The General Amino Acid Permease FfGap1 of Fusarium fujikuroi Is Sorted to the Vacuole in a Nitrogen-Dependent, but Npr1 Kinase-Independent Manner

The rice pathogenic fungus Fusarium fujikuroi is well known for the production of a broad spectrum of secondary metabolites (SMs) such as gibberellic acids (GAs), mycotoxins and pigments. The biosynthesis of most of these SMs strictly depends on nitrogen availability and of the activity of permeases of nitrogen sources, e.g. the ammonium and amino acid permeases. One of the three ammonium permeases, MepB, was recently shown to act not only as a transporter but also as a nitrogen sensor affecting the production of nitrogen-repressed SMs. Here we describe the identification of a general amino acid permease, FfGap1, among the 99 putative amino acid permeases (AAPs) in the genome of F. fujikuroi. FfGap1 is able to fully restore growth of the yeast gap1∆ mutant on several amino acids including citrulline and tryptophane. In S. cerevisiae, Gap1 activity is regulated by shuttling between the plasma membrane (nitrogen limiting conditions) and the vacuole (nitrogen sufficiency), which we also show for FfGap1. In yeast, the Npr1 serine/threonine kinase stabilizes the Gap1 position at the plasma membrane. Here, we identified and characterized three NPR1-homologous genes, encoding the putative protein kinases FfNpr1-1, FfNpr1-2 and FfNpr1-3 with significant similarity to yeast Npr1. Complementation of the yeast npr1Δ mutant with each of the three F. fujikuroi NPR1 homologues, resulted in partial restoration of ammonium, arginine and proline uptake by FfNPR1-1 while none of the three kinases affect growth on different nitrogen sources and nitrogen-dependent sorting of FfGap1 in F. fujikuroi. However, exchange of the putative ubiquitin-target lysine 9 (K9A) and 15 (K15A) residues of FfGap1 resulted in extended localization to the plasma membrane and increased protein stability independently of nitrogen availability. These data suggest a similar regulation of FfGap1 by nitrogen-dependent ubiquitination, but differences regarding the role of Fusarium Npr1 homologues compared to yeast.


Introduction
In the phytopathogenic ascomycete Fusarium fujikuroi, the biosynthesis of a broad spectrum of secondary metabolites (SMs), such as the plant hormones gibberellins (GA) and the red polyketide pigments bikaverin and fusarubins [1,2,3], is strictly repressed, while the biosynthesis of mycotoxins such as fusarin C, fusaric acid, and apicidin F is induced by high nitrogen concentrations [4,5,6,7]. Therefore, mechanisms of nitrogen sensing and nitrogen-mediated regulation of gene expression are in the focus of our research.
Uptake and eventually sensing of nitrogen by specific permeases play key roles in nitrogen metabolism [8,9]. Processes of nitrogen sensing and subsequent transduction of the signal to a nitrogen regulation network have been intensively studied in yeast. Beside the three ammonium permeases Mep1 to Mep3, yeast have about 20 amino acid permeases (AAPs) [10]. The expression of nitrogen permease-encoding genes such as GAP1 and the MEP genes is negatively influenced by the availability of favored nitrogen sources like glutamine, while poor nitrogen sources such as urea or proline support their transcription in a Gln3 and Gat1-dependent manner [11][12][13][14]. Some of the AAPs have been proposed to play a role as nitrogen sensors, e.g. the non-transporting amino acid carrier homologue Ssy1 [15] and the general amino acid permease Gap1 [16]. Yeast Mep2, and its orthologue from Candida albicans for instance, are proposed to play a signaling role as they are required for the filamentation process occurring upon nitrogen limitation [17][18][19][20][21].
