In plants, the hydroxymethylpyrimidine (HMP) and thiazole precursors of thiamin are synthesized and coupled together to form thiamin in plastids. Mutants unable to form HMP can be rescued by exogenous HMP, implying the presence of HMP transporters in the plasma membrane and plastids. Analysis of bacterial genomes revealed a transporter gene that is chromosomally clustered with thiamin biosynthesis and salvage genes. Its closest Arabidopsis homolog, the plastidic nucleobase transporter (PLUTO), is co-expressed with several thiamin biosynthetic enzymes. Heterologous expression of PLUTO in Escherichia coli or Saccharomyces cerevisiae increased sensitivity to a toxic HMP analog, and disrupting PLUTO in an HMP-requiring Arabidopsis line reduced root growth at low HMP concentrations. These data implicate PLUTO in plastidial transport and salvage of HMP.
Thiamin diphosphate (ThDP) is an enzyme cofactor required by virtually all forms of life. ThDP-dependent enzymes participate in crucial metabolic reactions that make or break C–C bonds. Plants, fungi, and many prokaryotes synthesize thiamin de novo but it is an essential vitamin for animals.
Thiamin is composed of hydroxyethylthiazole (HET) and hydroxymethylpyrimidine (HMP) moieties. These are synthesized separately as hydroxyethylthiazole monophosphate (HET-P) and hydroxymethylpyrimidine pyrophosphate (HMP-PP) and then condensed to form thiamin monophosphate (ThMP) (Figure 1). In plants, the enzymes responsible for these reactions are found solely in the plastid . The final conversion of ThMP to ThDP occurs in the cytosol [1–3].
Compartmentation of ThDP biosynthesis in plants and implied transporters
Most of the thiamin biosynthetic enzymes in Arabidopsis have been discovered using forward genetics. Mutants deficient in synthesis of the pyrimidine (py) or thiazole (tz) moieties, or thiamin itself (th-1, th-2) were isolated approximately 50 years ago , with the last of the corresponding enzymes having been cloned only recently . Exogenous supplementation of a thiamin precursor or thiamin itself rescued the mutant phenotype, allowing identification of the specific defect in each case.
The thiamin and precursor concentrations used in these exogenous supplementation experiments are minute , implying the existence of as-yet unidentified plasma membrane transporters for uptake of thiamin or its precursors into the plant. Moreover, to be incorporated into thiamin, HMP must be imported into the plastid from the cytosol, implying the existence of a plastidial HMP transporter (Figure 1). This transporter also remains to be identified.
In the present study, we identified a candidate transporter that is co-expressed with thiamin-related genes in Arabidopsis and whose prokaryotic homologs cluster with known thiamin biosynthetic genes. We show that expression of this transporter in Escherichia coli or Saccharomyces cerevisiae increases sensitivity to a toxic HMP analog, and that knocking out this transporter in an HMP-requiring Arabidopsis mutant results in a root growth defect when the level of HMP supplied is lowered.
Materials and methods
Nucleotide and amino acid sequences were from GenBank or SEED . A taxonomically diverse set of bacterial and archaeal genomes (>1000) were selected and analyzed using SEED tools. Transcriptional network analysis was done using GeneMANIA  restricting the data sources to ‘co-expression’ only and using the following genes as queries: At5g03555, At2g29630, At1g22940, At5g54770, At1g02880, At2g44750, At3g24030, At5g32470, At3g16990, At5g19460, At5g19470, At1g76730, At5g48970, At3g21390. The names and descriptions of these genes are given in Supplementary Table S1.
Reagents and chemicals were from New England Biolabs, Fisher Scientific or Sigma–Aldrich except for HMP ((4-amino-2-methyl-5-pyrimidinyl)methanol), oxy-HMP (oHMP; 5-(hydroxymethyl)-2-methylpyrimidin-4(1H)-one), which were purchased from Ark Pharm (Arlington Heights, IL) and [3H]thiamin hydrochloride (20 Ci/mmol, 1 mCi/ml in 1:1 ethanol:water) from American Radiolabeled Chemicals (Saint Louis, MO).
Synthesis of [3H]HMP
Briefly, [3H]thiamin (30 µCi, 1.5 nmol) was dried in vacuo and then hydrolyzed by Bacillus subtilis TenA_C in 250 µl as described previously . The reaction was then acidified by adding 25 µl of concentrated H3PO4 diluted to 500 µl with water and applied to Strata™-XL-AW column (Phenomenex, Belmont, CA) pre-washed with methanol, and pre-equilibrated with 100 mM potassium phosphate, pH 1.1. After sample application, the column was washed with 100 mM potassium phosphate, pH 1.1 and then methanol. HMP, HET, and unreacted thiamin were eluted with methanol:concentrated ammonium hydroxide (9:1) and dried in vacuo. The dried eluate was then dissolved in 20 µl of water and applied to a silica G60 F254 thin-layer chromatography (TLC) plate (0.1 mm thickness) (EMD Millipore), and developed with acetonitrile:water (4:1), pH 7.85. The radioactive HMP band (Rf 0.41) (well separated from thiamin, Rf 0.085 and HET, Rf 0.73) was scraped off, extracted thrice with water and lyophilized, yielding 0.56 nmol of [3H]HMP (4.45 µCi, 37% yield).
