfo-like genes with high sequence homology to pfo a, and these include pfo b, pfo c, pfo d, pfo e and pfo f (Carlton et al., 2007). These observations raise the following questions. (i) Does PFO localize to the surface of trichomonads and participate in adherence as the AP120 adhesin, and (ii) is PFO an enzyme with differential compartmentalization that is related to iron concentration?
Thus, we examined the relationship between AP120 and PFO. Specifically, we investigated whether these two proteins are encoded by the same ORF, the localization and function of PFO in the presence of iron, and the possibility of immuno-cross-reactivity between PFO and AP120. Here, we report that T. vaginalis PFO exhibits novel localization and functions as the AP120 adhesin. Furthermore, we report that PFO is positively regulated by iron and is involved in cytoadherence.
Methods
Parasite and HeLa cell cultures.
T. vaginalis parasites from the fresh clinical isolate CNCD 188 (Moreno-Brito et al., 2005) were cultured in trypticase-yeast extract-maltose (TYM) medium (Diamond, 1957) that was supplemented with 10 % heat-inactivated horse serum (HIHS) and incubated at 37 °C for at least 24 h and no longer than 2 weeks. TYM-HIHS medium contains iron at 20 µM (Gorrell, 1985; Alvarez-Sánchez et al., 2007). For parasites that were grown under iron-rich or iron-depleted conditions (250 or 0 µM iron, respectively), the culture medium was supplemented with 250 µM ferrous ammonium sulphate solution or 100 µM 2,2′-dipyridyl (Sigma), respectively, 24 h prior to inoculation with parasites (Alvarez-Sánchez et al., 2007). HeLa cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Invitrogen) that was supplemented with 10 % HIHS at 37 °C for 48 h in a 5 % CO2 atmosphere to obtain confluent cell monolayers (Arroyo et al., 1992).
RT-PCR assays.
Semiquantitative RT-PCR assays were performed using the SuperScript RNase H Reverse Transcriptase kit (Invitrogen). Total RNA from parasites (107) that were grown in 0, 20 or 250 µM iron was extracted using the TRIzol method and reverse-transcribed using reverse transcriptase and oligo(dT) primers as previously described (Moreno-Brito et al., 2005). Next, 10 % of the resulting cDNA was used as a template for the PCR. Gene-specific primers that were designed using the draft of the T. vaginalis genome (Carlton et al., 2007) were used for PCR amplification of each of the T. vaginalis pfo-like (pfo a, b, c, d, e and f) transcripts (Supplementary Table S1). A 112 bp fragment of the T. vaginalis β-tubulin transcript was amplified as an internal control (León-Sicairos et al., 2004).
As additional controls, a primer set that is specific for each pfo-like gene and primers for the entire pfo a ORF (3474 bp), including the ATG and TAA initiation and stop codons (Supplementary Table S1), were used to amplify fragments of the expected sizes from T. vaginalis CNCD 188 genomic DNA (gDNA). Each fragment was cloned into the pCRIITOPO vector (Invitrogen), and clones containing each pfo-like gene were selected and then sequenced by the dideoxy chain-termination method (Sanger et al., 1977) using an AB1377 Applied Biosystems automatic sequencer (UNAM, Institute of Cellular Physiology, Mexico). The DNA sequences were analysed using the blast (), ExPASy () and Workbench () search engine tools and deposited in GenBank under the accession numbers AY661465.1 GI: 56122435 (complete pfo a gene), AY652962 (pfo a amplicon; Moreno-Brito et al., 2005), HQ657199 (pfo bII amplicon), HQ657200 (pfo c amplicon), HQ657201 (pfo d amplicon) and HQ657202 (pfo f amplicon). Three independent RNA preparations that were obtained from separate parasite cultures were used for the semiquantitative RT-PCR assays and yielded identical results.
Expression and purification of a 1.1 kb fragment of the pfo a gene.
Using a plasmid containing the complete pfo a gene (pCRII TOPO-PFOA) and the primers that are listed in Supplementary Table S1, a 1.1 kb fragment encoding the C terminus of PFO A was amplified by PCR. The 1.1 kb amplicon was subcloned into the prokaryotic expression vector pColdI (Takara) according to the manufacturer’s instructions. The vector was transformed into the Escherichia coli BL21(DE3) strain, and positive clones (pPFO50) were induced for recombinant protein production by the addition of IPTG at a final concentration of 1 mM for 3 h at 37 °C. The bacteria were then harvested, and the pellet was suspended in Laemmli buffer and boiled for 3 min. The supernatants were separated by centrifugation at 15 000 g for 3 min at 4 °C and analysed by denaturing SDS-PAGE and Western blot (WB) analysis. The His-tagged recombinant PFO50 protein (~50 kDa) was purified by affinity chromatography using nickel-nitrilotriacetic acid (Ni-NTA) resin according to the manufacturer’s instructions (Invitrogen).
Ethics statement.
The protocols for antibody production in rabbits and mice were performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the Mexican Government NOM-062-ZOO-1999. The protocols were approved by the Internal Committee for the Care and Use of Laboratory Animals of the CINVESTAV-IPN (CICUAL) (permit number: 142-03). All bleeds were performed under xylazine hydrochloride local anaesthesia, except for the final bleed, which was performed under sodium pentobarbital anaesthesia. All efforts were made to minimize suffering.
Generation of antisera.