We have characterized all three ammonium permease-encoding genes in F. fujikuroi, MEPA, MEPB, and MEPC, regarding their participation in ammonium uptake and their potential involvement in signal transduction. The expression of all three genes is strictly regulated by the general nitrogen regulator AreA, the homologue of Saccharomyces cerevisiae GATA transcription factors Gln3 and Gat1 [13,14,22]. Severe growth defects of ΔMEPB mutants on low ammonium medium and deregulation of several nitrogen-repressed genes, such as the GA and bikaverin biosynthetic genes, suggested that MepB could function as an ammonium sensor in addition to its important role in ammonium transport in F. fujikuroi. However, the mechanism of ammonium sensing by MepB is not yet understood.
Plasma membrane permease activity can be modulated by transcriptional regulation, control of cell surface protein level and fine-tuning of the inherent transport activity via posttranslational modification. In yeast, the serine/threonine protein kinase Npr1 (nitrogen permease reactivator) is a regulator of the activity of permeases of nitrogenous compounds, being required for optimal uptake of nitrogen sources in cells grown on a non-preferred nitrogen source [21,23,24]. Npr1 is thus required for the activity of several permeases including the general amino acid permease Gap1 [24,25]. In the latter case, Npr1 is required to maintain Gap1 at the cytoplasmic membrane by preventing ubiquitination-dependent endocytosis under non-preferred nitrogen supply. In the npr1 mutant, newly synthesized Gap1 is further directly sorted to the vacuole and never reaches the plasma membrane whatever the quality of the nitrogen supply [26][27][28]. As a consequence, the uptake of citrulline, ornithine and tryptophan, all depending on the activity of Gap1, is affected in the npr1 mutant. Beside the stabilization/ destabilization of nitrogen permeases, there seem to be other ways for Npr1 to affect the activity of specific permeases. Npr1 is required for the inherent activity of ammonium transport by Mep1, Mep2 and Mep3, the npr1 mutant exhibiting a growth defect on low-ammonium medium similar to that of a mutant lacking the three ammonium permeases [18,25,29]. However, contrary to Gap1, plasma membrane localization of the ammonium permease Mep2 is not altered in the ΔNPR1 mutant of S. cerevisiae and Candida albicans [20,29,30] Under poor nitrogen supply, the Npr1 kinase mediates phosphorylation of the serine residue S457 in the Cterminal auto-inhibitory domain of Mep2 [29]. This phosphorylation relieves auto-inhibition and allows Mep2 to adopt an active conformation that enables high ammonium transport capacity. Due to the impairment of uptake of repressing nitrogen sources, a number of genes sensitive to nitrogen catabolite repression (NCR) are indirectly derepressed in Npr1-lacking cells growing in the presence of sufficient ammonium or glutamine [25,31]. The level of Npr1 phosphorylation, and presumably of its activity, is controlled by TORC1 (target of rapamycin complex1) according to the quality of the nitrogen supply [30,[32][33][34]. Treatment of yeast cells with the TORC1-inhibiting drug rapamycin as well as nitrogen starvation promote rapid though not complete dephosphorylation of Npr1, while the protein is hyperphosphorylated in a nitrogen-rich medium [34].
Not much is known about mechanisms for regulating subcellular localization of nitrogen permeases in filamentous fungi. So far no Npr1-like regulators have been characterized, and therefore, their putative involvement in regulating the activity of any nitrogen permease has not been studied in any filamentous fungus.
