Comprehensive characterization of purine and pyrimidine transport activities in Trichomonas vaginalis and functional cloning of a trichomonad nucleoside transporter

Trichomoniasis is a common and widespread sexually‐transmitted infection, caused by the protozoan parasite Trichomonas vaginalis. T. vaginalis lacks the biosynthetic pathways for purines and pyrimidines, making nucleoside metabolism a drug target. Here we report the first comprehensive investigation into purine and pyrimidine uptake by T. vaginalis. Multiple carriers were identified and characterized with regard to substrate selectivity and affinity. For nucleobases, a high‐affinity adenine transporter, a possible guanine transporter and a low affinity uracil transporter were found. Nucleoside transporters included two high affinity adenosine/guanosine/uridine/cytidine transporters distinguished by different affinities to inosine, a lower affinity adenosine transporter, and a thymidine transporter. Nine Equilibrative Nucleoside Transporter (ENT) genes were identified in the T. vaginalis genome. All were expressed equally in metronidazole‐resistant and ‐sensitive strains. Only TvagENT2 was significantly upregulated in the presence of extracellular purines; expression was not affected by co‐culture with human cervical epithelial cells. All TvagENTs were cloned and separately expressed in Trypanosoma brucei. We identified the main broad specificity nucleoside carrier, with high affinity for uridine and cytidine as well as purine nucleosides including inosine, as TvagENT3. The in‐depth characterization of purine and pyrimidine transporters provides a critical foundation for the development of new anti‐trichomonal nucleoside analogues.


| INTRODUC TI ON
Trichomoniasis can be said to be a neglected and underestimated sexually transmitted infection (STI), even though several hundred million people world-wide are infected annually (World Health Organization, 2012). Although symptoms are usually mild or even absent (Edwards et al., 2016), the potential sequelae are severe.
These include adverse pregnancy outcomes (Silver et al., 2014) and increased transmission of viral infections as a result of damage to the epithelial layers of the reproductive tracts, including HIV (Kissinger & Adamski, 2013), HSV-2 (Gottlieb et al., 2004;Kissinger, 2015), and HPV (Feng et al., 2018;Raffone et al., 2021), increasing the incidence of AIDS, genital herpes and cervical neoplasia, respectively.
Vertical transmission during birth has also been documented (Peters et al., 2021). Although the infection is particularly ignored in men, as a result of its often-asymptomatic nature, here too are severe longterm risks, including diminished fertility, urethritis, prostatitis and, again, higher risk of HIV infection (Van Gerwen et al., 2021).
The infection is routinely treated with the relatively cheap nitroimidazole drug metronidazole or, in a minority of cases, its more recent derivative tinidazole (Kissinger, 2015). However, reports of clinical resistance to these drugs have increased. Although the level of resistance is generally mild-to-moderate (Marques-Silva et al., 2021) the adverse effects of the drugs, including neurologic maladies, nausea, metallic taste, and hypersensitivity (Muzny et al., 2020), are such that dosage cannot easily be increased. The need for alternative anti-trichomonal therapies when nitroimidazoles are ineffective or not tolerated is therefore undisputed. Among the most promising targets in trichomonads is their nucleotide metabolism, as they lack the ability to synthesize either purines  or pyrimidines (Heyworth et al., 1984;Wang & Cheng, 1984) de novo. This makes the parasites vulnerable to inhibitors of key enzymes of the nucleoside salvage pathways and to subversive substrates. Nucleoside analogues with strong antitrichomonal activity have been identified and include formycin A (Munagala & Wang, 2003), adenine arabinoside, 2′-F,2′-deoxyadenosine, and 2′-F,2′-deoxyarabinoadenosine (Shokar et al., 2012). Very recently, we reported on a range of 7-substituted,7-deazaadenosine analogues with mid-nanomolar activity against T. vaginalis in vitro and one compound, 7-(4-Cl-phenyl),7-deazaadenosine, was shown to be efficacious in a murine model of vaginal trichomonad infection (Natto et al., 2021).
Trichomonas vaginalis must salvage extracellular purines and pyrimidines, and expresses 5'-ecto-nucleotidases and NTPDases to hydrolyse nucleotides to their corresponding nucleosides (Menezes et al., 2016;Tasca et al., 2003), which are then internalized by transporters. Most nucleoside antimetabolites rely on those same transporters and thus their substrate selectivity is a key determinant as to which analogues will be efficiently taken up (Campagnaro & De Koning, 2020). In Trypanosoma brucei, for instance, sensitivity to tubercidin (7-deazaadenosine) and cordycepin (3'-deoxyadenosine) depends on the expression of the TbAT1 aminopurine transporter (Geiser et al., 2005), while sensitivity to 7-Br,3'-deoxytubercidin does not (Hulpia et al., 2019). Similarly, sensitivity to tubercidin and formycin B in Leishmania donovani depends on the NT1 and NT2 nucleoside transporters, respectively (Galazka et al., 2006;Vasudevan et al., 2001), and sensitivity to adenine arabinoside (AraA) in Toxoplasma gondii on the TgAT1 adenosine transporter (Chiang et al., 1999). However, nucleoside and nucleobase transport have been poorly studied in T. vaginalis. A single report from 1988 describes two nucleoside transport activities, one that transports all nucleosides and one selective for adenosine, guanosine and uridine (Harris et al., 1988); neither was inhibited by nucleobases, although adenine and guanine have both been shown to be incorporated into the T. vaginalis nucleotide pool (Munagala & Wang, 2003). Yet, the genome of T. vaginalis contains nine genes of the Equilibrative Nucleoside Transporter (ENT) family (TrichDB.org), to which, to date, all protozoan nucleoside and nucleobase transporters have been attributed (Campagnaro & De Koning, 2020;De Koning et al., 2005), although there are some indications that there may be some protozoan nucleobase transport activities that are not encoded by ENT genes (Campagnaro, Alzahrani, et al., 2018;De Koning, 2007). We therefore performed a comprehensive examination of nucleoside and nucleobase transport in T. vaginalis trophozoites with the objective to begin the process of assigning specific transport activities to individual genes, as well assessing their relative levels of expression and their regulation in the presence and absence of substrate and feeder cells. Expression of the individual TvagENT genes in a Trypanosoma brucei cell line allowed us to identify the main high affinity, broad specificity nucleoside transporter, which turned out to be encoded by TvagENT3.