Affinity-purified recombinant hexokinase (HK) protein (Saavedra et al., 2005) (10.0 µg per animal) was used as an antigen to immunize 4 week-old male BALB/c mice by homogenizing with Freund’s complete adjuvant (Gibco) at a 1 : 1 ratio. The animals received four booster injections with the HK protein (10.0 µg per animal) in Freund’s incomplete adjuvant (Gibco) at 7 day intervals (Harlow & Lane, 1988). The resulting antiserum was used in WB assays. Pre-immune (PI) mouse serum was obtained prior to the immunization schedule and served as a negative control.
The generation of the anti-AP65, anti-PFO50 and anti-TvCP4 antibodies in rabbits was performed using the following antigens: 100 µg of a HeLa cell-bound AP65 adhesin band, as previously described (Arroyo et al., 1992); 300 µg of affinity-purified recombinant PFO50 protein; or the previously obtained His-tagged recombinant TvCP4 protein (Ramón-Luing et al., 2010). Similar schedules were used for antibody generation in rabbits and mice, except that the rabbits received injections at 15 day intervals (Harlow & Lane, 1988). Immune serum was obtained 7 days after the final immunization. The antisera were used in WB analysis, indirect immunofluorescence, co-localization, immunogold labelling, and cytoadherence inhibition assays. PI serum was obtained from each rabbit prior to the immunization schedule and was employed as a negative control in the assays using rabbit sera. For the cytoadherence inhibition assays, IgG fractions were obtained using the caprylic acid method (Harlow & Lane, 1988).
WB analysis.
After electrophoresis, recombinant PFO50 protein or total protein extracts from T. vaginalis that were grown in 20 or 250 µM iron were transferred onto nitrocellulose (NC) membranes and blocked using 5 % skimmed milk in PBS/0.05 % Tween 20 (PBS-T) buffer for 18 h at 4 °C. The NC membranes containing the recombinant PFO50 protein were incubated with anti-PFO50 (1 : 80 000) and anti-TvAP120 (1 : 100) (Moreno-Brito et al., 2005) antibodies for 18 h at 4 °C. The membranes containing total protein extracts were incubated with anti-PFO50 (1 : 7000) and anti-AP120 (1 : 20 000) antibodies, or PI rabbit serum (1 : 100). The NC membranes were then washed five times with PBS-T, incubated with an IgG peroxidase-conjugated secondary antibody (Bio-Rad) (1 : 3000) for 2 h at 37 °C, and developed with 4-chloro-1-naphthol (Bio-Rad). The respective PI rabbit and mouse sera were used as negative controls. Anti-α-tubulin (Zymed) or anti-AP65 adhesin antibodies at their appropriate concentrations were used as loading controls for the WB assays of total parasite extracts or proteins with affinity to the surface of HeLa cells, respectively. These experiments were performed at least three independent times with similar results.
Indirect immunofluorescence assays.
Parasites that were grown in 20 or 250 µM iron media were attached to coverslips and fixed with 4 % paraformaldehyde for 1 h at 37 °C, washed with PBS, permeabilized with 0.2 % Triton X-100 for 15 min at room temperature, washed with PBS, and then blocked with 0.2 M glycine for 1 h at 37 °C and 0.2 % fetal bovine serum for 15 min at 37 °C. The fixed trichomonads were incubated with the anti-PFO50 antibody (1 : 200) or PI rabbit serum (1 : 100) for 18 h at 4 °C, followed by washing with PBS. The parasites were then incubated with an IgG–FITC-conjugated anti-rabbit secondary antibody (Pierce, 1 : 100) for 1 h at room temperature, washed, counterstained with propidium iodide (10 µg ml−1), washed with PBS, and mounted with Vectashield mounting solution (Vector). The parasites were analysed by confocal microscopy (Leica, Microsystems).
For the surface co-localization assays, fixed but non-permeabilized trichomonads that had been grown in iron-rich conditions were blocked as described above, incubated with the anti-PFO50 antibody (1 : 200) for 18 h at 4 °C, and then washed with PBS. The parasites were then incubated with an IgG–FITC-conjugated anti-rabbit secondary antibody (Pierce, 1 : 100) for 1 h at room temperature, washed, and counterstained with Dil (Dil-CM-38; Molecular Probes, 1 : 2000) for 30 min and DAPI (0.5 µg ml−1) for 15 min at room temperature. Finally, the parasites were washed with PBS and mounted with Vectashield mounting solution (Vector). The samples were analysed by confocal microscopy (Leica, Microsystems). These experiments were performed at least three independent times with similar results.
Sample preparation for immunogold localization and transmission electron microscopy (TEM).
The parasites were fixed with 4 % paraformaldehyde/0.1 % glutaraldehyde in serum-free TYM medium for 1 h at 25 °C, embedded in LR White, and then polymerized under UV at 4 °C overnight. Thin (60 nm) sections were obtained, mounted on Formvar-covered nickel grids, and then incubated with the anti-PFO50 antibody (1 : 200) for 1 h at 25 °C and the 15 nm gold-conjugated secondary antibody (Zymed, 1 : 50) for 1 h at 25 °C. The sections were examined using a transmission electron microscope (Morgagni 268 D, Philips).
Hydrogenosome purification.
