Agp2, a Member of the Yeast Amino Acid Permease Family, Positively Regulates Polyamine Transport at the Transcriptional Level

Agp2 is a plasma membrane protein of the Saccharomyces cerevisiae amino acid transporter family, involved in high-affinity uptake of various substrates including L-carnitine and polyamines. The discovery of two high affinity polyamine permeases, Dur3 and Sam3, prompted us to investigate whether Agp2 directly transports polyamines or acts instead as a regulator. Herein, we show that neither dur3Δ nor sam3Δ single mutant is defective in polyamine transport, while the dur3Δ sam3Δ double mutant exhibits a sharp decrease in polyamine uptake and an increased resistance to polyamine toxicity similar to the agp2Δ mutant. Studies of Agp2 localization indicate that in the double mutant dur3Δ sam3Δ, Agp2-GFP remains plasma membrane-localized, even though transport of polyamines is strongly reduced. We further demonstrate that Agp2 controls the expression of several transporter genes including DUR3 and SAM3, the carnitine transporter HNM1 and several hexose, nucleoside and vitamin permease genes, in addition to SKY1 encoding a SR kinase that positively regulates low-affinity polyamine uptake. Furthermore, gene expression analysis clearly suggests that Agp2 is a strong positive regulator of additional biological processes. Collectively, our data suggest that Agp2 might respond to environmental cues and thus regulate the expression of several genes including those involved in polyamine transport.


Introduction
Agp2 is a 67.2-kDa plasma membrane protein that contains 12 potential transmembrane domains and belongs to the yeast Saccharomyces cerevisiae amino acid permease family that also includes many other members such as Put4, Alp1, Lyp1, Can1 and Gap1 [1,2]. Agp2 was initially shown to be involved in the plasma membrane transport of L-carnitine into yeast cells [3]. Lcarnitine serves as a carrier to transport acetyl-CoA generated by fatty acid b-oxidation via the peroxisomal-mitochondrial carnitine acetyltransferase (Cat2) shuttle for complete oxidation by the Krebs cycle [3]. AGP2 was subsequently shown to be negatively regulated by the high osmolarity glycerol response 1 gene (HOG1), encoding a MAP kinase, whereby under high osmolarity conditions AGP2 expression was downregulated [4]. Additional studies identified a function for Agp2 as a low-affinity, non-specific amino acid permease in poor nutrient conditions [1]. However, its low level of expression, its low affinity and specificity as well as its redundant function with that of several more active amino acid permeases, including Gap1, raised important questions about the relevance of a direct role for Agp2 in amino acid transport.
Using a high-throughput screen, we later discovered that single deletion of AGP2, similar to four other genes (PTK2, SKY1, FES1, and BRP1), causes enhanced resistance to the anticancer drug bleomycin-A5 (BLM-A5), a spermidine-conjugated bleomycin species which acts via DNA damage induction [5]. We further reported that Agp2 mediates high-affinity polyamine transport, using spermine and spermidine with a higher affinity than putrescine as substrates, a specificity pattern corresponding to that observed in wild-type cells [6]. This is supported by the fact that agp2-defective mutants exhibit a dramatically reduced initial velocity of spermidine and spermine uptake, and a very strong resistance to the toxicity of spermine and its analogs [6]. Introduction of a plasmid overexpressing AGP2 restored polyamine toxicity to the agp2-defective mutants [6]. Based on these observations, we reached the conclusion that AGP2, which has all the conserved features of amino acid permeases of its family, was the first eukaryotic gene identified as encoding a polyamine permease in yeast [6]. In fact, van Roermund et al. had identified Agp2 as a carnitine permease based on the same type of evidence as that used by our group to assign a direct role in polyamine transport to its gene product [3]. However, we were puzzled by preliminary observations indicating that addition of high concentrations of L-carnitine (up to a 1000-fold excess) did not compete for the uptake of radiolabeled polyamines nor prevented spermine toxicity [6] (see below). To account for these observations, we proposed various hypotheses for the actual transport function(s) of Agp2 that include (A) that Agp2 possesses two independent binding sites, specific for L-carnitine or polyamines, respectively, and thus acts as a polyspecific transporter accepting substrates with different sizes and molecular structures, (B) a role of Agp2 as a non transporting transceptor, i.e., a transporter which has lost the ability to transport any ligands and that may thus act as a sensor to regulate the expression of L-carnitine and polyamine permeases encoded by other, as yet undescribed genes, and (C) acting as a transporting transceptor of polyamines in a signal transduction pathway for the expression of other transporters [6,7].