In this work, we identified a homolog of the yeast Gap1, FfGap1, among the 99 annotated putative amino acid permeases in the genome of F. fujikuroi by complementing the S. cerevisiae gap1Δ mutant. We show that FfGAP1 is not expressed in media containing glutamine, and that its transcription strictly depends on the presence of the GATA transcription factor AreA. The FfGap1 protein is localized to the plasma membrane when internal amino acid levels are low. Upon addition of glutamine, the FfGap1-Gfp fusion protein is endocytosed and delivered to the vacuole. To test whether Npr1-like protein kinases are required for stabilization of FfGap1 at the plasma membrane as in yeast, we cloned three homologous genes of ScNpr1, FfNPR1-1, FfNPR1-2 and FfNPR 1-3, and studied their role in regulating FfGap1 activity. We show that complementation of the yeast npr1Δ mutant with the F. fujikuroi NPR1-1, NPR1-2 and NPR1-3 genes, respectively, resulted in partial restoration of growth on ammonium, arginine and proline by FfNPR1-1. However, the putative serine/threonine kinases had no impact on sorting of FfGap1. Instead, we demonstrate that two conserved lysine residues at the N-terminus of FfGap1 are involved in regulating FfGap1 protein abundance and its sorting to the vacuoles. Point mutations of these lysine residues, K7 and K15, of FfGap1 resulted in prolonged maintenance at the cytoplasmic membrane under high glutamine conditions, probably by preventing ubiquitination-dependent endocytosis.
To infer the evolutionary relationship between the 19 candidates and the four references we calculated a phylogenetic tree and selected closely related proteins in extracting the members of the smallest subtree that contains all four known reference GAPs (Fig 1A). This resulted in a refined candidate set of eleven putative F. fujikuroi GAP proteins (FFUJ_01137, FFUJ_05331,  FFUJ_ 08309, FFUJ_08705, FFUJ_08913, FFUJ_09118, FFUJ_09356, FFUJ_10091,  FFUJ_11370, FFUJ_11387, FFUJ_11624).
We next determined whether the expression of these putative Gap1-like encoding genes is repressed under nitrogen sufficient conditions as it has been shown for Gap1 in S. cerevisiae [13,40]. The wild-type was grown for three days under known conditions for nitrogen  [36], C. albicans [37], N. crassa [38], P. chrysogenum [39], and 19 F. fujikuroi proteins with the highest sequence homology to S. cerevisiae Gap1. Branches show bootstrap values (%), scale bar indicates amino acid substitutions per site. The fungal GAP proteins branched together with eleven putative GAP sequences of F. fujikuroi (orange box) (B) Expression analysis of putative GAP-encoding genes in F. fujikuroi. Total RNA was isolated from the wild-type and the ΔAREA and ΔAREB deletion mutants grown for 3 days in ICI liquid cultures with 6 mM (-Gln) or 60 mM (+Gln) glutamine as single nitrogen source and used for Northern blot analysis. 18S rRNA was visualized as loading control. deficiency and nitrogen sufficiency with 6 mM or 60 mM glutamine, respectively [41]. To further determine whether the two major nitrogen regulators, the GATA transcription factors AreA and AreB, are involved in regulation of Gap1-like AAP gene expression, the F. fujikuroi ΔAREA [42] and ΔAREB [41] mutants were grown under the same conditions. The expression of the eleven putative Gap1-encoding genes was analyzed by Northern blot analysis (Fig 1B). The expression of five of the AAP-encoding genes (FFUJ_08309, FFUJ_08705, FFUJ_08913, FFUJ_09356 and FFUJ_11370) was too low to be detected under any of the tested conditions (data not shown). Four of the genes, FFUJ_09118, FFUJ_05331, FFUJ_01137 and FFUJ_11624, were expressed only under nitrogen starvation conditions in the wild type and in the ΔAREB mutant, but not expressed in the ΔAREA mutant indicating that they are subject to nitrogen metabolite repression and are expressed in an AreA-dependent manner (Fig 1B). The expression of FFUJ_01137 and FFUJ_11624 is more elevated in the ΔAREB mutant than in the wild type under nitrogen limiting conditions indicating that AreB could act as a transcriptional repressor of these AAPs. The gene FFUJ_10091 shows low expression under all conditions while the expression of FFUJ_11387 depends on the presence of both AreA and AreB under nitrogen sufficient conditions. In S. cerevisiae, the GAP1 gene is sensitive to NCR, and its expression is activated by either the GATA transcription factor Gln3 or the GATA transcription factor Gat1 when grown under poor (proline) nitrogen conditions [43]. To determine whether the two nitrogen-repressed and AreA-dependent GAP1-like genes (FFUJ_09118, FFUJ_05331) share other analogies with the S. cerevisiae GAP1, we proceeded with heterologous functional expression in yeast. Full length cDNA fragments of these genes were cloned into the yeast expression vector yEX-C [39]. The vectors were transformed into S. cerevisiae gap1Δdip5Δ mutant strain M4276 [44] and the gap1Δssy1Δ mutant strain M4238 [10]. Dip5 is the only yeast transporter capable of glutamate transport beside Gap1 [44], while Ssy1 is not an active amino acid transporter, but acts as amino acid sensor that mediates the transcriptional activation of several amino acid permeases [15]. Additionally, the vector yEX-PcGAP1, expressing the PcGAP1 gene from P. chrysogenum [39], and yEX-11370 containing the non-expressed F. fujikuroi AAP gene FFUJ_11370 were transformed as controls.