Parasites (2×108) were harvested at 3250 g for 5 min, washed twice in PBS (pH 7.0), suspended in 20 ml of SMDI buffer (250 mM sucrose, 10 mM DTT, 10 mM MOPS, pH 7.2, 50 µg TLCK ml−1, 10 µg leupeptin ml−1), and then homogenized in a glass vessel with a teflon pestle for 15 min on ice. The suspension was centrifuged at 1000 g for 10 min at 4 °C. The pellet (PP1) was discarded, and the supernatant (SN1) was centrifuged at 12 180 g for 10 min at 4 °C in a Sorval RC-5B centrifuge (Waltham). The supernatant (SN2), which contained the membrane and cytosol fractions, was saved for further processing, and the pellet (PP2), which contained the hydrogenosomes, was resuspended in 5 ml SMDI buffer and mixed with an equal volume of Percoll (Amersham) that had been diluted (90 %, v/v) with SMDI buffer. Both the SN2 and PP2 samples were centrifuged at 153 000 g for 45 min at 4 °C in a Sorvall 100SE centrifuge. A white band that contained the purified hydrogenosomes was recovered from the PP2 sample. The supernatant from sample SN2 corresponded to the cytosol fraction, and the membrane fraction was recovered from the pellet (Opperdoes et al., 1984). Samples of each fraction were analysed by SDS-PAGE and WB using the appropriate antibodies; in addition, the samples were processed for TEM and were used to determine enzyme activity. These experiments were performed at least three independent times with similar results.
Enzyme activities.
The enzyme activities of PFO and HK (accession nos AF248652 and AF248653; Wu et al., 2001) were determined for each cellular fraction as previously described (Ramos-Martínez et al., 2009; Saavedra et al., 2007). PFO activity in the trichomonad Triton-extracted fraction was determined under an N2 atmosphere with an assay using the following: 100 mM Na2HPO4 buffer, pH 7.4 (previously purged with N2); 0.25 mM nitro blue tetrazolium (NBT), used as the electron acceptor; 20–60 µg of protein of the various trichomonad cellular fractions; and 10 mM pyruvate. The reaction was initiated by the addition of 0.1 mM CoA. NBT reduction was monitored at 560 or 604 nm for PFO and HK, respectively, using a spectrophotometer (Shimadzu). The absorbance baseline in the absence of one of the substrates was always subtracted. Care was taken to ensure that the activity was linearly dependent on the protein content of the sample (Pineda et al., 2010). The hydrogenase (HYD) activity (accession no. AAC47160.1; Bui & Johnson, 1996) of 20–60 µg of protein from each fraction was determined in 100 mM Na2HPO4 buffer (pH 7.4) that had been previously saturated with N2. The reaction was initiated by the addition of 0.25 mM NBT, and an increase in absorbance, which correlates with NBT reduction, was monitored at 560 nm using a spectrophotometer (Shimadzu). The basal NBT-reducing activity, which was determined from a control reaction prepared with N2-saturated buffer, was subtracted from each cellular sample. Each sample was analysed in triplicate, and the experiment was performed at least two independent times with similar results.
Cell-binding assays.
WB assays were performed on proteins that were eluted using standard cell-binding assays (Moreno-Brito et al., 2005) from total protein extracts of parasites that were grown in 20 or 250 µM iron media and glutaraldehyde-fixed HeLa cells, using the following antibodies: anti-PFO50 (1 : 40 000), anti-AP65 (1 : 8000) (positive control) and PI rabbit serum (1 : 100) (negative control). The membranes were developed by chemiluminescence. A cell-binding assay using the recombinant PFO50 protein was performed with a slight modification of the standard method (Solano-González et al., 2006). In brief, glutaraldehyde-fixed HeLa cells (2×106) were incubated with either 50 µg of the recombinant PFO50 protein or BSA as a control for specificity. Following incubation, the cells were washed seven times with PBS-T to remove any loosely associated proteins, and then boiled for 3 min in Laemmli buffer to elute the epithelial cell-binding proteins. These proteins were then analysed by SDS-PAGE on 10 % polyacrylamide gels that were stained with Coomassie brilliant blue. These experiments were performed at least three independent times with similar results.
Cytoadherence assay.
The cytoadherence assay was performed over confluent live HeLa cell monolayers on coverslips, using an adaptation of reported methods (Alderete & Garza, 1985, 1988; Bastida-Corcuera et al., 2005). Cell monolayers were seeded on 12 mm coverslips in 24-well plates at 5×105 cells per well in culture medium, and grown to confluence at 37 °C and 5 % CO2 for 2 days. The cells were washed once with PBS before the addition of living parasites (106 per well) that had previously been labelled with 25 µM CellTracker Blue CMAC (Molecular Probes) for 30 min at 37 °C in TYM medium according to the manufacturer’s instructions. For the inhibition assays, the labelled parasites were incubated for 30 min at 4 °C with 50, 100 or 150 µg ml−1 anti-PFO50 IgGs before interaction with the HeLa cell monolayers (Moreno-Brito et al., 2005). The same concentrations of the IgG fractions of anti-TvCP4 antibody or PI rabbit serum were used as negative controls. The parasites (106 per well) were washed and added to the HeLa cells in 1 ml (2 : 1) serum-free DMEM-TYM medium and incubated at 37 °C for 30 min and 5 % CO2. The coverslips were subsequently washed in PBS, fixed with 4 % paraformaldehyde, and mounted on slides. Each condition was performed in triplicate, and eight ×20 magnification fields were analysed per coverslip. Fluorescent parasites that had adhered to host cells were counted using an Eclipse 80i epifluorescence microscope (Nikon) and the NIS-Elements BR 2.1 software (Nikon). The experiment was repeated at least three independent times with similar results.