In S. cerevisiae, several non transporting transceptors (or sensors) have been previously described such as Ssy1, a plasma membrane protein belonging like AGP2 to the yeast amino acid permease family, but which is endowed with no or only limited transport function [8]. Ssy1 senses amino acid availability by direct interaction with extracellular amino acids and triggers the expression of several downstream target genes, that encode amino acid permeases such as AGP1, BAP2, BAP3, DIP5, and TAT1 via the formation of an intermediary complex, the SPS sensor, with the plasma membrane proteins Ptr3 and Ssy5 [8,9]. In this sensory system, Ptr3 becomes hyperphosphorylated allowing the activation of the protease Ssy5, which cleaves the latent form of the two homologous zinc-finger transcription factors, Stp1 and Stp2, present in the cytosol, which then enter the nucleus and induce the Ssy1-targets [10]. Other examples include transporting transceptors like Gap1 and Mep2, which not only act as a general amino acid transporter and ammonium permease, respectively, but also as a nitrogen sensor that signals through the rapid activation of protein kinase A (PKA) targets [11,12]. Thus, in view of the increasing number of sensors that are currently being identified, and since a permease with independent binding sites for similar substrates (i.e. model A above) is highly implausible, we assess here the hypothesis whether Agp2 may instead possess regulatory function. A regulatory model for Agp2 became even more plausible with the report by Uemura et al., that Dur3 and Sam3 are the major high-affinity permeases accounting for polyamine uptake in yeast [13]. This is supported by the fact that agp2defective mutants exhibit a dramatically reduced initial velocity of spermidine and spermine uptake, and a very strong resistance to the toxicity of spermine and its analogs [6]. In this study, we now document that the agp2D single mutant and the dur3D sam3D double mutant display similar phenotypes towards polyamines, in contrast to the dur3D or sam3D single mutants. Using a variety of approaches that included RT-PCR, microarray and gene expression profile analyzer (GeXP), we also demonstrate for the first time that in cells lacking a functional AGP2 gene nearly 172 genes are downregulated as compared to the parent. Amongst the downregulated genes, many encode membrane transporters and these include DUR3, SAM3, HNM1, TPO2 and HXT3. Moreover, we have identified the expression of SKY1, a SR kinase that positively regulates low-affinity polyamine transport via its indirect effect on membrane potential, as a target of Agp2 [13,14,15,16,17,18]. We propose that Agp2 mainly acts as a regulator of polyamine transport as well as other specific biological functions.

Spot Test Analysis of Cell Growth
Standard spot tests were performed as previously described [24].

Microscopy
For fluorescent localization studies, imaging was carried out with an Olympus FV1000 confocal laser scanning biological microscope, and images were processed with the accompanying software, FV10-ASW 2.0. Cells were in some cases stained with DAPI (2.5 mg/ml final concentration; Roche) for simultaneous localization of nuclei.

RNA extraction and reverse transcription (RT)-PCR analysis
Total RNA was prepared using the RiboPure-Yeast extraction kit (Ambion) from 3 ml of an overnight suspension culture and treated with TURBO DNA-free kit (Ambion) to eliminate genomic DNA contamination. Total RNA was quantified by measuring A 260 spectrophotometrically. The cDNA was synthesized using M-MLV reverse transcriptase from Invitrogen. The PCR primers used to amplify the target genes are listed in Table 2. The PCR program was 2 min at 95uC followed by 25 cycles of 1 min of denaturation at 94uC, 2 min of annealing at different temperature dependent of the gene fragment (e.g. 54uC for ACT1 fragment) and 4 min of primer extension at 72uC. This was followed by extension at 72uC for 7 min. The identity of PCR products was verified on a 1% agarose gel.

Microarray analysis
Total RNA was isolated from exponentially growing cultures of the parent and the agp2D mutant using the RiboPure kit above. RNA samples were then hybridized to a GeneChip Yeast Genome 2.0 array that includes approximately 5,744 probe sets for 5,841 of the 5,845 genes in S. cerevisiae (Affymetrix, Santa Clara, CA). Probe sets on the array include 11 oligonucleotide pairs to detect each transcript. Microarray was performed and analyzed using the services of Genome Quebec (The Innovation Centre, McGill University). Densitometry readings were exported to a project file and statistical analysis of the datasets was performed using the FlexArrayH program developed by Genome Quebec which embeds the RH statistical software. Raw data were first normalized using robust multi-array averages and statistical comparisons between the two sets of RNA-DNA chip hybrids were performed with two-sample Bayesian t-tests including a Bonferroni correction [25,26]. Genes that were significantly downregulated by $2-fold or upregulated by $2-fold can be found in Supplemental Tables S1 and S2, respectively, which include the gene name and a brief functional description, as well as the t-test and P values. The threshold for significant differences between the parent and agp2D mutant was set at P,0.001 for better stringency.