Expression constructs and strains
The pTAQ-PLUTO expression plasmid with the N-terminal signal peptide removed  was used without modification. The empty vector (EV) was obtained by digestion of pTAQ-PLUTO with KpnI and XhoI, blunted with E. coli T4 DNA polymerase and ligated with T4 DNA ligase, generating pTAQ-EV. These were then transformed into E. coli strain JD23420 lacking the endogenous uracil transporter uraA. In order to make a PLUTO S. cerevisiae expression plasmid similar to that described previously , PLUTO was amplified from pTAQ-PLUTO (forward primer, 5′–3′: ACTGCTCGAGAAAAAATGACCGGCTCAGAAATTAATG; reverse primer, 5′–3′: ACTGTCTAGATTACAAAAGCGGATGTGAAGA), digested with XhoI and XbaI and ligated into pRS1024  previously digested with XhoI and SpeI to generate pRS1024-PLUTO. The EV was obtained similarly to pTAQ-MAC-EV except that pRS1024 was digested with XhoI and SpeI, generating pRS1024-EV. The S. cerevisiae fur4 knockout strain Y03158 (MATα; ura3Δ0; leu2Δ0; his3Δ1; lys2Δ0; YBR021w::kanMX4; Euroscarf, Köhlerweg, Germany) was transformed with the full length ura3 amplified from S. cerevisiae strain S288C (GE Healthcare) (forward primer, 5′–3′: GAGTGAAACACAGGAAGACCAG; reverse primer, 5′–3′: GTTTTGTTCTTGGAAACGCTG) using the LiAc method  to generate a uracil prototroph. This strain was then transformed with either pRS1024-PLUTO or pRS1024-EV.
E. coli and S. cerevisiae growth experiments
E. coli was grown in M9-glucose medium supplemented with 50 µg/ml kanamycin at 37°C, with shaking at 220 rpm. S. cerevisiae was grown in synthetic complete medium without uracil and leucine at 30°C, with shaking at 250 rpm. Single colonies (three of each strain) were used to inoculate 2 ml of medium and grown until the OD600 reached 2–3. These were then used to inoculate, to an OD600 of 0.05, another 2 ml of media with or without oHMP and in the case of E. coli, with or without IPTG. These were then grown for 24 h, and their OD600 measured.
E. coli transport assays
E. coli harboring pTAQ-PLUTO or pTAQ-EV were grown and induced with IPTG as described previously for uracil transport assays  and resuspended to an OD600 of 10 in M9-glucose medium. Uptake assays were initiated by combining 500 µl of the cell suspension and 500 µl of M9-glucose containing [3H]HMP (25 nM, 0.1 nCi). Aliquots (50 µl) were removed immediately and after 2, 5, 10, and 20 min, passed through a pre-rinsed 0.45 µm cellulose nitrate filter (Whatman), and washed twice with 2 ml of M9-glucose medium. [3H]HMP uptake of the retained cells was then determined by liquid scintillation counting.
Arabidopsis PLUTO py double mutant lines
The homozygous PLUTO T-DNA Arabidopsis line WiscDsLox419Co3  (stock number CS854962) and the py-1 (ThiC) mutant  (stock number CS3491) were crossed. F1 seeds of the py-1 mutant line were grown on ½ Murashige and Skoog (MS) medium  plates supplemented with 2% (w/v) sucrose, 0.1% (w/v) MES, pH 5.7, and 0.6% (w/v) phytagel. Only plants containing the T-DNA insertion, verified using 3′ (5′–3′: GAATTTCCTTCCCTGTCCTTG) and 5′ (5′–3′: CCGGAGGGACTCAAAGAGTAC) gene-specific primers and a T-DNA-specific primer (p745; 5′–3′: AACGTCCGCAATGTGTTATTAAGTTGTC) were selfed and F2 seeds grown on the above medium supplemented with 100 µM thiamin. True-breeding lines were selected by verifying the py phenotype [4,7] on thiamin-free medium and the lack of a wild-type copy of PLUTO using the above primers.
