Higher plant plastids and cyanobacteria have folate carriers related to those of trypanosomatids

Cyanobacterial and plant genomes encode proteins with some similarity to the folate and biopterin transporters of the trypanosomatid parasite Leishmania . The Synechocystis slr0642 gene product and its closest Arabidopsis homolog, the At2g32040 gene product, are representative examples. Both have 12 probable transmembrane domains, and the At2g32040 protein has a predicted chloro-plasttransitpeptide.Whenexpressedin EscherichiacolipabApabB or folE , mutants, which are unable to produce or take up folates, the slr0642 protein and a modified At2g32040 protein (truncated and fused to the N terminus of slr0642) enabled growth on 5-formyltet-rahydrofolate or after streaking, others after 3 days. Similar results were obtained when 11 (cid:3) M folic acid replaced 5-CHO-THF except that growth was slightly slower. B , strain P1–7B (a folE deletant) harboring pACYC-RIL was transformed with V , slr , and At-c constructs as above. Four independent clones of each construct were streaked on LB medium plus 0.1% (w/v) sodium ascorbate, 1 m M dithio- threitol,andIPTGalone( control )orcontaining11 (cid:3) M 5-CHO-THF.Plateswereincubated for2daysat37°C.

Tetrahydrofolate (THF) 2 and its derivatives (collectively termed folates) are essential cofactors for one-carbon transfer reactions in almost all organisms (1,2). Bacteria and plants synthesize folates de novo, but higher animals do not and so need a dietary supply (1,3,4). Folates consist of a pteridine, p-aminobenzoate (pABA) and one or more glutamate residues (Fig. 1A). All these moieties are synthesized and assembled into folates in the cytosol of bacteria, but in plants the pathway is split among three subcellular compartments. The pteridine is made in the cytosol, and the pABA is made in plastids; both then enter mitochondria, where they are coupled together and glutamylated to yield folate (5)(6)(7)(8)(9)(10)(11)(12) (Fig. 1B). Much of the folate produced exits the mitochondria to supply folate pools in the cytosol, plastids, and vacuoles (13)(14)(15)(16)(17). This compartmentation of the synthesis pathway and its end products entails various transmembrane fluxes of folates and their precursors (Fig. 1B). Although pABA can cross membranes by simple diffusion (18), folate and pteridine fluxes are likely to be carrier-mediated, as in animals (19 -22) and bacteria (23)(24)(25). Plants, therefore, presumably have several as yet undiscovered organellar folate or pteridine transporters (Fig. 1B).
Multiple transporters that act on folates as well as antifolates such as methotrexate or aminopterin have been characterized in mammals. These transporters include the reduced folate carrier, which catalyzes exchange with organic anions, folate receptors, which mediate folate entry via endocytosis, folate efflux pumps of the ATP binding cassette family, and a protein from the mitochondrial carrier family (19 -21).
High affinity folate and pteridine transporters are also known from the trypanosomatid Leishmania, a parasitic protist that is auxotrophic for folate and pteridine (26). These carriers are a biopterin transporter (BT1) and 2 folate transporters (FT1, FT5), all 3 being similar proteins with about 12 putative transmembrane helices (27)(28)(29). Together, they define a small family (the FBT family) that is a distant member of the major facilitator superfamily (30). BT1 is the only carrier yet known to act on pteridines; it also has low affinity folate transport activity (22).
Besides proteins from trypanosomatids, the FBT family (also known as the BT1 family, Pfam03092) includes a highly conserved protein from Cyanobacteria and nine Arabidopsis proteins. In this study we used heterologous expression in Escherichia coli mutants to probe the functions of a representative cyanobacterial FBT protein and its closest Arabidopsis homolog. Having demonstrated that both have folate and antifolate transport activity, we investigated the subcellular location of the Arabidopsis protein, the expression of its gene, and the effect of ablating this gene. cago, IL), and strain NM1200 was supplied by N. Majdalani (NCBI, Bethesda, MD). The NM1200 strain was constructed by P1 transduction of the Mini-Cm R prophage carrying the Red gam, beta, and exo genes under the P L promoter (31) in MG1655 (E. coli K12, E. coli Genetic Stock Center). The folE gene was deleted by targeted homologous recombination using electroporated linear DNA and strain NM1200 as described (32). Briefly, the kanamycin cassette was amplified from plasmid pUC4K with primers FolE-5Ј and FolE-3Ј (see Supplemental Table I), which respectively included the start and stop codon of folE and 47-bp 5Ј-and 3Ј-flanking sequences. After inducing the recombination system at 42°C for 15 min, strain NM1200 was transformed with the PCR product. Clones with the desired replacement of the folE gene were selected on LB plates containing 300 M thymidine and 30 g/ml kanamycin. The kanamycin resistance locus was then transferred by P1 transduction to strain K12, to give strain P1-7B. Deletion of the folE gene was confirmed by PCR and sequencing. P1-7B was maintained on LB medium containing 300 M thymidine and 30 g/ml kanamycin.