Multiplex analysis by GeXP system
The procedure for sample preparation and subsequent gene expression analysis was performed according to manufacturer's instructions using the GenomeLab TM Gene eXpression Profiler (GeXP; Beckman Coulter, Canada) and as previously described [27]. Briefly, the primers used for GeXP are listed in Table 3 and were designed using the eXpress Designer module of the express Profiler software (Beckman Coulter). A total of 25 ng RNA was used in a 20 mL reaction volume for RT. Kanamycin RNA, an internal positive control was included. The RT reaction was performed using the GenomeLab TM GeXP Start Kit (Beckman Coulter) under the conditions: 1 min at 48uC, 60 min at 42uC, 5 min at 95uC, hold at 4uC, in a thermal cycler. After the RT, a PCR was performed under the conditions: 10 min at 95uC, followed by 35 cycles of 30 s at 94uC, 30 s at 55uC, and 1 min at 70uC. The data was normalized to Kanamycin before being expressed as a log 2 ratio.

Ontological Analysis of the Effects of AGP2 Inactivation on Gene Expression
Considering the large number of genes whose expression was found to be strongly affected upon AGP2 gene inactivation, we carried out a detailed analysis of the ontologic relationships between the genes found to be either downregulated or upregulated in agp2 D mutants from the Affymetrix microarray experiments. For that purpose, we relied on the Gene Ontology enRIchment anaLysis and visuaLizAtion (GOrilla) application (http:// cbl-gorilla.cs.technion.ac.il) [28] which presents the advantage of rigorous and highly reliable statistical criteria for gene ontology (GO) group assignments and graphic interface.

Determination of Spermidine and Carnitine Uptake
Prior to polyamine or carnitine uptake assay, cells were grown to the mid-logarithmic phase, washed three times in uptake buffer A (50 mM sodium citrate, pH 5.5, 2% D-glucose), and resuspended in 100 ml of the same buffer at 2610 7 cells/ml. The uptake assay was initiated by the addition of either [ 14 C]-spermidine or L-[methyl-14 C]-carnitine hydrochloride followed by incubation in a water bath at 30uC with shaking. The reaction was stopped at predetermined intervals by adding 1 ml of ice-cold uptake buffer. Cells were washed three times with uptake buffer and resuspended in 100 ml of this buffer. Five ml of scintillation mixture (Amersham Biosciences) were added to each sample, and the retained radioactivity was determined by liquid scintillation spectrometry.

Results
Simultaneous null mutation of SAM3 and DUR3 is needed to mimic the defective polyamine transport phenotype of agp2D cells In 2007, Dur3 and Sam3, belonging to the sodium solute symport family and the amino acid/polyamine/organocation (APC) family, respectively, were reported as the main high-affinity polyamine permeases in S. cerevisiae [13,29]. Although Dur3 and Sam3 virtually account for the total high-affinity influx of polyamines, all previous evidence along with its permease-like structure, point to Agp2 as a major factor promoting both Lcarnitine and polyamine transport [3,6]. In spite of its dual effect on L-carnitine and polyamine transport, a 1000-fold excess of Lcarnitine did not compete with the cytotoxic effect of spermine on a wild type yeast strain (Fig. S1). This excludes that Agp2 acts as a transporter with a single binding site for L-carnitine and polyamines, a possibility already considered unlikely in terms of the structure of H + -coupled symporters, which normally bear a single binding site for related substrates [2,30]. An alternative possibility is that Agp2 harbors a sensing function similar to that of Ssy1, which would regulate the expression of L-carnitine and polyamine transporters encoded by other distinct genes, without directly acting as a major influx pathway for substrate entry [30].
On the basis of the aforementioned possible role of Agp2, we tested whether either dur3D or sam3D single mutants were resistant to the toxicity of high spermine concentrations using spot test analysis. As shown in Figure 1A, neither dur3D nor sam3D single mutants showed enhanced resistance to spermidine or spermine toxicity or to the spermidine derivative bleomycin-A5 (BLM-A5) when compared to the parental strain. The agp2D mutant was used as a positive control strain which displayed striking resistance to polyamines and BLM-A5 (Fig. 1A) [6]. None of the mutants showed any differential sensitivity to 4-nitroquinoline-1-oxide (4-NQO), a DNA-damaging agent without structural relationship with polyamines, which was used as a control for BLM-A5, indicating that BLM-A5 likely acts as a polyamine analogue. On the basis of these findings, we predicted that the dur3D and sam3D single mutants would show parental levels of polyamine uptake. As expected, the agp2D mutant was completely defective in [ 14 C]spermidine uptake, whereas the dur3D and sam3D single mutants showed parental levels of spermidine uptake (Fig. 1B). The latter observation is in disagreement with the previous results reported by Uemura et al., i.e. that mutants disrupted for either the DUR3 or SAM3 genes displayed a lower velocity of spermidine uptake than the parental strain under similar uptake conditions with exponential cells [13] (see discussion).