| Nucleoside transport in T. vaginalis trophozoites
For all permeants (i.e., the substrate for which permeation is being measured, as opposed to inhibitors that are only potential permeants), time course experiments were first undertaken to establish (1) whether uptake could be discerned at a certain radiolabel concentration, and (2) that uptake is linear and through zero over a given period. If these conditions are not met, the parameters determined might be of the rate-limiting step, which could be a metabolic enzyme rather than the transporter for this radiolabel. In all cases, care was taken to use a very low starting permeant concentration so as to obtain the most accurate K m and K i values and Hill slopes, and avoid (partial) saturation of very high-affinity transporters as much as possible. This has the added benefit of extending the linear range of uptake as the low rate of permeant entry at those concentrations will not easily saturate the downstream metabolic reactions. For some permeants, the lower concentration limit was determined by detectability of low rates of uptake. Higher radiolabel concentrations were used when specifically probing the existence of loweraffinity transporters and in these cases, again, linearity of uptake was first established. Dose-response experiments used incubation times near the middle of the established linear uptake period.
Linearity was queried using the Prism runs test for deviation of linearity. Uptake was considered significant if the slope was significantly non-zero (F-test, Prism) and saturability by high levels of unlabeled permeant was tested with the function for significant difference between linear regression lines (F-test). All transporter data presented in figures are single experiments with data points representing the mean and SEM of triplicate determinations (unless otherwise indicated), and are representative of multiple similar repeats. For the transport experiments with T. vaginalis trophozoites, close attention was paid to the Hill slope of inhibition experiments, as a Hill slope above −1 (usually between −1 and −0.5) will indicate the uptake of permeants by more than one transporter with non-identical K m for the inhibitor. In a complex cellular system with multiple transporters of overlapping substrate selectivity this is an important parameter and a Hill slope that is consistently >−1 permits the plotting of the inhibitor data to a bi-phasic inhibitor model.

| Adenosine uptake Submicromolar concentrations of [ 3 H]-adenosine were rapidly
taken up by T. vaginalis trophozoites. Figure 1a shows that transport of 0.25 µM [ 3 H]-adenosine was linear over 60 s with a rate of 1.15 pmol(10 7 cells) −1 s −1 . The uptake was >99% inhibited by 1 mM unlabeled adenosine, and the remainder was not significantly different from zero (p = .35, F-test). The uptake was therefore saturable and most likely carrier-mediated. The inset in Figure 1a is a technical control showing that stopping the uptake with 1 ml of 1 mM ice-cold adenosine did stop all uptake, as previously reported for other cell types.