Plants and Growth Conditions-Arabidopsis (ecotype Columbia) plants were grown at 23-28°C in 12-h days (photosynthetic photon flux density 80 microeinsteins m Ϫ2 s Ϫ1 ) in potting soil irrigated with water. When roots were required, plants were grown hydroponically (33).
Expression Constructs-Synechocystis slr0642 and Arabidopsis At2g32040 amplicons preceded by a Shine-Dalgarno sequence were cloned between the NotI and SstI sites of pLOI707HE (34), which contains the tac promoter, the lacI q repressor gene, and the tet marker.
Folate Uptake Tests-BN1163 cells harboring pACYC-RIL and a pLOI707HE construct were streaked on minimal medium plates containing M9 salts (35), 0.1 mM CaCl 2 , 0.5 mM MgSO 4 , trace elements (36), and 0.4% (w/v) glucose supplemented with antibiotics as above and 0.5 mM IPTG. To this medium was added 3.6 M pABA or 11 M (6R,6S)-5-formyl-THF (5-CHO-THF), folic acid, or (6R,6S)-5-formyl-THF triglutamate. P1-7B cells harboring pACYC-RIL and a pLOI707HE construct were streaked on LB medium containing 0.1% (w/v) sodium ascorbate and 1 mM dithiothreitol, appropriate antibiotics and IPTG minus or plus 11 M 5-CHO-THF, dihydropteroate, hydroxymethyldihydropterin, or dihydropterin-6-aldehyde. Tests with the two latter compounds were made in a 10-l vessel flushed with N 2 at 25 ml/h. Antifolate Uptake Tests-Solid or liquid minimal media were supplemented with antibiotics, IPTG, and pABA as described above. For tests on solid medium, cells were streaked on plates containing a central 1-cm disk of Whatman No. 1 paper soaked with 10 l of dimethyl sulfoxide alone or containing 100 mM methotrexate or 80 mM aminopterin. For tests in liquid medium, cells from an overnight culture were inoculated to give an A 600 reading of 0.05 in 1.5 ml of fresh medium supplemented with various amounts of methotrexate or aminopterin. Growth was evaluated by measuring A 600 after incubating at 37°C for 9.5 h with shaking at 220 rpm.
Transient Expression of Green Fluorescent Fusion Protein in Arabidopsis Protoplasts-The entire At2g32040 sequence was amplified using primers At-gfp-5Ј and At-gfp-3Ј, digested with SalI and NcoI, and cloned in-frame upstream of the green fluorescent protein (GFP) gene in pTH2 (37). Protoplasts were prepared from cell cultures, transformed, and cultured as described at 22°C for up to 48 h (7, 38). Transformation efficiencies were typically 5-20%. Samples were analyzed by epifluorescence microscopy as described (11).
Chloroplast Import Assays-The full-length At2g32040 sequence and a version lacking the first 88 residues were, respectively, amplified using forward primer At-ifl-5Ј or At-it-5Ј and reverse primer At-i-3Ј. The amplicons were cloned as EcoRI-HindIII fragments into pGEM ® -4Z (Promega). Coupled in vitro transcription-translation was performed using a wheat germ TNT kit (Promega) and [ 3 H]leucine according to the manufacturer's protocol. Translation product import into pea chloroplasts was assayed as described (39). Briefly, translation product (50 l) was added to 50 l of 2ϫ import buffer (39) containing 100 mM A, folic acid. Aminopterin and methotrexate differ from folic acid in having an amino instead of an oxo group at C 4 , and N 10 is methylated in methotrexate. The pteridine ring of folic acid is fully oxidized; in natural folates it is reduced to dihydro or tetrahydro forms. A ␥-linked polyglutamyl tail of up to about six residues is usually attached to the first glutamate. B, distribution of folate biosynthesis steps between cytosol, mitochondria, and plastids and inferred carrier-mediated transport steps for folates (open symbols) and pteridines (gray symbol). Note that polyglutamylation of folates also occurs in the cytosol and plastids (7). DHN, dihydroneopterin; DHNTP, dihydroneopterin triphosphate; DHM, dihydromonapterin; HMDHP, hydroxymethyldihydropterin; DHP, dihydropteroate; DHF, dihydrofolate; ADC, aminodeoxychorismate; PP, pyrophosphate. unlabeled leucine; 100 l of 1ϫ import buffer containing 30 mM Mg 2ϩ -ATP and 200 l of purified chloroplasts (0.2 mg chlorophyll) were added, and samples were incubated for 15 min at 25°C in the light. Chloroplasts were re-purified as described (39), resuspended in SDS loading buffer containing 4 M urea, and held for 10 min at 37°C before loading.