Since Dur3 and Sam3 have been described as largely redundant high-affinity polyamine transporters, but could be deleted singly without affecting apparent spermidine uptake, we next examined the polyamine transport and toxicity phenotypes of dur3D sam3D double mutants. The dur3D sam3D cells obtained from a BY4741 parental background grew slightly slower in YPD media, unlike either dur3D or sam3D single mutants (data not shown). We observed that the dur3D sam3D double mutant was distinctly more resistant to polyamine toxicity than parental cells, but its tolerance was not nearly as pronounced as that found in agp2D cells upon growth on solid YPD medium containing spermine (Fig. 1C). The stronger tolerance to spermine toxicity observed in agp2D cells relative to the dur3D sam3D strain was independently confirmed by transferring exponentially growing cells to liquid YPD medium containing 1 mM spermine and monitoring the OD 600 at various times. Indeed, under these conditions, the double mutant showed no growth defect and was less tolerant to spermine toxicity when compared to the agp2D mutants (Fig. 1D). This suggests that in addition to the high-affinity polyamine carriers Dur3 and Sam3, yeast cells also contain one or more low-affinity polyamine transporters. We next verified that the enhanced tolerance of dur3D sam3D cells towards spermine cytotoxicity was related to a defect in polyamine uptake. As shown in Figure 1E, the dur3D sam3D double mutant was completely defective in [ 14 C]-spermidine uptake at a low substrate concentration (i.e. the high-affinity component, measured at 10 mM substrate), as compared to parental cells (see Discussion). Importantly, the defect in highaffinity spermidine uptake detected in the dur3D sam3D mutant was as pronounced as that found in the agp2D mutant (Fig. 1E). Taken together, these findings are consistent with the previous demonstration that Dur3 and Sam3 are high-affinity polyamine permeases [13] and reinforced the notion that Agp2 might act to regulate these two high-affinity transporters, as well as govern low-affinity polyamine transport. It is noteworthy that AGP2 expression level was not detectably altered in the dur3D sam3D double mutant, as determined by RT-PCR (Fig. 1F), excluding the possibility that this double mutant influences AGP2 expression.
Agp2 localizes at the plasma membrane in the absence of polyamine transporting activity The lack of high-affinity polyamine transport in dur3D sam3D prompted us to analyze Agp2 subcellular localization in this mutant. We investigated the localization of Agp2 by using a GFP C-terminally tagged version of the gene expressed under its own promoter, from the previously described plasmid pAGP2-GFP [6]. This plasmid expresses a functionally active fusion protein, which we have previously shown to rescue the polyamine uptake defect of the agp2D mutant. Cells growing exponentially in selective minimal medium showed a heterogeneous distribution of Agp2-GFP between plasma membrane, ER and endosomes ( Fig. 2A). Under these conditions, Agp2-GFP partially colocalized with the mRFP-tagged ER translocation subcomplex subunit protein Sec66, which shows localization in cortical ER and nuclear envelope [23]. Partial colocalization of Agp2-GFP was also observed with the DsRed-tagged FYVE domain (a phosphatidylinositol 3-phosphate-binding domain found on endosomal membranes) [22] (Fig. 2A). The existence of intracellular pools of Agp2 located in different subcellular compartments is consistent with the known distribution of proteins possessing both transporter and sensor functions [31,32]. This form of distribution may indicate that the activity of Agp2 is tightly regulated at the posttranslational level by intracellular sorting, i.e., recycling or endocytosis, for rapid adaptation to environmental changes in the nutrient conditions.
