Characterization of the apicomplexan amino acid transporter (ApiAT) family in Toxoplasma gondii

Apicomplexan parasites are auxotrophic for a range of amino acids which must be salvaged from their host cells, either through direct uptake or degradation of host proteins. Here, we describe a family of plasma membrane-localized amino acid transporters, termed the Apicomplexan Amino acid Transporters (ApiATs), that are ubiquitous in apicomplexan parasites. Functional characterization of the ApiATs of Toxoplasma gondii indicate that several of these transporters are important for intracellular growth of the tachyzoite stage of the parasite, which is responsible for acute infections. We demonstrate that the ApiAT protein TgApiAT5-3 is an exchanger for aromatic and large neutral amino acids, with particular importance for L-tyrosine scavenging and amino acid homeostasis, and that TgApiAT5-3 is critical for parasite virulence. Our data indicate that T. gondii expresses additional proteins involved in the uptake of aromatic amino acids, and we present a model for the uptake and homeostasis of these amino acids. Our findings identify a family of amino acid transporters in apicomplexans, and highlight the importance of amino acid scavenging for the biology of this important phylum of intracellular parasites. Author Summary The Apicomplexa comprise a large number of parasitic protozoa that have obligate intracellular lifestyles and cause significant human and animal diseases, including malaria, cryptosporidiosis, toxoplasmosis, coccidiosis in poultry, and various cattle fevers. Apicomplexans must scavenge essential nutrients from their hosts in order to proliferate and cause disease, including a range of amino acids. The direct uptake of these nutrients is presumed to be mediated by transporter proteins located in the plasma membrane of intracellular stages, although the identities of these proteins are poorly defined. Using a combination of bioinformatic, genetic, cell biological, and physiological approaches, we have characterized a family of plasma membrane-localized transporter proteins that we have called the Apicomplexan Amino acid Transporters (ApiATs). The family is found in apicomplexans and their closest free-living relatives. We show that TgApiAT5-3, a member of the family in the apicomplexan Toxoplasma gondii, is an exchanger for aromatic and large neutral amino acids. In particular, it is critical for uptake of tyrosine, and for parasite virulence in a mouse infection model. We conclude that ApiATs are a family of plasma membrane transporters that play crucial roles in amino acid scavenging by apicomplexan parasites.

ApiATs are broadly-distributed in apicomplexan parasites. To identify ApiAT-family 1 0 0 proteins in the apicomplexan parasites T. gondii, Neospora caninum, Eimeria tenella, P. (www.eupathdb. org, www.blast.ncbi.nlm.nih.gov; (9, 11)). In addition to the previously  To determine the relationships between ApiAT family proteins, we constructed a multiple 1 1 7 sequence alignment. This revealed the presence of a major facilitator superfamily (MFS) 1 1 8 signature sequence between transmembrane domains 2 and 3 of most ApiAT family protein 1 1 9 ( Fig S1; (14, 15)). Most ApiAT proteins were predicted to be polytopic membrane proteins 1 2 0 containing 12 transmembrane domains (www.cbs.dtu.dk/services/TMHMM/; (16)), and 1 2 1 exhibited highest sequence similarity in the regions encompassing these transmembrane We next performed a maximum likelihood phylogenetic analysis. This revealed the presence 1 2 6 of multiple ApiAT subfamilies (Fig 1). Orthologs of the ApiAT2 subfamily were present in 1 2 7 all organisms in the study, with the exception of Cryptosporidium parvum and the free living 1 2 8 Chromerid species (Fig 1). Members of the ApiAT3, ApiAT5, ApiAT6 and ApiAT7 1 2 9 subfamilies were restricted to coccidians (a group of apicomplexans that includes T. gondii 1 3 0 and N. caninum), and the ApiAT9 family was restricted to the piroplasms T. annulata and B. bootstrap support for this association was weak (Fig 1). The ApiAT8 subfamily comprises and TgApiAT1 appear not to be orthologous.  Western blotting indicated that TgApiAT2, 1 4 5 TgApiAT5-3, TgApiAT6-1, TgApiAT6-2, TgApiAT6-3, and TgApiAT7-2 proteins were 1 4 6 expressed in tachyzoite stage parasites (Fig 2A-E). We were unable to detect expression of 1 4 and glycine to the growth medium, suggesting they can also acquire exogenous amino acids Cryptosporidium spp, an early-diverging lineage of apicomplexans (29 amongst the major apicomplexan lineages (Fig 1), suggesting its presence before these Much of the expansion of ApiAT proteins, then, appears to have occurred subsequent to the 4 2 5 evolution of parasitism in this phylum. An intriguing possibility is that expansion within ApiAT subfamilies is linked to expansion of these parasites into different hosts, and cell are expressed in the tachyzoite stage of the life cycle (Fig 2; (5)). Of the ApiAT5 subfamily, we could only detect expression of TgApiAT5-3 in tachyzoites (Fig 2), and only TgApiAT5- 3 is important for growth of the tachyzoite stage (Fig 3). This raises the possibility that other 4 3 4 TgApiAT5 subfamily proteins are expressed, and function, at other stages of the life cycle.