Yeast transformants were grown on minimal media each containing single amino acids as nitrogen source (Fig 2A, shown for FFUJ_05331, FFUJ_09118, FFUJ_01137, FFUJ_11370).
First of all, we tested the ability of the potential AAPs to restore growth on minimal L-citrulline medium as transport of this amino acid is a peculiarity of the yeast Gap1 permease [45,46]. Of note, the gap1Δdip5Δ double mutant could only grow on citrulline when expressing the P. chrysogenum PcGAP1 or the F. fujikuroi AAP-encoding gene FFUJ_09118, suggesting that the latter F. fujikuroi permease acts as a Gap1 orthologue. Aspartic acid could obviously be utilized by other permeases than Dip5 and Gap1, as the double mutant displayed no growth defect with this amino acid. Furthermore, the gap1Δssy1Δ double mutant strain grew better on isoleucine and phenylalanine when expressing PcGAP1 and the gene FFUJ_09118, indicating that both gene products are broad spectrum permeases that can complement many growth defects of this mutant. In addition, the capability to import tyrosine was completely lost in the gap1Δs-sy1Δ background, most likely due to the reported Ssy1-mediated activation of the main tyrosine transporter Bap2 [15]. Interestingly, only FFUJ_09118, but not PcGap1, was able to restore tyrosine uptake. To further address the specificity of this functional permease towards amino acids, strains were dropped in ten-fold dilution steps onto minimal media containing 1 mM ammonium sulfate as positive control, or 1 mM glutamate as sole N-source ( Fig 2B). All of the tested yeast mutants (shown for FFUJ_09118, FFUJ_05331 and FFUJ_11370) grew well with ammonium as nitrogen source, while only the yeast transformants carrying PcGAP1 and FFUJ_09118 were able to efficiently utilize glutamate, suggesting that FFUJ_09118 encodes a functional F. fujikuroi permease further able to transport glutamate.
Taken together, these growth assays support that the gene FFUJ_09118 encodes a likely Gap1 orthologue of F. fujikuroi. The gene will be referred to as FfGAP1 in the following.

Functional characterization of FfGap1 in F. fujikuroi
To test the substrate specificity and function of FfGap1, we created a deletion mutant in the F. fujikuroi wild-type strain. The deletion did not result in any growth defects on minimal medium, neither on glutamine, glutamate, proline, citrulline or tryptophane (data not shown). These data indicate that transport of these amino acids can be ensured by other permeases than FfGAP1 and that the latter is not the sole pathway of citrulline uptake unlike the situation in yeast. The tested amino acids are obviously transported by one or more of the numerous additional AAPs present in the genome of F. fujikuroi [13].