Since AGP2 is normally expressed at the transcriptional level in the dur3D sam3D mutant, the absence of polyamine transport in this mutant strongly suggests that Agp2 lacks the ability to directly transport polyamines as a plasma membrane transporter. Alternatively, the possibility remains that in the dur3D sam3D mutant Agp2 does not accumulate at the plasma membrane. In order to discern between these two possibilities, we monitored the localization of Agp2-GFP in the agp2D dur3D sam3D triple mutant (Fig. 2B). Our results showed that Agp2-GFP still accumulated at the plasma membrane of the agp2D dur3D sam3D cells growing exponentially in selective minimal medium. Under this condition, Agp2-GFP did not sensitize the dur3D sam3D double mutant to spermine or spermidine (data not shown), again excluding the possibility that Agp2 possesses the ability to directly transport polyamines. These findings, therefore, support the hypothesis that Agp2 might act as a plasma membrane polyamine sensor/ regulator. Addition of polyamines (shown in Fig. 2B for the sublethal concentration of spermidine, 5 mM) did not cause apparent changes in the localization of Agp2-GFP, in either the agp2D or agp2D dur3D sam3D mutants, at least during an incubation period of up to 4 h (similar results were obtained with putrescine 5 mM, and spermine 0.05 mM; data not shown). However, Agp2-GFP plasma membrane localization was strongly dependent on the presence of an external nitrogen source. Lack of nitrogen caused vacuolar sorting of Agp2-GFP in both agp2D and agp2D dur3D sam3D mutants. Part of this vacuolar sorting is due to endocytosis of Agp2-GFP, since in the endocytosis deficient strain agp2D end3D, Agp2-GFP accumulated at the plasma membrane of nitrogen-starved cells (Fig. 2C).

Expression of the genes DUR3, SAM3 and HNM1 is downregulated in the agp2D mutant
To address the hypothesis that Agp2 behaves as a regulator and controls the expression of the two high-affinity polyamine permease genes DUR3 and SAM3, we performed RT-PCR analysis of total RNA isolated from the agp2D mutant and the parental strain. While the expression level of a control gene (ACT1) was identical in both parental and agp2D cells, DUR3 and SAM3 expression was considerably reduced in the agp2D mutant, with DUR3 being the most profoundly affected (Fig. 3A). Since Agp2 was originally described as a L-carnitine transporter [3], we assessed whether its strongly positive action on L-carnitine uptake might in fact be explained by an indirect mechanism, i.e. by upregulation of a different transporter gene. We predicted that the choline permease Hnm1 might also act as an L-carnitine transporter on the basis of the similar quaternary amine structures of L-carnitine and choline, and as also suggested by an earlier report that L-carnitine is an effective competitive inhibitor of choline transport [33]. As shown in Figure 3B, the velocity of L-  uptake was decreased by 40-50% in hnm1D mutant cells, as compared to the control strain, whereas agp2D cells were completely deficient in L-carnitine transport activity, confirming a previous report [3]. These data suggest for the first time that Hnm1 is a significant pathway for L-carnitine import, although another Agp2-dependent uptake mechanism must be invoked to fully account for L-carnitine transport activity (see Discussion). Consistent with a biologically significant role for Hnm1 in Lcarnitine uptake, HNM1 expression was barely detectable in the agp2D mutant (Fig. 3A). This confirms that a substantial fraction of L-carnitine is transported by Hnm1 into the cells and that the expression of this permease is tightly controlled by Agp2. Taken together, these findings strongly suggest that Agp2 acts on polyamine and L-carnitine transport via an indirect mechanism, i.e., through regulation of expression of at least three permease genes, namely DUR3 and SAM3 in the case of high-affinity polyamine transport, and HNM1 for L-carnitine transport. This regulation seems to share at least one similarity to the mode of action described for the amino acid Ssy1 and the glucose Snf3-Rgt2 sensors, namely the ability to regulate downstream targets at the transcriptional level [9,34,35].

Microarray analysis reveals that Agp2 is a major regulator of gene expression
To investigate whether the postulated regulatory role of Agp2 extends to genes other than polyamine and L-carnitine permease genes, we performed three independent microarray analyses from total mRNA derived from the wild-type parent and agp2D mutant strains (see Materials and Methods). The inactivation of AGP2 caused either the down-or upregulation of a much larger number of genes than anticipated (Table S1 and S2). Ontogeny analysis identified four major clusters of high statistical significance that were the most strongly depressed upon AGP2 inactivation with a very strong bias (P = 7.5610 28 ) towards transmembrane transport. Thus, in addition to DUR3, SAM3 and HNM1, a large number of other transporter genes including at least four hexose permeases (HXT1, HXT2, HXT3, and HXT13) which function in maintaining a steep inwardly directed D-glucose gradient, as well as TPO2, which encodes one of the four plasma membrane drug/H + exchangers of the major facilitator superfamily that excrete polyamines in yeast, were downregulated in agp2D mutants (Table 4 and Table S3). The three other clusters of downregulated genes represent those involved in (i) mating, cell adhesion and agglutination, several of which are under mating factor control and are essential components of the pheromone response pathway, (ii) vitamin and cofactor biosynthesis, and (iii) sulfur assimilation and methionine biosynthesis (Table 4 and Table S3).