| Guanosine uptake
The above findings suggest the expression of two similar adenosine transporters with, judging by the Hill slopes, slightly different affinities for adenosine, cytidine, and guanosine. This should be mirrored when using [ 3 H]-guanosine as radiolabel, unless there is a separate adenosine-insensitive guanosine transporter as well. Uptake of 1 µM [ 3 H]-guanosine was linear for at least 60 s and almost completely saturable (98.6% reduced) by 250 µM guanosine (Figure 1e). Sigmoid inhibition plots showed that the Hill slopes for nucleoside inhibitors were indeed indicative of two transporters with somewhat different affinity: −0.67 ± 0.03 (adenosine), −0.78 ± 0.08 (guanosine), and −0.72 ± 0.06 (cytidine), respectively. Using the EC 50 values from plotting to a monophasic sigmoid curve with variable slope apparent K i,app values of 4.1 ± 0.9 µM and 18.6 ± 0.9 µM were obtained for adenosine and cytidine, respectively, and a K m,app of 9.0 ± 0.6 µM was calculated for guanosine (n = 3 for all) (Figure 1f). It is acknowledged that most of these K i,app values are highly likely to be composites of at least two separate transport activities. Also consistent with the [ 3 H]-adenosine transport data was that uridine again displayed a Hill slope close to −1 (−0.91 ± 0.02) and high affinity (K i,app 7.8 ± 1.3 µM) while the affinity for thymidine was much lower (K i,app 206 ± 62 µM). Finally, the Hill slope for inosine (−0.75 ± 0.07) was again indicative of at least two separate transport activities, with K i,app 6.3 ± 2.6 and 146 ± 26 µM, respectively. It can be tentatively concluded on the evidence that guanosine is taken up by the same transporters as adenosine, at least at low permeant concentrations.  In the presence of unlabeled adenosine uptake over 60 s was not significantly different from zero (0.0091 ± 0.0085 pmol(10 7 cells) −1 s −1 , p = .35, F-test). In contrast, uninhibited uptake was highly significant (1.15 ± 0.06 pmol(10 7 cells) −1 s −1 ), p < 0001). Inset: Same data, but also showing two extra data points representing samples that were incubated for 20 s with the [ 3 H]adenosine but left uncentrifuged for an additional 120 or 240 s after the addition of ice-cold 1 mM adenosine to stop the uptake. Uptake over the additional 240 s was not significantly different from zero (−0.0021, pmol(10 7 cells) −1 s −1 ), p = .62). (b) Dose-dependent inhibition of 0.15 µM [ 3 H]-adenosine transport over 30s, with unlabeled adenosine, plotted to a sigmoid curve, Hill slope −0.974. Inset: Conversion of the same data to a Michaelis-Menten saturation plot. (c) Dose-dependent inhibition of 0.15 µM [ 3 H]-adenosine transport over 30s, with unlabeled inosine, plotted to a biphasic curve for two-site competition, or with guanosine, plotted to a one-site inhibition curve with variable slope. (d) As (c), showing monophasic inhibition of adenosine transport with three pyrimidine nucleosides. Hill slopes were −0.99, −0.89 and −0.82 for cytidine, thymidine and uridine, respectively. (e) Transport of 1 µM [ 3 H]-guanosine with a rate of 0.22 ± 0.01 pmol(10 7 cells) −1 s −1 , significantly different from zero (p < .0001) and not significantly non-linear (p = .71). In the presence of 0.25 mM unlabeled guanosine, the rate was reduced to 0.0030 ± 0.0004 pmol(10 7 cells) −1 s −1 (r 2 =0.96; significantly different from zero (p = .004), not significantly non-linear (p = .90).

| Low affinity adenosine transport
The presence of a low-affinity adenosine transport capability was suspected not just on the basis of inhibition on the low affinity thymidine flux but on the Hill slope for adenosine inhibition of multiple permeants being above −1, including the inhibition curves for 150 nM [ 3 H]-adenosine (section 1). We probed the presence of this low-affinity carrier using a concentration of 20 µM [ 3 H]-adenosine to fully saturate the high-affinity adenosine uptake. The K m,app was determined as 59.4 ± 6.1 µM and V max as 31 ± 2 pmol(10 7 cells) −1 s −1 (n = 3; Figure 4a), not statistically different from the K i value for adenosine on the low-affinity thymidine transport activity (p = .29, Student's unpaired t-test). Similarly, the K i,app for thymidine, at 557 ± 135 µM (n = 4; Figure 4b) was not significantly different from the thymidine low affinity K m,app (p = .66), an indication that this may be the same transporter. Apparent K i values for uridine ( Figure 4b) and cytidine were 376 ± 63 µM and 116 ± 5 µM, respectively (n = 3).
No inhibition above 50% was observed at the highest tested concentrations of guanosine (250 µM, n = 2), adenine (1 mM, n = 3), hypoxanthine (1 mM, n = 3), or uracil (10 mM, n = 2). We conclude that this is a high capacity transporter with modest affinity for adenosine and low affinity for the pyrimidine nucleosides.

| Inosine transport
Inosine uptake could be measured at 1 µM at a rate of 0.0081 pmol(10 7 cells) −1 s −1 , which was significantly (p < .0001, F-test) but incompletely inhibited by 1 mM unlabeled inosine (88%, significantly non-zero uptake, p = .0018; Figure 5a). Dose-response experiments for guanosine and uridine were 7.41 ± 2.92 µM and 3.60 ± 1.14 µM, respectively (n = 3), both with Hill slopes of approximately −1. This inhibition pattern is very close to that of the high-affinity adenosine transport described above. However, the high-affinity adenosine transport split in two parts, with high and with low affinity for inosine ( Figure 1c). The uptake described in this section clearly delineates the transport component with the lower inosine affinity. We did not observe any high-affinity inosine transport, the rate of which may have been insufficient to be clearly observed over the high capacity lower affinity uptake activity.