3 5
Interestingly, proteomic studies identified TgApiAT5-5 in the oocyst proteome 4 3 6 (www.toxodb.org), and it could be that this transporter has particular importance at this stage 4 3 7 of the parasite life cycle. Of the fifteen ApiAT family proteins that we were able to disrupt genetically, only the 4 3 9 TgApiAT1, TgApiAT2 and TgApiAT5-3 mutants exhibited defects in tachyzoite growth ( proteins are not important for parasite growth in vitro (Fig 3). We were unable to disrupt the 4 4 7 reading frame of TgApiAT6-1, despite multiple attempts using a guide RNA that targets the 4 4 8 TgApiAT6-1 locus (Table S1). TgApiAT6-1 has a phenotype score of -5.4 (31). It is likely, then, that our inability to generate a TgApiAT6-1 knockout is because it is essential for 4 5 0 parasite growth.

5 1
Our studies of TgApiAT5-3 demonstrate that this protein is important for parasite growth.

5 9
Our oocyte studies indicated that TgApiAT5-3 can function as an exchanger, with the rate of 4 6 0 uptake of L-Tyr and other aromatic and large neutral amino acids enhanced when equivalent 4 6 1 amino acids were present on the trans side of the membrane (Fig 6). performs an unusual dual function in facilitating both the net uptake of L-Tyr into the  large neutral amino acids such as L-Leu (Fig 6). This raises the possibility that TgApiAT5-3 4 7 2 also functions in the net uptake of these amino acids in the parasite. We saw no differences in WT parasites (Fig 4), implying that the uptake of these branched-chain amino acids is we did observe a defect in the uptake of [ 14 C]Phe in the mutant strain ( Fig 7C). A possible 4 7 8 explanation for this discrepancy lies in the different uptake conditions for these experiments. The [ 13 C]-labelled amino acid uptake experiments were performed in medium containing a concentrations (as we observed in the oocyte experiments, and as appears to be the case in 4 8 8 mammalian cells (32)). This is consistent with the hypothesis that TgApiAT5-3 plays little transport pathways (Fig 9). The importance of TgApiAT5-3 for L-Trp uptake in T. gondii is less clear. We were unable TgApiAT5-3 when present in equimolar amounts (Fig S5A), suggesting that TgApiAT5-3 4 9 5 has similar affinities for L-Trp and L-Tyr. We observed a defect in growth of suggesting that TgApiAT5-3 may have a role in the uptake of L-Trp at low concentrations.

9 8
The estimated plasma concentration of L-Trp in humans and mice is ~60 µM (33), a 4 9 9 concentration at which ∆ apiAT5-3 parasites grow optimally in vitro ( Fig 8C). However, T. host response to parasite infection. L-Phe and L-Trp uptake (Fig 9). At high concentrations of L-Phe and L-Trp, uptake of L-Tyr which we showed that high levels of other cationic amino acids inhibit parasite growth in the 5 1 9 absence of the selective L-Arg transporter (5).