Real-time Quantitative RT-PCR-Total RNA was extracted using RNeasy plant mini kits (Qiagen) and treated with DNase (DNA-free TM kit, Ambion, Austin, TX). RT-PCR was performed on 250 ng of RNA in 25-l reactions using Taq-Man One-Step RT-PCR Master Mix reagents and a GeneAmp 5700 system (Applied Biosystems, Foster City, CA). The primers (At-rt-5Ј, At-rt-3Ј) and Taq-Man probe (At-rt-probe) were designed with Applied Biosystems Primer Express software. The amplicon length was 74 bp. Primer At-rt-5Ј spanned two exons to avoid amplifying contaminating genomic DNA, and controls without reverse transcriptase were run to verify that no such amplification occurred. RT-PCR conditions were 48°C for 30 min and 95°C for 10 min followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. An At2g32040 RNA standard was synthesized from a cDNA template using primers At-SP6std-5Ј and At-std-3Ј and labeled using [5,6-3 H]UTP as described (10). An internal standard RNA was added to each sample before RT-PCR to estimate recoveries, which were 34 -100%. Data were corrected for recovery.
Insertional Mutant-An At2g32040 mutant (Salk 109024) was identified in the Salk Arabidopsis (ecotype Columbia) T-DNA insertion collection (40). Wild type or homozygous mutant segregants were identified by PCR using gene-specific primers located 5Ј or 3Ј of the T-DNA insertion (At-s-5Ј and At-s-3Ј, respectively) and the T-DNA-specific primer SalkLBb1. Genomic DNA was extracted as described (41). The insertion site was confirmed by sequencing the amplicon from mutant homozygotes. Homozygous mutant and wild type segregants were selfed, and their progeny were used for experiments. Kanamycin resistance was shown to co-segregate with the At2g32040 mutation, indicating the absence of T-DNA insertions at other loci.
Chloroplast Folate Analyses-Chloroplasts were prepared from the leaves of 4-week-old Arabidopsis plants as described (42,43). Folates were extracted from samples (ϳ1 mg of chlorophyll), deglutamylated, affinity-purified, and determined by HPLC with electrochemical detection as described (41).

FBT Family Genes in Cyanobacteria and Plants-Blast searches of
GenBank TM confirmed and extended the delineation of the FBT family given in the Pfam data base. Homologs of the Leishmania FT and BT1 transporters were found only in trypanosomatids, apicomplexans, Cyanobacteria, and plants. Cyanobacterial genomes have a single FBT gene specifying a highly conserved (Ն56% identity) protein. In contrast, plant genomes encode several diverse FBT proteins, there being nine in Arabidopsis. Multiple alignment and phylogenetic analysis showed that one of these, At2g32040, clustered with Cyanobacteria, whereas the others fell into two groups of four each (see Supplemental Fig. 1, A and C). The At2g32040 protein shares 52-54% identity with the cyanobacterial proteins but differs in having an ϳ90-residue N-terminal extension with the features of a plastid targeting peptide (44)  Because trypanosomatids and apicomplexans have apparently acquired numerous genes from photosynthetic organisms (45, 46) and plants acquired many genes from Cyanobacteria (47), the cyanobacte-rial protein is potentially the archetype of the FBT family. We accordingly chose a representative cyanobacterial FBT protein, Synechocystis slr0642 (henceforth slr0642) for further study along with its close plant relative, Arabidopsis At2g32040 (henceforth At2g32040).