| Uracil transport
We next probed whether T. vaginalis is able to salvage nucleobases from its environment, starting with uracil. Uptake of 0.5 µM a K m,app of 257 ± 58 µM, V max 1.58 ± 0.08 pmol(10 7 cells) −1 s −1 ( Figure 5d). Other pyrimidine nucleobases were similarly able to inhibit this transporter, with K i,app for cytosine, thymine and 5Furacil at 334 ± 73 µM, 57 ± 8 µM and 265 ± 57, respectively (all n = 3). In contrast, pyrimidine nucleosides (uridine, thymidine, up to 10 mM) and purine nucleosides (adenosine, inosine, up to 1 mM) were unable to inhibit uracil transport, even at high concentrations. We conclude that this is a low-affinity pyrimidine nucleobase transporter. However, in preliminary experiments saturable uptake of cytosine and thymine over 120 s was not significant at 0.25 µM and 0.5 µM, respectively.
2.1.9 | Transport of purine nucleobases Transport of [ 3 H]-guanine was measurable at 2 µM (as lower concentration yielded insufficient signal), producing a rate of 0.017 ± 0.002 pmol(10 7 cells) −1 s −1 (Figure 6a). The slope of the linear regression line was significantly non-zero (p = .0035) and not significantly non-linear (p = .50; r 2 = 0.96) but limits to guanine solubility did not allow stringent testing of saturability or a dose-dependent inhibition with unlabeled guanine. It appears that most natural purines and pyrimidine nucleosides can be taken up by multiple transporters (Table 1). This is an important conclusion but these observations make construction of an unambiguous model for purine and pyrimidine salvage based on analysis of whole-organism cellular uptake studies challenging.
Equally, it is difficult to construct a full model of nutrient salvage by an organism through a reductive approach of characterizing single transporters through heterologous expression aloneclearly, the two approaches are complementary in order to arrive at a full understanding of purine/pyrimidine salvage. Nonetheless, the cells express at least two high affinity, broad specificity transporters for purine and pyrimidine nucleoside, one of which has high affinity for inosine, one low affinity. Neither has significant affinity for nucleobases. Of the potential substrates, adenosine appears to be taken up more efficiently than the others, as expressed by V maz / K m,app (Table 1) and uridine appears to be the best pyrimidine substrate. Thymidine was salvaged relatively poorly, and through separate higher and lower affinity transporters, which may make T. vaginalis vulnerable to antifolates. We also found low affinity uptake of cytidine but are unable to say, based on the current data, whether this is a separate transport activity or the same carrier that is responsible for the low affinity thymidine uptake. Separate nucleobase transporters were also found, including a high-affinity and a low-affinity adenine carrier as well as likely uptake of guanine, although this remains largely uncharacterized. The only pyrimidine nucleobase uptake we observed was low affinity, low capacity uracil uptake.

| ENT genes of T. vaginalis
In order to gain further insights into purine and pyrimidine salvage in T. vaginalis, we next probed its genome for genes of the Equilibrative Nucleoside Transporter (ENT) family, which to date is the only gene family linked to nucleoside/nucleobase transport in protozoa (Campagnaro & De Koning, 2020). BLAST and keyword searches of TrichDB identified nine putative ENT family members, which we designated TvagENT1-9, with 1,011-3,377 bp and 9-11 predicted transmembrane domains (TMDs) ( Table 2). Their sequences and a multiple alignment of the nine genes are shown in the Supporting Information ( Figure S1). The alignment shows that most of the variation in length is from high variability in the size of the N-terminal domain, most pronounced for TvagENT2, which lacks 135 N-terminal amino acids relative to TvagENT1, including all of TMD1 and part of TMD2, posing the question whether this transporter can be fully functional; it is the only TvagENT with less than 10 TMDs. From a phylogenetic tree (Figure 7) it can be seen that all TvagENTs are more similar to each other than to other protozoan or human ENT genes, and that the TvagENTs essentially split into two clades, with one consisting of TvagENT1, 2, 3, 6, and 8, and the other of TvagENT4, 5, 7, and 9 (Table 3).