Arabidopsis root growth experiments
Seeds were sown on plates of the above medium without thiamin, supplemented with various concentrations of HMP. They were then left in darkness at 4°C for 3 days before being placed vertically under fluorescent lights (Sylvania F40/CWP 40W cool-white plus, 100–150 µmol photons m−2 s−1) on a 12-h-light/12-h-dark cycle at 22°C. Images were captured 6 days after germination and root length was determined using ImageJ .
Prediction of PLUTO as an HMP transporter
Exploiting the fact that genes operating in the same pathway are often clustered on bacterial chromosomes , we used the SEED database and its tools  to search for uncharacterized transporter genes adjacent to known thiamin biosynthetic genes. This approach identified genes encoding a candidate transporter, CytX, which has also been assigned a putative role in HMP transport by others . The CytX-thiamin association is found in several different configurations and in at least four phyla, making it very robust  (Figure 2A). The presence of CytX in organisms that lack the HMP synthesis enzyme ThiC (Figure 2A) suggest that CytX transports HMP rather than HET or thiamin.
PLUTO and its prokaryote homolog CytX are associated with thiamin synthesis in prokaryotes and Arabidopsis
We used Pyrococcus furiosus CytX as a BLASTp query to identify an Arabidopsis homolog. PLUTO (At5g03555), a member of the nucleobase cation symporter 1 (NCS1) family, was the bi-directional best hit with 23% identity. This lies within the ‘twilight zone’ of protein sequence pair alignments (20–35%), within which function may be conserved . Additional support for PLUTO as a potential HMP transporter came from building a co-expression network with GeneMANIA  using PLUTO and the known thiamin biosynthetic and related transport and salvage genes. In a network containing six thiamin-related genes (Figure 2B), PLUTO connects directly to three thiamin-related genes (ThiC, COG0212, and At3g21390) and to a fourth (Thi1) through an unrelated gene (LCAT1). Other thiamin genes such as TPPK (At1g02880), TH1 (At1g22940), and TenA_E (At3g16990) also used as queries were not included in the network. Taken together, this genomic and transcriptomic evidence implicates PLUTO in thiamin metabolism and specifically in HMP transport.
Expression of PLUTO in E. coli or S. cerevisiae causes hypersensitivity to oxy-HMP
We chose to use PLUTO heterologously expressed in both E. coli and S. cerevisiae as a model because it has been reported to be active in these systems in the transport of nucleobases [8,9]. The S. cerevisiae strain was modified to be prototrophic for uracil to avoid competitive inhibition of PLUTO with supplied uracil. We used oxy-HMP (oHMP) as a substrate analog of HMP because it is thought to be incorporated into an inactive form of ThDP, resulting in toxicity and slower growth .
E. coli cells expressing PLUTO showed reduced growth in the presence of 1 mM oHMP, while uninduced cells or those harboring the EV showed no reduction of growth (Figure 3A). Similarly, S. cerevisiae expressing PLUTO had significantly less growth after 24 h in the presence of 3 mM oHMP, while no significant change was measured in the empty-vector control (Figure 3B). These data demonstrate that PLUTO can mediate uptake of oHMP.
Expression of PLUTO in E. coli or S. cerevisiae increases sensitivity to the toxic HMP analog oxy-HMP
[3H]HMP was synthesized and used in E. coli transport assays to attempt to confirm the HMP transport activity of PLUTO. However, due to high endogenous HMP transport activity, no difference between PLUTO-expressing cells and empty-vector controls was detected (Supplementary Fig. S1).
Arabidopsis py PLUTO double mutants have a distinct phenotype rescuable by HMP
Because a PLUTO knockout is able to synthesize HMP in the plastid, we created an Arabidopsis py PLUTO double knockout to test for reduced growth or a thiamin-deficient phenotype at various concentrations of exogenously supplemented HMP. The double mutant was not lethal at HMP concentrations needed to rescue the py mutant, suggesting that PLUTO is not the sole plastidial HMP transporter. As roots of various plants are known to have a limited capacity to synthesize thiamin [19–21], we checked for root growth phenotypes. In the absence of exogenous HMP, and at HMP concentrations below 10 nM, py PLUTO double mutant seedlings had significantly shorter roots than py single mutants (Figure 4 and Supplementary Fig. S2). Increasing the HMP concentration to 10 or 30 nM resulted in similar root growth of both mutants, indicating that PLUTO deficiency can be overcome by an excess of HMP.
Disruption of PLUTO in the HMP-requiring py mutant of Arabidopsis leads to a root growth defect that is rescued by supplementing with high concentrations of exogenous HMP
All the thiamin biosynthetic enzymes and several salvage enzymes have been identified in Arabidopsis, but the only transporters identified are mitochondrial ThDP transporters [1,2]. As HMP-requiring mutants can be grown and propagated when HMP is exogenously supplied [4,7], the presence of HMP transporters on the plasma membrane and the plastid envelope can be strongly inferred. Here we show that PLUTO acts as a plastidial HMP transporter, albeit a partially redundant one.