Statistical analysis.
All summary data are presented as the mean±sem of triplicate samples. A statistically significant difference between means was determined by analysis of variance (ANOVA) (P) using Microsoft Excel 2007 software.
Results
Induced expression of pfo-like genes under iron-rich conditions
To investigate which of the pfo-like genes that are present in the T. vaginalis genome are expressed and positively regulated by iron (as is the case with AP120 adhesin), semiquantitative RT-PCR assays using specific primers for each pfo-like gene (Supplementary Table S1) were performed on RNA that was isolated from trichomonads that were grown in the presence of 0, 20 or 250 µM iron. Although each of the expected pfo-like gene fragments, except for the pfo e gene (Fig. 1, lane 2), was amplified from the gDNA, only pfo a and pfo bII were transcribed under the iron concentrations that were tested. The pfo a transcript was observed primarily in parasites that were grown in 250 µM iron (Fig. 1a, lane 5), whereas the pfo bII gene was constitutively expressed at all iron concentrations that were tested (Fig. 1b, lanes 3–5). The level of β-tubulin transcripts in the same RNA samples at each iron concentration was used to normalize the transcript quantities (Fig. 1f), and an RT-PCR that lacked the reverse transcriptase enzyme was used to exclude genomic contamination (Fig. 1g, lanes 3–5). Additionally, the identity of each amplicon was verified by DNA sequencing, and these sequences were deposited in GenBank. Therefore, in this trichomonad isolate, not all of the pfo-like genes in the trichomonad genome are expressed in the presence of iron. Indeed, only pfo a and pfo bII are expressed. For convenience, the product of the pfo bII gene will be hereafter referred to as PFO B.
Effect of iron on the transcription of T. vaginalis pfo-like genes. (a–e) RT-PCR with specific primers for each pfo gene (pfo a, pfo b, pfo c, pfo d and pfo f; see Supplementary Table S1) using cDNA from parasites grown under iron-depleted (0 µM; lane 3), normal (20 µM; lane 4) and iron-rich (250 µM; lane 5) conditions. (f) RT-PCR with specific primers for the β-tubulin gene using the same cDNA (a–e) as an internal control. (g) Reverse transcriptase minus [RT (−)] reaction used to verify the lack of gDNA contamination in the cDNA samples (lanes 3–5). (a–g) PCR with the same pairs of primers for each pfo gene using gDNA as positive controls [+; lane 2, except in (g), where no DNA was added]. PCRs without gDNA were used as negative controls (–; lane 1). The amplicons were separated on ethidium bromide-stained 1 % agarose gels. The sizes of the amplicons are given in bp. The RT-PCR product in (b) corresponds to the pfo bII isoform.
Cross-reactivity between expressed PFO and the AP120 protein
To check for cross-reactivity between the expressed PFO genes and the AP120 adhesin, a 1.1 kb pfo a gene fragment that encodes the C-terminal region of the 120 kDa PFO A protein (amino acid residues 759–1157) was cloned and expressed in a bacterial system (Fig. 2a). This PFO A fragment, which is common to PFO A and PFO B, contains 7 of the 22 peptides in the AP120 adhesin as identified by MS (Moreno-Brito et al., 2005). This recombinant protein of ~50 kDa, hereafter termed PFO50, was then overexpressed in E. coli, purified by Ni-affinity chromatography (Fig. 2a), and used for polyclonal antibody production in rabbits (anti-PFO50). The affinity-purified PFO50 recombinant protein was transferred onto NC membranes and was recognized by the polyclonal anti-PFO50 antibody (Fig. 2c, lane 1) and by the anti-TvAP120 antibody (Moreno-Brito et al., 2005) in WB assays (Fig. 2d, lane 2). Likewise, the anti-PFO50 antibody reacted with a 120 kDa protein band in total protein extracts from parasites that were grown in media containing 20 or 250 µM iron (Fig. 2b, lanes 1 and 2; Fig. 2c, lanes 2 and 3) and reacted with higher intensity under iron-rich conditions. The positive and negative controls reacted as expected (Fig. 2b, d). These data demonstrate that there is cross-reactivity between anti-PFO50 and anti-TvAP120 antibodies in recognizing a 120 kDa protein, suggesting that iron-induced PFOs and the AP120 adhesin share sequence homology in their C termini. These data also demonstrate that the anti-PFO50 antibody recognizes PFOs that are expressed in parasites that are grown under 20 and 250 µM iron conditions.
Cross-reactivity of the anti-PFO50 and anti-TvAP120 antibodies. (a) SDS-PAGE on 12 % polyacrylamide gels of bacterial extracts harbouring the pColdI-PFO50 plasmid before (lane 2) or after IPTG induction (lane 1). Affinity-purified recombinant PFO50 protein (lane 3, PFO50). Gels were stained with Coomassie brilliant blue (CBB). (b) Total protein extracts from parasites (TvEx) grown in 250 µM (lane 1) or 20 µM (lane 2) iron, and WB using the anti-α-tubulin monoclonal antibody of duplicate gels transferred to NC membranes. The asterisk indicates the α-tubulin (50 kDa) protein band as a loading control (lanes 3 and 4). (c) WB of the recombinant PFO50 protein (lane 1) and total extracts from parasites grown in 250 µM (lane 2) or 20 µM (lane 3) iron using the anti-PFO50 antibody (lanes 1–3). (d) WB of the recombinant PFO50 protein (lanes 1 and 2) and total extracts from parasites grown in 250 µM iron (lane 3) using the control PI rabbit serum (lane 1) or the anti-TvAP120 antibody (lanes 2 and 3). The arrowheads show the position of the 120 kDa protein band (b–d) or the PFO50 recombinant protein band (a, c and d).