| Expression levels of TvagENTs in trophozoites
The relative level of expression of the TvagENTs in trophozoites was determined using quantitative real time polymerase chain reaction (qRT-PCR), standardized to the glycolytic enzyme glyceraldehyde 3-phosphate dehydrogenase (GAPDH), in a panel of three metronidazole-sensitive and three resistant T. vaginalis isolates. The expression pattern of the TvagENTs was very similar in all six isolates, with the highest expression consistently found for TvagENT5 ( Figure 8). An exception was TvagENT9, which was expressed at a low level in isolates G3 and S1469 but robustly in all other strains

| Heterologous expression of TvagENTs in Trypanosoma brucei brucei TbAT1-KO
To start assigning gene identities to the observed nucleoside and nucleobase transport activities of T. vaginalis trophozoites, each individual gene was to be expressed in a T. b. brucei strain, TbAT1-KO, in which we had successfully characterized nucleoside transporters from several other protozoan species including Trypanosoma cruzi (Campagnaro, de Freitas Nascimento, et al., 2018), Trypanosoma congolense , Leishmania spp , and Toxoplasma gondii (Campagnaro et al., manuscript in preparation).
As T. b. brucei is able to synthesize its own pyrimidine nucleotides (Ali, Tagoe, et al., 2013), bloodstream forms are known to have a very limited ability to take up any pyrimidines (Ali, Creek, et al., 2013;Ali, Tagoe, et al., 2013;Gudin et al., 2006) except uracil, which is efficiently taken up by the U3 transporter (Ali, Tagoe, et al., 2013). However, the P1 transporter sub-family, consisting of at least eight distinct ENT-family genes (Al-Salabi et al., 2007;Campagnaro & De Koning, 2020;Sanchez et al., 2002), is also capable of some thymidine uptake (Ali, Creek, et al., 2013) with affinity of 44 µM (De Koning & Jarvis, 1999), although P1-type transporters are principally high affinity, broad specificity purine nucleoside carriers (Campagnaro & De Koning, 2020;. For these reasons, the TbAT1-KO strain is particularly useful for the expression and characterization of transporters of uridine and, particularly, cytidine, which was fortuitous as these are high-affinity substrates of the main broad specificity nucleoside transporter(s) of T. vaginalis (Table 1).
The nucleotide sequences of the nine TvagENTs were codonoptimized for T. brucei codon usage bias using the Codon Adaptative Index described by Rashmi and Swati (2013). The thus optimized sequences were custom synthesized and cloned into the pHD1336 expression vector (Biebinger et al., 1997;Munday et al., 2015) for expression in the T. b. brucei strain TbAT1-KO, from which the aminopurine transporter TbAT1 had been deleted by homologous recombination (Matovu et al., 2003). Plasmid DNA was then used to transfect bloodstream forms of T. b. brucei TbAT1-KO followed by selection on blasticidin, and cloned out by limiting dilution. The presence and integrity of the constructs were verified by Sanger sequencing. For each TvagENT, three clones were screened by qRT-PCR for expression. The highest-expressing clone was selected for each gene of interest, to be used in nucleoside transport experiments.
As expected, a screen using a low concentration of [ 3 H]-thymidine was not particularly revealing, due to the relatively high background, although cells expressing TvagENT3 again displayed the highest rate, closely followed by those expressing TvagENT6 (p > .05; Figure 11a). Uptake of 0.25 µM [ 3 H]-uridine suffered from similar issues but highlighted TvagENT6 as the most efficient uridine transporter (Figure 11b).
The high background rate in control cells was caused by uridine uptake through a uracil transporter . Thus, the addition of 250 µM of uracil to the assay buffer reduced background uridine uptake by ~85% (Figure 11c) and allowed the determination of K m and K i values (Figure 11d). The uridine K m,app for TvagENT6 was 169 ± 39 µM with a V max of 0.092 ± 0.021 pmol(10 7 cells) −1 s −1 (n = 5). K i,app values for adenosine and inosine were 47.2 ± 6.4 µM (n = 4) and 40.0 ± 9.5 µM (n = 5), respectively (Figure 11d). The transporter was not significantly inhibited by adenine up to 1 mM (n = 2). From this partial characterization, it appears that TvagENT6 encodes a transporter with moderate affinity for purine nucleosides, lower affinity for uridine, and probably some limited capacity for thymidine uptake.