2 0
We propose a model whereby the uptake of L-Tyr into T. gondii parasites is mediated  and L-Tyr (36). Knockout of these enzymes revealed that they are not required for tachyzoite growth, but are particularly important for producing oocysts following the sexual stages of across the entire life cycle will be of particular interest.

3 7
In this manuscript, we describe an apicomplexan-specific family of plasma membrane 5 3 8 transporters that appear to be primarily involved in amino acid uptake. Our findings highlight 5 3 9 the evolutionary novelties that must arise to enable parasites to scavenge essential nutrients Reciprocal protein BLAST searches in www.eupathdb.org were used to identify orthologues Chromera velia and Vitrella brassicaformis. Gene IDs are listed in Table S3.

3
The sequences of the 67 identified ApiAT proteins were aligned using ClustalX 2. values were generated by running the alignment file through the 'seqboot' program, which 5 5 9 was used to generate 1000 pseudosamples of the alignment. Next, multiple phylogenetic trees were generated from the pseudosamples using the 'proml' tree algorithm, using a randomised 5 6 1 order of entry and three jumbles. Finally, the multiple phylogenetic trees were converted to a the '\fingerprint' command (https://ctan.org/pkg/texshade?lang=en). Parasites were maintained in human foreskin fibroblasts (HFFs; a kind gift from Holger Schülter, Peter MacCallum Cancer Centre) cultured at 37°C in a humidified 5 % CO 2 5 7 1 incubator. Unless otherwise noted, parasites were cultured in Dulbecco's modified Eagle's 5 7 2 medium (DMEM) supplemented with 1 % (v/v) fetal calf serum and antibiotics. 'Homemade' 5 7 3 media were generated as described previously (5), with amino acids at the concentrations  TATi/∆ku80 (39), and RH∆hxgprt/Tomato (22) parasites were used as parental strains for the 5 7 7 genetically modified parasites generated in this study. Guide RNA (gRNA)-encoding sequences specific to target genes were introduced into the 5 8 1 vector pSAG1::Cas9-U6::sgUPRT (Addgene plasmid # 54467; (18)) using Q5 site-directed 5 8 2 mutagenesis (New England Biolabs) as described previously (18). A list of the forward 5 8 3 primers used to generate gRNA-expressing vectors for introducing frame-shift mutations are 5 8 4 described in Table S4. In each instance, the reverse primer 5'- genes, and transfected into parasites on a vector that also expressed Cas9-GFP. Transfections 5 8 8 were performed as described previously (40). GFP positive parasites were selected and 5 8 9 cloned using flow cytometry 2-3 days following transfection using a FACSAria I or FACSAria II cell sorter (BD Biosciences). The region of the candidate genes targeted by the 5 9 1 gRNAs were sequenced in clonal parasites, and clones in which the target gene had been primers listed in Table S5. In addition, a donor DNA sequence encoding a 3x HA tag was either side of the stop codon. Template DNA encoding the 3x HA tag was generated as a 5 9 8 gBlock (Integrated DNA Technologies), with the sequence listed in Table S6. Forward and 5 9 9 reverse primers used to amplify the HA tag for each target gene are also listed in Table S6.  3' replacement plasmids were created to epitope tag TgApiAT2, TgApiAT3-2, TgApiAT3-3, 6 0 5 TgApiAT5-3, TgApiAT6-1, TgApiAT6-2 and TgApiAT7-1 using conventional crossover 6 0 6 recombination methods as described previously (41). Regions of DNA homologous to the 3' 6 0 7 ends of the genes were amplified by PCR using primers described in Table S7, and ligated  (Table S7), then transfected into TATi/∆ku80 parasites. Parasites were selected on 6 1 1 chloramphenicol or pyrimethamine as described (40). In cases where we were unable to 3' ends (TgApiATs 5-1, 5-2, 5-4, 5-5 and 5-6; not shown), or by PCR screening (TgApiAT7-6 1 5 1). For assessing HA integration into the TgApiAT7-1 locus, DNA was extracted from 6 1 6 TgApiAT7-1-HA parasite clones, and used as template in a PCR with the primers 5'-  To complement the ∆ apiAT2 mutant with a constitutively-expressed copy of TgApiAT2, we constitutively-expressed copy of TgApiAT5-3, we amplified the open reading frame of 6 2 7 TgApiAT5-3 with the primers 5'-6 2 8 GATCGGATCCAAAATGGAGTCGACCGAGGCGACTAT and 5'- and selected on chloramphenicol. Immunofluorescence assays and western blotting 6 3 6 Immunofluorescence assays and western blotting were performed as described previously (5). For western blotting, membranes were probed with rat anti-HA antibodies (clone 3F10, 6 3 8 Sigma) at dilutions between 1:1,000 to 1:3,000, mouse anti-GRA8 (a kind gift from Gary 6 3 9 Ward, U. Vermont, (42)) at a dilution of 1:80,000, or rabbit anti-TgTom40 antibodies (43)  For immunofluorescence assays, samples were probed with the following primary antibodies: Oocyte data analysis and statistics. All oocyte data were analyzed using OriginPro (2015). All data displayed in figures represent oocytes is included in figures, uptake in uninjected oocytes was subtracted from uptake in 8 0 9 TgApiAT5-3-injected oocytes to give the 'TgApiAT5-3-mediated uptake'. All data sets were hoc test and significance tested at the P < 0.05 level. Time-course analysis of uptake and oocyte retention data of L-Tyr in TgApiAT5-3-injected 8 1 6 oocytes were fitter to 1 st order integrated rate equations: (eq. 1, retention) 8 1 8 Or: ሻ (eq. 2, uptake) substrate amount, and k is the 1 st order rate constant. Steady-state kinetic data collected under initial rate conditions were fitted to both the 8 2 4 Michaelis-Menten equation: And a Scatchard linear regression equation: In the Scatchard regression, the apparent Michaelis constant (K 0.5 ) is derived from the slope (1/െK 0.5 ) and the maximal rate (V max ) from the ordinate intercept (V max /K 0.5 ).