In a similar ontology analysis for the genes that were negatively regulated by AGP2, we identified three general functional classes consisting of genes involved in (i) lipid/fatty acid oxidation and mitochondrial oxidative metabolism (Krebs cycle and respiration), (ii) plasma and peroxisomal membrane transport and (iii) meiosis and sexual spore formation (Table 5 and Table S4). Thus, Agp2 seems to influence the expression of specific subgroups of genes.
To further confirm that Agp2 plays a major role in the regulation of gene expression, we measured mRNA levels for a set of genes identified from the microarray analysis using two independent approaches. In the first approach, we used a multiplex quantitative assay, namely the Gene eXpression Profiler (GeXP), to confirm the differences in the expression of 24 genes, which included both targets and non-targets, between agp2D mutants and the parental strain in a single reaction. This strategy utilizes RT-PCR and PCR followed by gene product analysis via capillary electrophoresis (see ''Materials and Methods'') [27]. As shown in Figure 4, the key target genes SKY1, SAM3, DUR3, and HNM1 were downregulated in the agp2D mutant as compared to the parent. In contrast, the non-target genes such as ACT1, FES1, AGP1, PDR13 and ZOU1 were not affected in the agp2D mutant (Fig. 4).
In the second approach, we used RT-PCR and examined the expression pattern of a few individual genes. As shown in Figure 5A, the target genes SKY1, FIG1, TPO2, and HXT3 were downregulated in the agp2D mutant, as compared to the control gene ACT1. In contrast, the non-target transporter genes AGP1 and GAP1 were not affected in agp2D mutants (Fig. 5B). We next verified whether ectopic expression of AGP2 could restore the expression of target genes in the agp2D mutant. Introduction of the single copy plasmid pAGP2 restored nearly normal expression of the target genes DUR3, SAM3 and SKY1, while having no effect on ACT1 expression (Fig. 5C). Collectively, the above analyses

Sky1 is not involved in the transcriptional control of the polyamine permeases by Agp2
From all the gene expression analyses data, SKY1 was the most strongly downregulated gene (43-fold decrease from the microarray analysis), which led us to assess whether this SR kinase, shown to be a key factor responsible for both high-and low-affinity polyamine transport, could mediate Agp2 regulatory function [16]. As illustrated in Figure 6A, SKY1 deletion did not alter the expression of the Agp2 target genes, DUR3 and SAM3, thus eliminating Sky1 as an intermediate in the signaling mechanism leading to the positive action of Agp2 on the transcriptional expression of the two major polyamine permeases. It is noteworthy that the expression of SKY1, as in the case of AGP2, was not altered in the dur3D sam3D double mutant (Fig. 6B), excluding the possibility that these two permeases can influence this Agp2 target.
Ste12, Sok2, Sfp1, Yap1 and Msn2 are predicted to transduce a major portion of the signal from Agp2 In order to obtain some insight on the potential pathway(s) responsible for the regulation of gene expression by Agp2 as a sensor, we performed an analysis of the frequency of (directly or indirectly) documented transcription factor recruitment for the expression of genes that were either downregulated by $2.5-fold (n = 140) or upregulated by $2.5-fold (n = 152), using the YEASTRACT algorithm (http://www.yeastract.com/ formgroupbytf.php) [36]. More than 50% of the subset of genes found to be strongly downregulated in agp2D mutants were under the control of either Sfp1 or Ste12 (Table S5) at a frequency that was 57% and 51% higher than for the full list of ORFs tested, respectively. Ste12 was also the most highly represented transcription factor (48%) among the most strongly upregulated genes in agp2D cells (Table S6). In addition, a few other transcription factors (Yap1.Sok2.Msn2.Kar4.Dal80.Sut1. Msn4) and (Yap1.Sfp1.Aft1.Sok2.Msn2.Hap1.Rme1.I-me1.Ume6) were markedly overrepresented among the list of genes downregulated and upregulated, respectively, in the agp2D mutants. Importantly, the highest enrichment factor for potential transcription factor utilization for the regulation of genes that were either up-or downregulated in agp2D cells was found for Sok2 (2.44-and 2.46-fold, respectively) and Yap1 (2.54-and 2.12-fold, respectively). However, single mutation in any of the above transcription factor genes did not reveal the polyamine phenotypes displayed by the agp2D mutants (data not shown), raising the possibility that some of these factors could perform redundant functions in polyamine homeostasis. It is noteworthy that a subgroup of these dysregulated transcription factors, e.g., Kar4, which specifically regulates meiosis and the karyogamy stages of the pheromone response pathway, controls target genes [37] that are among the most strongly affected by AGP2 inactivation. Thus, these observations support a model that transcription factors are indeed involved in the signalling pathway by which Agp2 influences polyamine uptake.