| D ISCUSS I ON
In this study, we have attempted to map out the nucleoside and nucleobase transport activities of T. vaginalis. The effort is complicated as trophozoites express multiple such transporters, with overlapping substrate selectivities and a wide range of substrate affinities. This makes it hard to characterize single carriers by kinetic analysis of whole-cell transport assays only. Yet, the whole-cell analysis helped to establish which purine and pyrimidines are salvaged by the parasite and with what affinities and rates. The efficiency of uptake, defined as V max /K m , was highest for high-affinity adenosine transport, at 1.5, compared to 0.10 for high-affinity guanosine uptake, making adenosine the preferred purine substrate. Uptake of adenine was relatively weak with low V max , with an efficiency of 0.22. For pyrimidines the highest affinity and efficiency was observed for uridine followed by cytidine and then thymidine. Although uracil uptake is robust in kinetoplastid parasites Papageorgiou et al., 2005) and Toxoplasma gondii (Natto and De Koning, unpublished), as well as other microbes including Saccharomyces cerevisiae, Aspergillus nidulans and E. coli (Campagnaro & De Koning, 2020;De Koning & Diallinas, 2000), uptake of uracil (efficiency 0.0061) and other pyrimidine nucleobases was marginal. F I G U R E 7 Phylogenetic tree of human and protozoan ENT family genes including T. vaginalis. Multiple alignment (MUSCLE), phylogeny (PhyML) and tree rendering (TreeDyn) were performed using phylogeny.fr (http://www.phylo geny.fr/ index.cgi). Accession codes for all genes are listed in the Supporting Information. The Scale bar indicates the number of substitutions of a given site The observations of efficient uptake of purine nucleosides rather than nucleobases are largely consistent with observations and models from before the T. vaginalis genome was published (Carlton et al., 2007). Hypoxanthine has been reported to not or barely be incorporated into the T. vaginalis nucleotide pool Munagala & Wang, 2003) and indeed we found little or no uptake of this nucleobase. However, adenine and guanine could be salvaged and incorporated into nucleotides (Munagala & Wang, 2003). Miller and Lindstead (1983) reported that they were unable to detect any phosphoribosyltransferase activity with any of the purine nucleobases, only with uracil, but the same authors did find that the purine nucleobases, including hypoxanthine and xanthine, could be converted by cell-free extracts of T. vaginalis to nucleosides using purine nucleoside phosphorylase (PNP). This enzyme has been isolated from extracts (Miller & Miller, 1985) and cloned (Munagala & Wang, 2002), and the structure has been elucidated (Rinaldo-Matthis et al., 2007). For the pyrimidine nucleobase, a uracil phosphoribosyltransferase activity was reported in T. vaginalis extracts by Miller and Lindstead (1983) but not by Wang and Cheng (1984), who attributed the incorporation of uracil to a uridine phosphorylase. We were unable to identify a candidate gene for uracil phosphoribosyltransferase in the T. vaginalis genome but did identify a candidate uridine phosphorylase (XP_001323814.1; TrichDB).
We observed very little overlap between nucleoside transporters and nucleobase transporters in T. vaginalis, as is generally the case for protozoan purine transporters (Campagnaro & De Koning, 2020;De Koning et al., 2005), although there are some notable exceptions like the TbAT1 adenosine/adenine transporter of T. b. brucei (De Koning & Jarvis, 1999), the UUT1 uridine-uracil transporter of Leishmania major  and the Plasmodium falciparum NT1 carrier that transports both hypoxanthine and adenosine (Quashie et al., 2008). The strict separation of nucleoside and nucleobase transport activities in T. vaginalis appears to fit well with the organism's overall preference for nucleosides over nucleobases.
Indeed, it is possible that the nucleobase carriers have more of a sensory/regulatory role than one of providing significant amounts of nutrients, judging by their very low V max and/or efficiency, and the observation that the presence of hypoxanthine influenced the expression of some of the TvagENTs, especially TvagENT1.
The most active nucleoside transporter in T. vaginalis trophozoites appears to most efficiently transport adenosine, in keeping with adenosine in the form of ATP being most likely the most abundant purine available to it for salvage and T. vaginalis expressing a number of apyrases and ecto-phosphohydrolases to convert extracellular nucleotides to the corresponding nucleosides (de Aguiar Matos et al., 2001;de Jesus et al., 2006). The same transporter also displayed high affinity for uridine, guanosine and cytidine (Table 1).
Our analysis strongly suggests there are at least two such transporters expressed in trophozoites, one with high affinity for inosine, here provisionally designated NT-a until the gene ID is definitively known, and one with low affinity, NT-b. In addition, we found a lower affinity adenosine >cytidine > uridine >thymidine transporter (NT-c), and a thymidine transporter with similar affinity for uridine and moderate affinity for adenosine and guanosine (NT-d). Together, our analysis suggests a minimum of 4 nucleoside transporters and up to three nucleobase transporters in T. vaginalis (NT-e for adenine, NT-f for uracil, NT-g for guanine) ( Figure 12). However, the characterization of guanine transport has been incomplete, without K m or inhibition profile and we cannot exclude the possibility that it overlaps with the NT-e adenine transporter activity or one of the nucleoside transport activities.