Localization of the expressed PFOs in iron-rich T. vaginalis parasites
To investigate whether iron-induced PFOs localize to the parasite hydrogenosome (the typical site of trichomonad PFO localization), immunofluorescence and immunogold-labelling assays using the anti-PFO50 antibody were performed on fixed, permeabilized whole parasites (Fig. 3a, b) and thin sections (Fig. 3c, d) of parasites that were grown in 20 or 250 µM iron media. Confocal microscopy revealed that this antibody, visible as rings around cytoplasmic organelles that appeared to be hydrogenosomes (Fig. 3a), reacted more strongly in iron-rich parasites than in iron-restricted parasites (Fig. 3b). This was confirmed with immunogold-labelling TEM. The anti-PFO50 antibody reacted primarily with hydrogenosomal proteins in parasites that were grown under either iron condition (Fig. 3c, d); however, the gold particles were more abundant (up to five times more) in iron-rich parasites (15–22 gold particles per hydrogenosome, Fig. 3c) than iron-restricted parasites (one to four gold particles per hydrogenosome, Fig. 3d). These data confirm the predicted localization of PFOs in trichomonads. As expected, the PI rabbit serum that was used as a negative control yielded no reaction (Fig. 3e).
Typical localization of T. vaginalis PFO in hydrogenosomes. (a, b) Parasites grown in 250 or 20 µM iron, fixed, permeabilized and incubated with the anti-PFO50 antibody, followed by a FITC-coupled secondary antibody (in green), and counterstained with propidium iodide (in red). (a) and (b) are representative of 100 cells. Fluorescence was observed in more than 80 % of cells; bar, 25 µm. (c, d) TEM immunogold localization of PFO in thin sections from parasites grown in 250 µM (c) or 20 µM (d) iron incubated with the anti-PFO50 antibody, followed by a gold-coupled secondary antibody. The arrows point to gold particles in the hydrogenosomes (H) and vacuoles (V). (e) Parasite thin sections incubated with PI rabbit serum as a negative control. (c) and (d) are representative of 50 cells. Gold particles were observed in more than 90 % of cells. Bars, 0.22 µm (c, d) or 0.071 µm (e).
Co-localization of PFO on the surface of T. vaginalis
Based on the high sequence homology between AP120 adhesin and PFO A, we co-performed localization experiments on fixed, non-permeabilized parasites to examine whether the expressed PFOs were localized on the parasite surface under iron-rich conditions. Confocal microscopy revealed that lipids (which were stained with DiI-CM-38; Fig. 4c, red) and PFO (which was immunolocalized using the anti-PFO50 antibody; Fig. 4b, green) co-localized at the surface of T. vaginalis (Fig. 4d, yellow). Parasite nuclei were also observed (stained with DAPI; Fig. 4d, blue). Nomarski microscopy showed the integrity of the parasites (Fig. 4a). Either slight surface fluorescence or no immunoreactivity was observed using the anti-PFO antibody or control PI serum, respectively, in non-permeabilized parasites that were grown under iron-restricted or iron-rich conditions (data not shown).
Co-localization of PFO at the T. vaginalis cell surface under iron-rich conditions. (a–d) Parasites grown in 250 µM iron, fixed, non-permeabilized, incubated with the anti-PFO antibody, followed by a FITC-coupled secondary antibody (b, green). (a) Nomarski microscopy. The membrane lipids and nuclei were stained with Dil (c, red) and DAPI (d, blue), respectively. Co-localization of the surface molecules is shown in yellow (d). Bar, 7 µm.
In immunogold-labelling TEM assays, the anti-PFO50 antibody also detected gold particles in vacuoles that contained complete or partially degraded hydrogenosomes that were close to the parasite surface and on the surface of parasites that were grown in iron-rich conditions (Fig. 5a–d). However, more gold particles were observed in free hydrogenosomes (52 %) than in hydrogenosomes in vacuoles (18 %), in vacuoles (11 %) or on the parasite surface (19 %) (Fig. 5). Taken together, these data reveal the differential compartmentalization of trichomonad iron-induced PFOs into hydrogenosomes, vacuoles and plasma membranes, and this may be related to the presence or absence of iron. Thus, these results suggest that PFO is a novel hydrogenosomal metabolic enzyme that is localized to the trichomonad surface and may serve an additional function in the presence of iron.
Immunogold localization of PFO by TEM in distinct cellular compartments. PFO hydrogenosomal (a), vacuolar (b, c) and membrane (d) localization detected with the anti-PFO50 antibody in thin sections of iron-rich parasites using an immunogold TEM assay as in Fig. 3. The redistribution of PFO from hydrogenosomes (a) to the parasite membrane (d) could be occurring via an autophagy-like pathway (a–d). Arrowheads point to gold particles. H, hydrogenosome; V, vacuole. Bars: (a, b and d), 0.125 µm; (c), 0.22 µm.