Prokaryotic homologs of PLUTO, previously predicted to be HMP transporters and given the name CytX , occur in clusters with thiamin biosynthetic enzymes in several bacterial phyla. Furthermore, CytX is found in organisms that almost certainly cannot make HMP because they lack the biosynthetic enzyme ThiC. Additionally, the gene specifying CytX is often found in a cluster encoding ThiD (HMP kinase), ThiE (ThMP synthase), ThiM (HET kinase), and in some cases a putative HET transporter, which – if CytX transports HMP – would together provide all the machinery needed to make ThMP. It should, however, be noted that CytX is not the only prokaryotic HMP transporter .
PLUTO is known to mediate plastidial nucleobase transport in Arabidopsis [8,9], but it is also co-expressed with genes involved in thiamin metabolism. Arabidopsis PLUTO has been shown to transport uracil [8,9] and PLUTO homologs from other organisms have been shown to transport cytosine ; both uracil and cytosine are pyrimidines and are thus chemically similar to HMP. Additionally, S. cerevisiae thi7 is structurally related to PLUTO  and transports thiamin .
Expressing PLUTO heterologously in E. coli or S. cerevisiae increased sensitivity to the HMP analog oHMP. oHMP is probably toxic because it can be phosphorylated by ThiD  and then converted to oxy-thiamin phosphates. However, it was shown not to be toxic in E. coli, likely due to lack of uptake . The lack of an effective endogenous transporter for oHMP enables PLUTO-expressing and wild-type E. coli to be distinguished by exposing them to 1 mM oHMP (Figure 3A).
We could not demonstrate transport of [3H]HMP in these same E. coli strains due to a background uptake of HMP (Supplementary Fig. S1), presumably mediated by unidentified specific HMP transporters that do not transport oHMP (and hence do not interfere with oHMP toxicity tests). Many prokaryotic vitamin or vitamin precursor transporters have substrate affinities in the high picomolar to low nanomolar range , which mirrors the typical concentrations of these compounds in natural environments [26,27]. PLUTO has a Km value in the micromolar range for uracil [8,9]. However, as cytosolic HMP concentrations are probably well above the trace levels found in the environment, PLUTO may not need a submicromolar affinity for HMP.
Arabidopsis PLUTO mutants show no thiamin deficiency symptoms , probably because the HMP moiety of thiamin is synthesized de novo in plastids (Figure 1), making HMP import dispensable . If PLUTO is the sole plastidial HMP transporter, a py PLUTO double mutant should be lethal when HMP is supplied (but viable if thiamin is supplied). That the double mutant is not lethal and has only a moderate root growth defect (Figure 4) when HMP is supplied indicates that other plastidial transporters can act on HMP. This defect is unlikely to be due to a depletion of other PLUTO substrates such as uracil because the phenotype can be rescued with exogenous HMP. That the root growth defect of py PLUTO double mutant seedlings is apparent when no HMP is supplied (Figure 4) most likely reflects the partial dependence of these seedlings on salvaged HMP, which can come only from the action of extraplastidial enzymes (At3g16990 and At5g32470) that hydrolyze thiamin breakdown products [2,7]. Thus, the root phenotype establishes that PLUTO is a significant HMP transporter, at least in roots. In this connection, it is interesting that roots of various species have long been known to require exogenous thiamin or thiamin precursors for growth in culture and, in planta, to import these compounds from shoots [19–21]. While HMP and HET synthesis genes are expressed in Arabidopsis roots, their expression is weak  and spatially restricted . This pattern is consistent with heavy reliance on HMP import into plastids to meet the demand for thiamin synthesis.
Finally, there is a clear parallel between our finding that the PLUTO plastidial nucleobase transporter doubles as an HMP transporter and the recent evidence that the Arabidopsis PUT3 polyamine transporter mediates phloem transport of both thiamin and polyamines . Perhaps other unidentified transporters for B vitamins and their precursors  will likewise prove to be known transporters with moonlighting activities.
JD23420 and pTAQ-PLUTO were generous gifts from Dr. Torsten Möhlmann and pRS1024 was a generous gift from Dr. Adrian Goldman. We thank Dr. Rémi Zallot for assistance in early stages of this work.
The authors declare that there are no competing interests associated with the manuscript.
This work was supported by the U.S. National Science Foundation [grant number IOS-1444202] and by an endowment from the C.V. Griffin Sr. Foundation.
ADH conceived and supervised the study. TSJ identified PLUTO and generated the py pluto double mutant. GAWB designed and performed the experiments, analyzed the data and wrote the paper.