Dual functions as nitrogen transporters and sensors have been described for several fungal nitrogen permeases, including the ammonium permeases Mep2 of S. cerevisiae [17,20] and MepB of F. fujikuroi [22], as well as the yeast Gap1 [16,47]. To investigate a possible dual function of FfGap1, we examined whether this permease plays a sensing role similarly to what was shown for MepB [22]. In this case the nitrogen-repressed SM genes would be highly expressed in the ΔFfGAP1 mutant under nitrogen-sufficient conditions. We cultivated the F. fujikuroi wild-type and ΔFfGAP1 mutant with low (6 mM) and high (60 mM) proline and glutamine levels for 15 h and 72 h. Expression of two SM biosynthesis genes, the ent-copalyldiphosphate/ ent-kaurene-synthase gene (CPS/KS) of the GA and the monooxygenase gene (BIK2) of the bikaverin biosynthesis pathways, were analyzed by Northern blot analysis (Fig 3). The GA and bikaverin biosynthesis genes are subject to nitrogen metabolite repression in an AreA-dependent and AreA-independent manner, respectively [1,48]. In general, the observed expression patterns were similar in both the wild-type and ΔFfGAP1 strains. At 15 h, when the fungus has not yet consumed the majority of the nitrogen source, the genes are not expressed. After 72 h the CPS/KS and BIK2 genes are highly expressed under low, but not under high glutamine conditions as shown before [41,49,50]. However, with proline as nitrogen source BIK2 is even stronger expressed at high compared to low proline concentrations while the GA gene is strongly repressed. These data underline the different nitrogen regulation mechanism for GA (AreA-dependent) and bikaverin (AreA-independent) genes.
Taken together, the nitrogen repression of two SM gene clusters is not released in the ΔFfGAP1 mutant indicating that FfGap1 is not involved in regulating expression of these genes in contrast to the ammonium permease MepB [22].

Nitrogen-dependent intracellular localization of FfGap1
The S. cerevisiae Gap1 is actively transported to the plasma membrane, when no rich nitrogen source is available [27]. We next assessed whether FfGap1 could be similarly regulated by the quality of the nitrogen supply. To follow up the subcellular localization of FfGap1 under different nitrogen conditions, a FfGAP1-GFP fusion construct was generated and transformed into the ΔFfGAP1 deletion strain. The correct integration of the full length FfGAP1-GFP construct at the FfGAP1 locus was shown for several transformants by diagnostic PCR (data not shown). The wild-type and one of the FfGAP1-GFP strains were grown for 48 hours in liquid ICI medium under nitrogen-limiting (6 mM glutamine) conditions. At this time point, glutamine is exhausted and the fungus suffers already from nitrogen starvation [51]. Fluorescence microscopy revealed that the FfGap1-Gfp fusion protein was localized to the plasma membrane, but also in small cellular structures in the cytoplasm. We next assessed the impact of nitrogen resupplementation on FfGap1 localization. After addition of 12 mM glutamine, the subcellular localization of FfGap1 was studied for a period of 5 hours (Fig 4A). At 30 min after addition of glutamine, membrane localization is notably weaker compared to the initial nitrogen starvation conditions. Furthermore, the size of fluorescent intracellular structures, was continuously increasing and small moving structures, probably endosomes, were observed. Five hours upon glutamine addition, the FfGap1-Gfp signal was completely absent at the plasma membrane and instead was concentrated in large, circular structures, probably the vacuoles. Exactly the same effects on FfGap1 localization were observed when proline was used as a nitrogen source instead of glutamine (data not shown). These data indicate that intracellular sorting of FfGap1 strictly depends on nitrogen availability as it is the case in S. cervisiae. However, in contrast to yeast where newly synthesized Gap1 is also delivered to the plasma membrane when cells are grown on poor substrates such as proline or urea as the sole nitrogen source [26], proline is recognized as a good nitrogen source in F. fujikuroi leading to FfGap1 sorting from the plasma membrane to the vacuoles.
In summary, FfGap1 appeared to be localized mainly at the plasma membrane under nitrogen starvation conditions, while the protein was internalized and sorted to the lumen of the vacuole under nitrogen-sufficient conditions as it has been shown for the S. cervisiae Gap1.