Folate Uptake by E. coli Expressing slr0642 or At2g32040-Folates cannot normally cross the inner membrane of E. coli cells, preventing rescue of folate synthesis mutants by supplied folates (48). We, therefore, evaluated folate transport activity by expressing FBT constructs in a pabA pabB mutant, which is auxotrophic for pABA, and testing for growth on minimal medium with or without 5-CHO-THF or folic acid (the two most stable folates) in place of pABA. The mutant also harbored a plasmid bearing genes for three rare E. coli tRNAs (argU, ileY, leuW, corresponding to AT-rich codons); this plasmid was found to be needed to express At2g32040 constructs. The N-terminal extension that contains the predicted targeting peptide was removed from At2g32040 constructs, basing the truncation point (after residue 88) on alignments with slr0642 (see Supplemental Fig. 1C) and with orthologous plant sequences derived from ESTs (not shown). Because the truncated At2g32040 protein had a more hydrophilic N-terminal region than slr0642 and this region may affect membrane targeting or translation in E. coli (49), we also constructed a chimeric version in which residues 1-40 of the truncated protein were replaced by residues 1-37 of slr0642.
After two passages on minimal medium plus or minus folate, the growth of mutant cells harboring the empty pLOI707HE vector ceased ( Fig. 2A). In contrast, cells expressing slr0642 continued to grow when given folates ( Fig. 2A), showing that the latter are transported. The truncated At2g32040 did not confer the ability to grow on folates, but the chimeric version was effective ( Fig. 2A). Cells expressing slr0642 or chimeric At2g32040 did not grow when 5-CHO-THF triglutamate replaced the monoglutamyl form (not shown), indicating that both FIGURE 2. Folate uptake tests in E. coli expressing Synechocystis slr0642 or modified Arabidopsis At2g32040. A, strain BN1163 (a pabA pabB mutant) harboring pACYC-RIL was transformed with pLOI707HE with no insert (V) or containing slr0642 (slr) the truncated At2g32040 protein (At-t) or a chimeric At2g32040 protein with residues 1-128 replaced by residues 1-37 of slr0642 (At-c) (see Supplemental Fig. 1C). Six independent clones of each construct were streaked on minimal medium plus IPTG alone (control) or containing 3.6 M pABA or 11 M 5-CHO-THF. Clones were streaked once on pABAsupplemented plates and three times in succession on the others. Incubation was at 37°C. The pABA plates were photographed 1 day after streaking, others after 3 days. Similar results were obtained when 11 M folic acid replaced 5-CHO-THF except that growth was slightly slower. B, strain P1-7B (a folE deletant) harboring pACYC-RIL was transformed with V, slr, and At-c constructs as above. Four independent clones of each construct were streaked on LB medium plus 0.1% (w/v) sodium ascorbate, 1 mM dithiothreitol, and IPTG alone (control) or containing 11 M 5-CHO-THF. Plates were incubated for 2 days at 37°C. transporters, like most other folate carriers, prefer monoglutamates (note, however, that the triglutamate approaches in size the exclusion limit of the outer membrane porins of E. coli, which may also have impeded uptake). Cells cured of their plasmids were not rescued by folates, confirming that the transport trait is plasmid-encoded.
Although 5-CHO-THF and folic acid are relatively stable, some spontaneous breakdown is inevitable, releasing pABA-Glu. To rule out the possibility that the growth seen on folates was due to such release, we confirmed that pABA-Glu did not support growth of mutant cells expressing slr0642 or chimeric At2g32040 (not shown). Because E. coli cannot normally take up pABA-Glu but can utilize it once absorbed (50), this result further shows that neither protein is able to transport pABA-Glu.
The results from the pabA pabB mutant were corroborated using a second mutant, a folE deletant, which lacks the first enzyme of the pteridine branch of folate synthesis (GTP cyclohydrolase I). This deletant could grow on rich medium supplemented with 5-CHO-THF when expressing slr0642 or chimeric At2g32040 but not when harboring the empty vector (Fig. 2B). The folE deletant could not be used to test slr0642 or At2g32040 for transport of dihydropteroate or dihydropterins (hydroxymethyldihydropterin, dihydropterin-6-aldehyde) as these compounds supported growth of deletant cells harboring the empty vector (not shown).
Antifolate Uptake-The antifolates aminopterin and methotrexate are close structural analogs of folic acid, differing only in one and two groups, respectively (Fig. 1A). E. coli normally has little capacity to take up these compounds, and they are inhibitory only at high concentrations (50,51). Therefore, we tested antifolate transport activity by evaluating growth of cells exposed to antifolate gradients on plates (Fig. 3A) or to various levels of antifolates in liquid medium (Fig. 3, B and C). Expression of slr0642 or the chimeric At2g32040 construct imparted large increases in sensitivity that were qualitatively obvious on plates and were quantified as 30 -100-fold in liquid culture. The truncated At2g32040 protein was again ineffective.