Discussion
We had previously documented that Agp2 affects high-affinity polyamine transport, although it was unclear whether this protein acts directly as a polyamine permease or as a regulator of polyamine transport, or both [6]. According to Uemura et al., (2007) S. cerevisiae has only two high-affinity polyamine transporters, namely Dur3 and Sam3, while Agp2 plays virtually no role in polyamine uptake (see Fig. 6 in [13]), raising the question of whether bona fide Agp2 was actually examined in their study, as the text refers to Agp1 (e.g., see p. 7 of Ref. 13, where the mutants are designated as agp1D). In a separate study, Porat et al., (2005) claimed to re-isolate Agp2 from a high throughput screen designed to identify factors involved in protecting cells against polyamine toxicity and, in a similar manner as our previous report, demonstrated that Agp2 is indeed required for high affinity polyamine transport [6,38]. As such, it is difficult to account for the major discrepancies between the results obtained by Uemura et al. and the data presented herein on Agp2 function. In this current study, we confirm and extend our previous demonstration that Agp2 plays an essential role in polyamine transport in yeast, and show that it acts as a novel regulator by sustaining the expression of several plasma membrane transporter genes, including DUR3, SAM3 and HNM1. Hnm1 is shown for the first time to be involved in the uptake of L-carnitine into cells. In addition, Agp2 controls, either as an activator or a repressor, the expression of a relatively small subgroup of genes that are involved in well-defined biological processes such as the pheromone response pathway and vitamin and cofactor biosynthesis. Moreover, we clearly demonstrate the functional redundancy between Dur3 and Sam3 as polyamine permeases, since deletion of both genes is necessary to generate an agp2D-like phenotype, which is characterized by a strong defect in polyamine uptake as well as a marked resistance to  toxic polyamines. We conclude that Agp2 is unlikely to be a highaffinity transporter with a dual role in polyamine and L-carnitine transport since (i) high concentrations of L-carnitine did not block polyamine uptake, as confirmed by the lack of protection afforded by L-carnitine against polyamine toxicity and (ii) the dur3D sam3D double mutant did not show a significant level of high-affinity polyamine uptake activity as would have been expected if Agp2 acts as a bona fide high-affinity transporter. The fact that Agp2 expression level and plasma membrane localization is normal in the dur3D sam3D double mutant and that this double mutant is much less tolerant to high spermine concentrations as compared to the agp2D mutant, raises the possibility that the additional resistance caused by agp2D mutant might be explained by the lack of a functional low-affinity polyamine permease. Consistent with the existence of a lowaffinity transporter, Uemura et al. showed that the dur3D sam3D double mutant displayed an intermediate level of polyamine transport activity in the presence of a high (100 mM spermidine) substrate concentration. It is unlikely that Agp2 might directly act as a low-affinity permease, as overexpression of Agp2-GFP did not sensitize sky1D or the dur3D sam3D mutant to polyamines (data not shown). A more likely possibility is that Agp2 regulates the lowaffinity polyamine uptake activity of an as yet unidentified permease. This is supported by several facts (i) Agp2 exerts stringent positive control on SKY1 gene expression, encoding a SR kinase essential for both high-and low-affinity polyamine transport activities, (ii) deletion of either AGP2 or SKY1 leads to strong resistance against the toxicity of polyamine and polyamine analogues, (iii) sky1D mutants are strongly defective in polyamine  Agp2 Regulates Polyamine Transporters transport, similar to that found in the agp2D mutant and the dur3D sam3D double mutant, and (iv) AGP2 gene expression, as confirmed by Agp2-GFP, is normal in sky1D mutants precluding the possibility that Agp2 plays a direct role in low-affinity polyamine transport [5,16,17,38,39]. To date, the exact role of Sky1 in regulating both high-and low-affinity polyamine transports is not known. In the case of the high-affinity transport, Sky1 does not act downstream of Agp2 to positively regulate DUR3 and SAM3 gene transcription, but it could serve to activate Dur3 and Sam3 transport activities by phosphorylation [39]. In the case of the low-affinity transport, Sky1 is believed to activate the type 1 phosphatases Ppz1 and Ppz2, which inhibit the activity of the Trk1,2 K + transporters, thus hyperpolarizing cells and increasing polyamine transport [16,40,41]. In the absence of Sky1, the increased activity of the Trk1,2 K + permeases is believed to depolarize cells by opposing the proton motive potential built up by the Pma1 H + -pump leading to diminished uptake and consequently resistance to polyamine. Therefore, the positive action of Agp2 on Sky1 may ultimately increase low affinity polyamine uptake activity via the general increase in membrane potential that results from the inhibition of Trk1,2 K + permeases as illustrated by the model shown in Figure 7. The latter explanation might also account for the fact that agp2D mutants are more strongly resistant to polyamine than dur3D sam3D double mutants since the former positively regulates both high-affinity, via Dur3 and Sam3, and low-affinity polyamine uptake via Sky1. Consistent with the model (Fig. 7), ppz1D ppz2D double mutants have been reported to be highly resistant to spermine in the presence of functional Trk1 and Trk2 [16,40]. However, these studies did not directly examine if the ppz1D ppz2D double mutant is defective in polyamine uptake and whether the defect can be rescued by deletion of the TRK1 and TRK2 genes.