The low-affinity [ 3 H]-inosine transport is most likely mediated by the NT-b activity as the K i values for adenosine, guanosine, and uridine are all very close to those obtained with [ 3 H]-adenosine.
Similarly, high affinity uptake of uridine and cytidine appears to be mediated jointly by NT-a and NT-b. The low affinity [ 3 H]-thymidine uptake could well be mediated by NT-c, considering the adenosine K i value is highly similar to the K m for NT-c and the NT-c K i for thymidine is similar to the K m for low affinity thymidine uptake.
Although our comprehensive transporter characterizations give a well-supported view of nucleoside and nucleobase uptake in T.
vaginalis, the data cannot resolve how many and which ENT-family (or other gene family) genes are involved in the various transport processes. As a complement to the functional transport studies in trophozoites, we therefore identified the TvagENT genes from the TA B L E 3 Percent amino acid identity among T. vaginalis ENT proteins   TvagENT1  TvagENT2  TvagENT3  TvagENT4  TvagENT5  TvagENT6  TvagENT7  TvagENT8  TvagENT9   TvagENT1  100  TrichDB database, had them synthesized in the Trypanosoma brucei codon preference and introduced them individually in the TbAT1-KO strain of T. b. brucei. The new strains were screened for the transport of pyrimidine nucleosides and uptake of cytidine was the most effective tool, identifying TvagENT3 as the carrier that mediated its uptake most strongly, at conditions designed to identify the high-affinity flux. The inhibition profile aligned quite closely to that of NT-a, with high affinity for inosine, adenosine, guanosine, and uridine but not thymidine or nucleobases. TvagENT3 is functionally related to the broad specificity T. brucei P1-type transporters (De Koning & Jarvis, 1999), although the relatively high affinity for cytidine may be unique amongst protozoan transporters characterized to date, and like P1, its broad specificity/high affinity should be valuable in targeting cytotoxic nucleoside analogues to the T.
vaginalis interior (Geiser et al., 2005;Hulpia et al., 2019;Ranjbarian et al., 2017). Indeed, we have very recently reported the identification of a series of strongly antitrichomonal nucleoside analogues (Natto et al., 2021) and other nucleoside analogues with activity against this parasite have been reported (Munagala & Wang, 2003;Shokar et al., 2012). Definitive assignment any gene ID to a specific transport activity in T.
vaginalis trophozoites, however, will require further studies, such as gene deletions or the identification of sufficiently specific inhibitors.
In this work, we present a model of multiple high-and low-affinity, broad specificity nucleobase transporters and separate nucleobase transporters in T. vaginalis. Only a single study of purine or pyrimidine transport in this species has previously been reported (Harris et al., 1988). That study did not address nucleobase uptake but did describe the expression of at least two high affinity, broad specificity nucleoside transporters with overlapping substrate specificity.
Indeed, their K m values for adenosine (3.9 µM), guanosine (13.9 µM), and uridine (2.5 µM) are remarkably similar to those we report here for the NT-a activity (6.2, 12.2, and 3.7 µM, respectively using [ 3 H]adenosine as substrate). In addition, we identified three additional nucleoside transport activities and three nucleobase transport activities in trophozoites and identified the genes encoding two of the main nucleoside transporters through heterologous expression. This study thus presents the most complete model yet of nucleoside/nucleobase transport in a trichomonad, or indeed any protozoan that is auxotrophic for both purines and pyrimidines. We find that the main transporters of T. vaginalis are able to take up purine nucleosides and pyrimidine nucleosides with similar affinity. In keeping with most protozoa no pyrimidine-specific nucleoside transporters were identified (the sole documented exception is the T. cruzi NT2 transporter (Campagnaro, de Freitas Nascimento, et al., 2018)), although NT-d did display higher affinity for uridine and thymidine than for adenosine and guanosine. Mixed purine-pyrimidine nucleoside transporters have been described in Leishmania (Vasudevan et al., 1998) and Toxoplasma gondii (De Koning et al., 2003), whereas purine-specific nucleoside transporters are the norm in T. brucei (De Koning & Jarvis, 1997 and Plasmodium falciparum (Quashie et al., 2008).
A uracil transporter, NT-f, was also identified and uracil-specific transporters were previously reported in L. major (Papageorgiou et al., 2005) and T. brucei . The one new feature of the T. vaginalis transporters, relative to those of other protozoa, is the high affinity for cytidine by NT-a and NT-b.
The characterization of the T. vaginalis purine and pyrimidine transporters provides a critical foundation for the design of novel trichomonacidal agents that target the nucleoside metabolism in the parasite.