3 0
All curve fittings were evaluated using adjusted R 2 values as indicated in the text and figure   8 3 1 legends. All non-linear fitting was conducted using the Levenburg-Marquardt algorithm, with 8 3 2 iteration numbers varying from 4 to 11 before convergence was attained. Amino acid uptake assays were carried out as described previously (5) points. Parasite samples were centrifuged through an oil mix to separate parasites from 8 4 0 unincorporated radiolabel as described previously (5). L-Arg uptake was measured by in each amino acid tested were fitted by a single exponential function and the initial rate of 8 4 5 transport was estimated from the initial slope of the fitted line. for performing some of the early studies to characterise the role of TgApiAT5-3 in aromatic . v a n D  o  o  r  e  n  G  G  ,  S  t  r  i  e  p  e  n  B  .  T  h  e  a  l  g  a  l  p  a  s  t  a  n  d  p  a  r  a  s  i  t  e  p  r  e  s  e  n  t  o  f  t  h  e  a  p  i  c  o  p  l  a  s  t  .  A  n  n  u  R  e  v  9  0  5   M  i  c  r  o  b  i  o  l  .  2  0  1  3  ;  6  7  :  2  7  1  -8  9  .  9  0  6   2  5  .  C  o  p  p  e  n  s  I  .  E  x  p  l  o  i  t  a  t  i  o  n  o  f  a  u  x  o  t  r  o  p  h  i  e  s  a  n  d  m  e  t  a  b  o  l  i  c  d  e  f  e  c  t  s  i  n   T  o  x  o  p  l  a  s  m  a   a  s  t  h  e  r  a  p  e  u  t  i  c  9  0    transmembrane domains two and three has been highlighted.