Subcellular Localization of At2g32040-This was investigated first by transient expression of an At2g32040::GFP fusion construct in Arabidopsis protoplasts. Consistent with the presence of a putative plastid targeting peptide, expression of the fusion protein resulted in green fluorescence that colocalized with the red autofluorescence of chlorophyll; the green fluorescence was strongest at the chloroplast margins, suggesting that the protein is concentrated in the envelope (Fig.  4A). Expression of GFP alone gave green fluorescence throughout the cytoplasm (Fig. 4A).
Import assays with isolated pea chloroplasts substantiated these results (Fig. 4B). After the import reaction with the full-length protein, chloroplasts contained a labeled product that was smaller than the fulllength protein and was protected from attack by thermolysin, as expected for a translocated protein. In contrast, a truncated version lacking the predicted targeting peptide was not processed and was ther-   molysin-sensitive, indicating that it had not been taken up. The latter protein (calculated molecular mass 51 kDa) behaved in SDS-PAGE as a 51-kDa species as expected, but the full-length protein (calculated mass 61 kDa) migrated as a 72-kDa species (Fig. 4B)). Other FBT family proteins likewise migrate anomalously (52). Thus, although the migration of the imported protein as a 67-kDa band (Fig. 4B) implies that the point we chose to truncate At2g32040 is not identical to the normal processing site, electrophoretic anomalies preclude quantifying the discrepancy.
Expression of At2g32040 in Arabidopsis Organs-Quantitative RT-PCR analysis indicated that the At2g32040 gene is constitutively expressed throughout the plant at a low level (sequencing of the RT-PCR product obtained from root and young leaf RNA confirmed that it corresponded to At2g32040 mRNA). Thus, the abundance of At2g32040 transcripts varied rather little among major organs, from 5 to 14 ϫ 10 4 copies/250 ng of total RNA (Fig. 5A), which corresponds to a frequency between 1 in 20,000 and 1 in 60,000 mRNAs (assuming that mRNA is 1% of total RNA and has an average size of 1.5 kb). Ubiquitous, low level expression is also reported for At2g32040 in the GENEVES-TIGATOR Arabidopsis microarray data base (53). The presence of At2g32040 transcripts in roots as well as aerial organs implies that the protein is expressed in non-green plastids as well as chloroplasts.
Identification and Analysis of a T-DNA Insertion Mutant-To explore the function of At2g32040 in vivo, an Arabidopsis line bearing a T-DNA insertion in the coding region of the gene was identified in the Salk collection (40). This mutant has a single T-DNA insertion in exon 1 at codon 24 (Fig. 5, B and C), as determined from the flanking DNA sequence and co-segregation of kanamycin resistance with the T-DNA. The insertion creates at least two in-frame stop codons and, therefore, is predicted to result in a complete knock-out.
Homozygous mutant plants did not differ noticeably from their wild type siblings in growth rate, morphology, leaf color, or fertility. Because At2g32040 is located in the plastid (Fig. 4), we compared the folate profiles of chloroplasts from wild type and mutant plants (Fig. 6). As in pea (15)(16)(17), total folate levels in wild type chloroplasts were low, about 50 pmol/mg of protein or 150 pmol/mg of chlorophyll, which is equivalent to a stromal concentration of ϳ2 M (taking stromal volume as 66 l/mg chlorophyll (54)). The most prominent folate forms were 5-methyl-and 5,10-methenyl-THF (the acidic HPLC mobile phase used in the analysis converts 10-formyl-THF to 5,10-methenyl-THF, so that the 5,10-methenyl-THF measured includes 10-formyl-THF plus any preexisting 5,10-methenyl-THF). Relative to wild type, mutant chloroplasts contained significantly (p Ͻ 0.05) more 5,10-methenyl-THF and  total folate per unit protein (Fig. 6A) and, when data were expressed on a percent basis, a significantly lower proportion of 5-methyl-THF (Fig. 6B).

DISCUSSION
We report here the functional expression of an FBT family protein in E. coli, an advance that should help studies of transport mechanisms. Expressing the Synechocystis slr0642 protein in E. coli was straightforward, but success with Arabidopsis At2g32040 required the presence of a plasmid bearing rare AT-rich codons and modification of the protein.