Distribution of PFO enzyme activity
Cell fractionation coupled with TEM analysis was performed to determine the enzyme activity of the iron-induced PFOs in the various cellular compartments of iron-rich parasites. After cell fractionation, PFO enzyme activity was detected primarily (80 %) in the hydrogenosomal fraction, and low levels of PFO enzyme activity were detected in the membrane (~4 %) and cytosol (~1 %) fractions (Fig. 6a). The enzyme activities of HYD and HK, which served as controls, were as expected in the following fractions: HYD was in the hydrogenosomes and HK was in the cytosol (Fig. 6a). The purity of the cellular fractions was verified by TEM (Supplementary Fig. S1) and WB assay (Fig. 6b) using the anti-PFO50 and anti-HK antibodies or control PI serum. These data show that PFO enzymic activity occurs in hydrogenosomes, as we expected. However some questions still remained. For example, what role does PFO play on the parasite’s surface, and does PFO participate in the host–parasite interaction, as is the case for other metabolic enzymes?
PFO enzyme activity of parasites grown under iron-rich conditions. (a) Enzyme activity of PFO in the hydrogenosome, membrane and cytosol. HK and HYD enzyme activities were used as markers for the cytosol and hydrogenosome fractions, respectively. Error bars, sem of two experiments performed in triplicate. The total enzyme activities in the homogenates were 5.6 U (PFO), 0.49 U (HYD) and 13.2 U (HK), and these values were normalized to 100 %. A statistically significant difference was found for PFO enzyme activity in the membrane and cytoplasm fractions (P = 0.050) (marked with asterisks) compared with the hydrogenosome fraction. (b) WB of the same samples as in (a) separated by SDS-PAGE on 7 % polyacrylamide gels, transferred to NC membranes, and incubated with anti-PFO50 (1 : 7000) and anti-HK (1 : 1000) antibodies, or control PI rabbit serum (1 : 1000) as a negative control. Total extracts (T), hydrogenosome (H), cytosol (C) and membrane (M) fractions. A Coomassie brilliant blue-stained gel (CBB) from the 120 kDa region is also shown.
PFO binds to the surface of HeLa cells
To determine whether PFO participates in host–parasite interactions, HeLa cell binding assays were performed as previously described (Arroyo et al., 1992) (Fig. 7a, b). Trichomonad or recombinant proteins that were eluted from the surface of the HeLa cells were analysed by SDS-PAGE directly or were transferred onto NC membranes for WB analysis using the anti-PFO50 antibody. The antibody reacted with a 120 kDa protein with affinity for the surface of HeLa cells from parasites that were grown in 250 µM iron (Fig. 7a, lane 1) but not in 20 µM iron (Fig. 7a, lane 2). The anti-AP65 antibody and the PI serum, which served respectively as positive and negative controls, yielded the expected results (Fig. 7a). Likewise, recombinant PFO50 protein was bound to the surface of the HeLa cells (Fig. 7b, lanes 4 and 5) but did not bind BSA, which was used as a non-related protein to test for specificity (Fig. 7b, lanes 2 and 3). These data suggest that the C-terminal PFO fragment may contain a cell-binding domain through which PFO mediates host–parasite interactions.
PFO interacts with the surface of HeLa cells (a, b) and participates in trichomonal adherence (c, d). (a) WB analysis of proteins eluted after a cell-binding assay using fixed HeLa cells and total extracts from parasites grown in 250 µM (lane 1) or 20 µM (lane 2) iron and probed with the anti-PFO50 antibody or the anti-AP65 antibody and PI serum as positive and negative controls, respectively. Arrowheads indicate the 120 kDa or the AP65 protein band (65). (b) Proteins eluted after a cell-binding assay visualized on Coomassie brilliant blue-stained 10 % polyacrylamide gels using the recombinant PFO50 protein (lane 5), fixed HeLa cells as a mock control (lane 1) or BSA for specificity of the interaction (lane 3). BSA and PFO50 were directly loaded as additional controls (lanes 2 and 4). (c) Fluorescence microscopy of a representative cytoadherence inhibition assay (over live HeLa cell monolayers) of fluorescently labelled parasites pre-treated with 100 µg ml−1 IgG from the anti-PFO50 serum (panels iii and iv) or PI serum as a representative negative control (panels i and ii); bar, 78 µm. (d) Data from inhibition experiments, showing the percentage cytoadherence of T. vaginalis to HeLa cell monolayers in the presence of 50, 100 or 150 µg ml−1 IgG from the anti-PFO50 serum (black bars) or from the anti-TvCP4 serum as a non-related control antibody (grey bars). The percentage of parasite cytoadherence (white bar) was obtained from the number of adhered fluorescent parasites (~1210 per coverslip) of control samples without treatment, which was normalized to 100 % adherence for comparative purposes. Bars represent the mean and sem of triplicate samples from a representative experiment. Statistically significant differences were found for the data for 100 and 150 µg ml−1 IgG from the anti-PFO serum (P = 0.050) (marked with asterisks) compared with IgGs from control anti-TvCP4 serum.
PFO participates in adherence
To test for alternative, non-enzymic roles that surface-associated PFO might play in host–parasite interactions, we performed cytoadherence inhibition assays. Fluorescently-labelled parasites that were grown in 250 µM iron were incubated in increasing concentrations of anti-PFO50 IgG in preparation for cytoadherence assays using live HeLa cell monolayers (Fig. 7c, d). Anti-PFO50 IgG reduced the levels of T. vaginalis adherence by up to ~45 % in a concentration-dependent manner (Fig. 7c, panels iii and iv; Fig. 7d), whereas no such effect was observed using control IgG from PI serum (data not shown) or from antiserum that was raised against the trichomonad surface protein TvCP4 (up to ~11 % reduction) (Solano-González et al., 2007), which does not participate in adherence (Fig. 7c, panels i and ii; Fig. 7d). Thus, our data show that, under iron-rich conditions, surface-associated PFO appears to play a role in trichomonal cytoadherence, thereby suggesting that PFO may be the AP120 adhesin.