| Transport assays
Results were plotted and analyzed with GraphPad Prism (versions 8 and 9), using the inbuilt statistics packages to determine line-

| RNA extraction and real-time RT-PCR of T. vaginalis
Total cellular RNA was extracted from lysates made with TRIzol (Invitrogen, 1 ml per 10 7 cells or confluent 10 cm 2 flask) and puri- F I G U R E 1 2 Diagram of nucleoside (red boxes and arrows) and nucleobase (orange boxes and arrows) transporters in Trichomonas vaginalis. HA, high affinity; LA, (relatively) low affinity. Substrates indicated in brackets are of substantial lower affinity than those without brackets. Arrows indicate bi-directional traffic (equilibrative) but it is not yet known whether (some of) the transporters are active transporters, which would imply mono-directionality. The guanine uptake may be mediated by a separate transporter, which would be designated NT-g. However, at this juncture the possibility that the guanine uptake is mediated by one of the other transporters rather than a separate gene product cannot be excluded

| Expression of TvagENTs in T. brucei
The full-length coding sequences of 9 TvagENT genes (Table 2) were retrieved from TrichDB.org by BLASTP and keyword searches. The nucleotide sequences were optimized for expression in T. brucei using a codon optimization algorithm at https://www.idtdna.com/CodonOpt (Rashmi & Swati, 2013). The thus optimized sequences were synthesized by BaseClear BV, except for TvagENT8, which was synthesized by Genewiz and all delivered in the vector pUC57-Amp. Each gene was amplified by PCR using primers that introduced 5'-HindIII and 3'-BamH1 restriction sites (Table S2); the same enzymes were used to digest pHD1336 (Biebinger et al., 1997) and the PCR products were ligated in using T4 DNA ligase (NEB). The constructs were used to transform E. coli XL1-blue competent cells (Agilent), which were then grown on ampicillin agar plates. Colonies were screened by PCR using the vector specific primer HDK528 (CTCTAGAGGATCCTATGCGTGACTGAGTGAGCC) and the relevant reverse primer for the inserted gene (Table S2) and, if yielding the correct size of amplicon, further verified by Sanger sequencing (Source BioScience, Nottingham, UK). Plasmid DNA from selected colonies was linearized by digestion with NotI before transfection into T. brucei strain TbAT1-KO (Matovu et al., 2003) using an Amaxa Nucleofector, program X-001 (Burkard et al., 2007). Transfectants were grown and cloned out, by limiting dilution, in complete HMI-11/FBS media containing 5 µg/ml blasticidin S. Correct integration of the expression cassettes was analyzed by PCR.

| qRT-PCR of TvagENTs expressed in T. brucei
Cells were harvested from 11 ml flasks, each containing ~3 × 10 6 cells/ ml for one clonal line expressing a TvagENT in T. brucei TbAT1-KO, by centrifugation at 2,000 RPM for 10 min. The supernatant was discarded and RNA was isolated using the NucleoSpin ® RNA isolation kit in accordance with the manufacturer's instructions. RNA was eluted off the column with 15 µl RNAase-free water and the concentration determined using a Nanodrop spectrophotometer.
cDNA was produced using the Precision nanoScript2 Reverse Transcription kit (Primer Design). Primers were annealed to 9 μl of RNA sample using 0.5 μl oligo-dT primers and 0.5 μl Random nonamer primers for 5 min at 65°C and subsequently cooled on ice for 5 min. The extension master mix consisted of 5 μl nanoScript2 4x buffer, 1 μl 10 mM dNTP, 2.5 μl RNAse free water, and 1 μl nano-Script2 enzyme. This was added to 10 µl RNA in water, briefly vortexed, vortexed and incubated at room temperature for 5 min before extending at 42°C and a heat denaturation step at 75°C for 10 min.
qPCR primers (Table S1) were designed using Primer Express 3000 software, selecting for primers with >50% GC content and the forward and reverse being of similar length. Primers were diluted in RNAse free water to 100 µM stock concentration and stored at −20°C. The forward and reverse primers were then combined at a concentration of 3 μM each. 15 μl of master mix, consisting of 10 μl PrecisionPLUS, 1 μl primer mix and 4 μl RNase free water, was added to wells in a 96-well plate. cDNA was diluted to 4 ng/μl using RNAse free water and 5 μl of cDNA from each clone was added to the master mix. Plates were designed to have three repeats for the TvagENT and the GPI8 control gene, and four repeats for the empty vector control as well as four no cDNA controls for both the TvagENT and the GPI8 primers. The qPCR was run on a 7500 Real Time PCR system for 40 cycles in accordance with the PrecisionFAST Mastermix protocol and analyzed using SDS software.

| Data analysis
All values for IC 50 , K m and V max are presented as means ± SEM of at least three independent determinations in triplicate but individual plots shown as Figures are always single experiments with data points being the means ± SEM of the triplicates. Hill slopes were calculated by plotting dose-response inhibition data to an equation for a sigmoid plot with variable slope (4 parameter) in GraphPad Prism.
Rates of transport in timecourses were calculated by linear regression in Prism and the in-build statistical analysis were used to determine deviation from linearity, significance of deviation from zero slope (flat line), and significance of difference in slope between two lines within the same experiment.
For mRNA measurements, experiments were repeated at least three times, fold changes were expressed as mean and SEM of log10transformed expression values. Significance was evaluated by ANOVA with Dunnett's post-hoc test using Prism (Graphpad software).

ACK N OWLED G M ENTS
The and National Institutes of Health grants DK120515 and AI158612.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest regarding the publication of this paper.

DATA AVA I L A B I L I T Y S TAT E M E N T
Sequences used are those of the publicly available database trichDB.
All other relevant data are presented in the main text and Supporting Information sections of the paper.