Deleting its N-terminal extension did not suffice; it was necessary also to replace the N terminus of the truncated protein with the corresponding sequence from slr0642, a region that may affect membrane insertion or translation in E. coli. This modification altered only the initial part (the first third) of the predicted transmembrane helix number 1, which part is not involved in substrate binding in other major facilitator superfamily proteins (55,56).
As shown in Supplemental Fig. 1C, there is a remarkably high sequence conservation between the cyanobacterial carrier and At2g32040 that extends over the entire protein except the region between transmembrane helices 6 and 7, a large loop that separates the two halves of the molecule. The residues that are conserved both in these proteins and the Leishmania FT1, FT5, and BT1 carriers are likewise spread throughout the sequence except for helices 6 and 7 and the region between them (see Supplemental Fig. 1C). In this interstitial region the Leishmania sequences are predicted to have two additional transmembrane ␣-helices, an uncommon configuration in the major facilitator superfamily (57).
Folate transport was evaluated via bacterial growth, which is a sensitive way to detect transport activity but not to assess its rate or substrate specificity. For example, growth on 5-CHO-THF or folic acid lagged behind that on pABA, but this does not necessarily imply that the uptake rate of either folate limited growth. Both 5-CHO-THF and folic acid need enzymatic conversion to metabolically active forms (5,10-methenyl-THF and THF, respectively) before use (1, 58) so conversion rates could have been limiting. Moreover, 5-CHO-THF inhibits folate-dependent enzymes (58) and may have slowed growth on this account. Respecting substrate specificity, growth tests could not show whether or not slr0642 and At2g32040 act on dihydropteridines because E. coli cells took such compounds up unaided. These proteins, thus, remain unconfirmed candidates for the pteridine transport activities observed or inferred in Cyanobacteria and chloroplasts (59 -61).
Although E. coli is suspected to have a cryptic gene (abgT), not normally expressed, that can allow transport of folate analogs and perhaps folates (51), three arguments exclude activation and selection of this or any other E. coli gene as an explanation for our data. First, curing transport-proficient cells of the expression plasmid abolished folate transport capacity. Second, the antifolate uptake results of Fig. 3 could not have involved selection because transport-proficient cells were ipso facto unable to divide. Third, when cells harboring empty expression vector underwent prolonged incubation on folate media, the growth of colonies was far too infrequent to compromise the folate uptake experiments of Fig. 2, where all cells expressing an active FBT carrier grew.
The subcellular localization and expression pattern of At2g32040 imply that it has a housekeeping function in green and non-green plastids alike. That this function is folate transport is supported by the finding that ablating At2g32040 caused a 23% rise in chloroplast total folate and a 34% fall in the proportion of 5-methyl-THF. That these changes were not more marked implies that there are other chloroplast folate transporters, and there is indeed recent evidence for a second one in Arabidopsis (62).
Because At2g32040 acts on 5-CHO-THF, folic acid, and antifolates, it is clearly not highly specific and so can presumably transport the monoglutamyl form of any folate found in plastids. In this connection it is noteworthy that plastid folate transport fluxes include not only net import but also an exchange with the cytosol driven by methionine synthesis. Plastids synthesize homocysteine and contain an isoform of methionine synthase, which converts homocysteine to methionine using 5-methyl-THF as methyl donor (63). However, plastids cannot make 5-methyl-THF due to lack of the enzyme that produces it from 5,10-methylene-THF and so must import it from the cytosol (64) in exchange (not necessarily via the same carrier) for other folates in a shuttle reaction. Because stoichiometric, not catalytic, amounts of 5-methyl-THF are needed for methionine synthesis, this methyl group shuttle must greatly exceed the net import of folates into plastids.
Finally, this work establishes conservation of folate transport function among FBT family proteins from species representing three kingdoms of life, monera (Synechocystis), plants (Arabidopsis), and protists (Leishmania). Given the evidence for lateral transfer of genes from Cyanobacteria to plants (47) and from Cyanobacteria or plants to trypanosomatids and apicomplexans (45,46), it seems likely that the FBT family arose in Cyanobacteria and passed to the other groups, the evolutionary driver being the selective advantage conferred by folate transport. Whatever the case, these folate carriers that function across kingdom boundaries may offer new options for engineering folate production in bacteria and plants (1,65). Such carriers might be used, for example, to enhance folate export from bacteria or folate sequestration in storage compartments in plant cells.