Discussion
PFO is a metabolic enzyme localized to the membrane of T. vaginalis hydrogenosomes (Hrdý & Müller, 1995), and catalyses the oxidative decarboxylation of pyruvate in a CoA-dependent reaction to yield acetyl-CoA and CO2 (Williams et al., 1987). This enzyme is present in protozoan parasites but is absent in humans (Upcroft et al., 2006). Its expression in T. vaginalis is positively regulated by iron (Gorrell, 1985; Vanácová et al., 2001). Interestingly, MS analysis after tryptic digestion of the AP120 adhesin, which localizes to the parasite surface under iron-rich conditions, revealed sequence homology with PFO A (Moreno-Brito et al., 2005), suggesting a possible multifunctional nature of the T. vaginalis PFO. This finding was not surprising, given that the other adhesins in T. vaginalis (Alderete et al., 2001; Arroyo et al., 1992, 1995; Engbring et al., 1996) are classified as multifunctional proteins (Collingridge et al., 2010).
Thus, to explore the possibility of a second function for PFO in trichomonads, we addressed this homology from the perspective of PFO. First, we examined the effect of iron on the expression of pfo-like genes in T. vaginalis (Upcroft et al., 2006). Only two of the seven pfo-like genes were expressed in the presence of iron (Fig. 1), suggesting that different environmental conditions are required for the expression of the other pfo-like genes, which is consistent with an analysis of pfo cDNAs and expressed sequence tag (ESTs) (Carlton et al., 2007; Trichdb database). We were unable to amplify the pfo e gene from gDNA or cDNA of the CNCD 188 trichomonad isolate that was used in this study, despite using several combinations of primers that were synthesized based on the pfo e sequence that was reported for the T. vaginalis genome of the G3 isolate (Carlton et al., 2007). This result suggests that there are differences in the pfo genes between trichomonad isolates, as occurs in the vast bspa-like gene family (Noël et al., 2010). However, we cannot exclude the possibility that pfo e is a pseudogene, as pfo a, pfo b and pfo e are >80 % identical at the amino acid level. It is interesting to note that the PFO-encoding genes belong to a multigene family that includes other metabolic enzymes with adhesin functions (Alderete et al., 1995; Arroyo et al., 1995; Engbring & Alderete, 1998a, b; Garcia et al., 2003) and whose expression appears to be differentially regulated by iron (Alderete et al., 1998).
We next determined the cellular localization of the iron-induced PFOs. Our data indicate that in trichomonads, these PFOs are differentially compartmentalized in hydrogenosomes, in vacuoles, and at the cell surface (Figs 3–5). These results are consistent with previous reports of protozoan PFOs (Hrdý & Müller, 1995), and with the two-compartment localization (i.e. cytosol and surface) of Entamoeba histolytica PFO (EhPFO; Rodríguez et al., 1998; Thammapalerd et al., 1996). Moreover, differential compartmentalization of PFO in the presence of iron has been observed for other trichomonad adhesins (e.g. AP65) that have high sequence homology with hydrogenosomal enzymes (Alderete et al., 1995; Arroyo et al., 1995; Engbring & Alderete, 1998a, b; Garcia et al., 2003) and for other trichomonad surface-associated enzymes with no enzyme activity, such as enolase and GAPDH (Alderete et al., 2001; Lama et al., 2009; Mundodi et al., 2008). This partial redistribution of metabolic enzymes from the cytosol to the cell periphery has also been reported in Toxoplasma gondii tachyzoites (Pomel et al., 2008) and in P. falciparum, in which enolase and aldolase are two of the most clearly delineated examples of enzyme moonlighting in parasites (Collingridge et al., 2010).
The hydrogenosomal localization of iron-induced PFOs in T. vaginalis was both expected and consistent with previous studies (Bradley et al., 1997; Dyall et al., 2000; Mentel et al., 2008; Williams et al., 1987). The pfo genes encode a signal peptide that is present in the majority of hydrogenosomal enzymes (Bradley et al., 1997; Dyall et al., 2000), which are synthesized in free polyribosomes (Lahti & Johnson, 1991) and may be involved in guiding expressed PFOs to hydrogenosomes under normal and iron-rich growth conditions (Mentel et al., 2008).
Interestingly, our immunogold TEM data are similar to the observation in Tritrichomonas foetus (Benchimol, 1999), suggesting that the relocalization of PFO and other hydrogenosomal enzymes to the parasite surface could be done by a hydrogenosome autophagy-like pathway (Fig. 5), allowing these enzymes to participate in host–parasite interactions. The presence of an autophagy-like pathway in T. vaginalis was first suggested by Hirt et al. (2007) and recently supported by the identification in the T. vaginalis genome of the genes encoding several of the conserved ATG proteins involved in autophagy (Brennand et al., 2011; Carlton et al., 2010). Thus, it is possible that high iron, as another stress signal triggering autophagy, might lead to a rapid intracellular remodelling of organelles and cytoskeletal architecture to initiate organelle turnover within minutes of the cell receiving the appropriate cues (Arroyo et al., 1993; Herman et al., 2008; Lehker et al., 1991; Collingridge et al., 2010). Future work will explore this hypothesis further.