How Agp2 transmits a signal to modulate the expression of Dur3, Sam3, and Sky1 in order to maintain polyamine homeostasis is not known, although it could occur through one or more transcriptional activators in a manner similar to the Ssy1 transceptor [10]. In such a scenario, Agp2 would behave as a plasma membrane sensor or non transporting transceptor for polyamines. While our haploid genome-wide screen did not yield any such transcription factor [5], the microarray analysis performed herein has identified a number of transcription factor genes that could be involved in either the downregulation or upregulation of Agp2 target genes. According to the YEAS-TRACT algorithm [36], which predicts the frequency of putative transcription factors for the most strongly regulated genes identified in agp2D mutants, there is a strong bias in favour of several transcription factors including Sfp1, Ste12, Yap1, Sok2, Msn2, Aft1 and Gcn4 that could receive signals from Agp2 (Tables  S5 and S6). However, the actual factor that transmits the signal from Agp2 remains elusive, perhaps due to functional redundancy.
The sum of the above observations on the pattern of gene regulation by Agp2 consistently supports the following tentative Figure 7. A model illustrating Agp2 function in regulating polyamine uptake. Agp2 could act to sense several nutrients such as polyamines and maintain the transcriptional expression of many genes including the regulatory kinase Sky1 and the two high affinity polyamine transporters Dur3 and Sam3. Similar to Agp2, Sky1 controls both high and low affinity polyamine transport. Since Sky1 does not affect DUR3 or SAM3 expression, we propose that it may activate these transporters by post-translational modification, e.g., by phosphorylation. Furthermore, Sky1 has been shown to negatively regulate the Trk1 and 2 potassium transporters, perhaps via activation of the Ppz1 and 2 phosphatases, leading to polyamine uptake. ppz1D ppz2D mutants activate Trk1, 2 potassium uptake and contribute to polyamine resistance. The transcription factors (TFs) responsible for conveying Agp2 sensory function and culminating in gene activation remain to be identified. doi:10.1371/journal.pone.0065717.g007 model that explains the physiological mechanism of action of Agp2 as a cell surface regulatory protein (see Fig. 7). An important unresolved issue of this model is the exact nature of the ligand(s) that stimulates Agp2 function in order to regulate the activation of the signalling pathways. The requirement for a functional AGP2 gene for polyamine transport via its stringent positive regulation of DUR3, SAM3 and SKY1, as well as for L-carnitine uptake at least through expression of HNM1, suggest a plausible connection of Agp2 with a metabolic shift towards gluconeogenesis and other carbohydrate storage. This metabolic shift would require an increased level of both coenzyme A and L-carnitine in order to transfer acetyl groups across the mitochondrial membrane [42]. Coenzyme A must be synthesized de novo from spermine, as spermine is the sole source of b-alanine that provides the initial precursor pantothenic acid for coenzyme A synthesis [43,44,45,46]. For L-carnitine, the intracellular pool derived from the degradation of e-trimethyllysine-containing proteins (e.g., cytochrome c) that provides the initial substrate for L-carnitine biosynthesis might be insufficient, and thus must be attained through import from extracellular sources [47]. In this respect, the primary function of Agp2 might be to act as a sensor to expand the intracellular pool of both coenzyme A and L-carnitine. Whether polyamines and L-carnitine are indeed the actual ligand molecules sensed by Agp2, and the nature of the mechanism responsible for the activation of cellular signaling that ultimately leads to the down-or up-regulation of specific gene expression are still unanswered questions worth future investigation. Figure S1 The effect of L-carnitine on polyamine toxicity of parent cells. Exponentially growing wild type cultures were coincubated without or with 10 mM L-carnitine and various concentrations of spermine for 1 h, followed by serial dilution and then plated onto solid yeast-peptone-dextrose (YPD) agar to score for survivors. The results are the averages of three independent experiments. (TIF)