Having established that distinct PFO compartmentalization occurs, we next examined its enzyme activity in the hydrogenosome, cytosol and membrane cellular fractions. As a metabolic enzyme, PFO exhibited enzymic activity in hydrogenosomes. However, no significant enzyme activity was detected for PFO that was localized to the cytosol or at the surface of the parasite (Fig. 6a, b), despite the controls, which ensured that the same protein concentration was measured in each fraction and which measured the levels of PFO and HK in each fraction (Fig. 6b). This decreased PFO activity could be due to inappropriate environmental redox conditions, which might inhibit the activity of the enzyme (Williams et al., 1987) when PFO is located at the parasite surface. Indeed, EhPFO enzyme activity is highly sensitive to the presence of oxygen (Ramos-Martínez et al., 2009; Pineda et al., 2010). However, we cannot exclude the possibility that the membrane extraction procedure itself affected PFO enzyme activity. Thus, the functional diversity of trichomonad PFO might be associated with its distinct cellular compartmentalization, which is influenced by changes in iron concentration and the redox state in each compartment. Such a model is consistent with the iron-induced differential compartmentalization and function of the AP65 adhesin (Garcia et al., 2003).
Given that surface-associated PFO lacks enzyme activity, its possible role in host–parasite interaction was explored by cell-binding and cytoadherence inhibition assays using live cells (Fig. 7). Our results support a novel role for the surface-localized PFO in trichomonal adherence. An antibody that was generated against a recombinant C-terminal region of PFO specifically inhibited trichomonal adherence to HeLa cell monolayers in a concentration-dependent manner, reaching saturation at three times the concentration of IgG from the rabbit anti-PFO50 antibody that was used. This is consistent with a multifactorial process involving distinct molecules. Five adhesins, two cysteine proteinases, and lipophosphoglycan have been described as involved in trichomonal adherence (Arroyo & Alderete, 1989, 1995; Arroyo et al., 1992; Bastida-Corcuera et al., 2005; Hernández et al., 2004; Mendoza-López et al., 2000; Moreno-Brito et al., 2005). Although the AP120 adhesin represents a minority protein fraction, it contributes to nearly half of adherence under iron-rich conditions, and this behaviour is similar to that of other adhesins (Arroyo et al., 1992). Taken together, these data indicate that PFO is another example of a metabolic enzyme that is typically found in hydrogenosomes and that can additionally localize to the parasite surface, where it has a novel function in trichomonal adherence as the AP120 adhesin. Thus, PFO can be added to the growing list of moonlighting proteins in parasitic protists such as T. vaginalis.
Many of the known moonlighting proteins are highly conserved enzymes and are also known to be ancient enzymes. In particular, moonlighting occurs in enzymes that are involved in sugar metabolism. It has been suggested that as many as seven of the 10 proteins in the glycolytic pathway and seven of the eight enzymes in the tricarboxylic acid cycle also have a moonlighting function (Huberts & van der Klei, 2010; Pomel et al., 2008). The rate at which multifunctional or moonlighting proteins are being identified is increasing, particularly with respect to the group of functionally diverse metabolic enzymes that additionally localize to the cell surface (Alderete et al., 2001; Collingridge et al., 2010), where they have previously unsuspected or unappreciated ‘second jobs’ (Jeffery, 1999, 2003, 2005, 2009). The multifunctionality of trichomonad PFOs that are expressed under iron-rich conditions is similar to that of the enolase in mammalian cells and parasites (Mundodi et al., 2008; Pal-Bhowmick et al., 2007; Redlitz et al., 1995). Thus, the data that we obtained in this study of PFO are in agreement with the following statement by Collingridge et al. (2010): ‘...in the post-genomic analysis of gene function for some, maybe many, metabolic enzymes their cellular functions extend beyond classic roles in intermediary metabolism’.
In conclusion, in this study, we demonstrate that, in the presence of iron, T. vaginalis PFOs behave as moonlighting proteins. PFOs have enzyme activity when localized to the hydrogenosomes, and appear to play a role in the host–trichomonal interaction as the AP120 adhesin when localized to the parasite surface, and this is likely through an autophagy-like pathway.
Acknowledgements
This work was partially supported by the Department of Infectomics and Molecular Pathogenesis at CINVESTAV-IPN, as well as grants 33044 M, 58611 and 68949 from CONACYT-México (to R. A.) and from the Instituto de Ciencia y Tecnología del D.F. (ICyT-299), México (to R. A.). P. M.-C. and R. E. C.-G. were scholarship recipients from CONACYT-México. We are grateful to Dr Ricardo Jasso-Chávez of the Instituto Nacional de Cardiología for his technical assistance in cell fractionation, Ivan J. Galván-Mendoza for his technical assistance in the confocal microscopy analysis, MVZ Manuel Flores Cano of the ‘Unidad de Producción y Experimentación de Animales de Laboratorio (UPEAL)’ at CINVESTAV-IPN for his help with rabbit handling for antibody production, Q. F. B Leticia Ávila-González for her technical assistance, Mar Sarai Hernández-García for TEM gold particle quantification, and Martha Aguilar for her secretarial support. We also thank the reviewers for constructive comments.