Identification of the Npr1 homologue of F. fujikuroi
After demonstrating that FfGap1 sorting is regulated by nitrogen availability, we next asked whether a Npr1-like protein kinase could be involved in such a process. So far, no Npr1homologs are described in any filamentous fungus. A BLASTX search with the S. cerevisiae Npr1 protein sequence against the F. fujikuroi genome database [7] revealed three best hits (FFUJ_04942, FFUJ_02924 and FFUJ_05668), all possessing serine/threonine protein kinase InterPro motifs. FFUJ_04942 showed the highest amino acid identity to ScNpr1 (33.1%), while FFUJ_02924 and FFUJ_5669 displayed only 22.7% and 21.3% identity, respectively. While FFUJ_04942 (774 amino acids) and FFUJ_05668 (736 amino acids) are similar in size to the S. cerevisiae Npr1 protein (790 amino acids), FFUJ_02924 is much smaller than the yeast Npr1 (500 amino acids). The N-terminal region of the protein sequences is rich in serine residues, albeit not as rich as in the case of yeast Npr1 (S1 Fig) [25]. The three Npr1 homologues will be referred to as FfNpr1-1 (FFUJ_04942), FfNpr1-2 (FFUJ_02924) and FfNpr1-3 (FFUJ_05668) in the following. To determine if one of the three ORFs encodes a functional F. fujikuroi Npr1-analogue, full length cDNA clones of all three genes were cloned into the YEplac195-npr1-Term expression vector containing the S. cerevisiae NPR1 promoter and terminator regions. The three resulting vectors and the S. cerevisiae native NPR1 gene in the yeast expression vector pMV33 as a positive control, were transformed into the wild-type yeast strain 23344c and the npr1Δ yeast mutant [54]. Transformants were screened for growth on different nitrogen sources. It has been previously shown that the yeast npr1Δ mutant is not able to grow on several nitrogen sources because Npr1 plays essential but differing roles in regulating the activity of nitrogen permeases. While the inability of the npr1Δ mutant to grow on a specific substrate of Gap1 (e.g. citrulline, and to lesser extend tryptophan) is due to the role of the Npr1 kinase to oppose the ubiquitination and subsequent vacuolar degradation of Gap1 [23,26,27], the inability to grow on low ammonium and the reduced growth on some other nitrogen sources (e.g. proline) is additionally due to inefficient retrieval of catabolic ammonium escaping from the cells, a consequence of the Mep inactivity [18].
To compare the growth behavior of the yeast npr1Δ mutant with that of the transformants, all strains were plated on minimal medium containing different nitrogen sources (Fig 5). The wild-type was able to efficiently use all supplied nitrogen sources, while the yeast npr1Δ mutant did not or only weakly grow on all nitrogen sources except glutamate. Complementation of S. cerevisiae with the native NPR1 gene fully restored growth of the npr1Δ yeast mutant on all nitrogen sources. Although none of the three F. fujikuroi NPR1-like genes was able to restore growth on all nitrogen sources, they were intriguingly able to mediate growth on specific nitrogen sources, respectively. FfNpr1-1 mediated growth on ammonium, urea and partially on arginine. FfNpr1-3 seemed to slightly improve growth on valine and, like FfNpr1-1, on urea while growth on ammonium was only partially restored after 7 days of incubation. FfNpr1-2 was only able to slightly support growth on valine.
The absence of growth restoration on citrulline indicates that none of the potential Npr1-like proteins from Fusarium can re-establish yeast Gap1 activity. To further examine if one of the three FfNpr1-like kinases has an impact on Gap1 stability in yeast, a Western analysis was performed with membrane enriched protein extracts of the npr1Δ strain and npr1Δ strains transformed with F. fujikuroi NPR1 homologues. All strains were grown with urea which is a non-preferred N-source like proline (Fig 6). Hybridization with a Gap1 antibody revealed a stronger abundance of the permease in the npr1Δ cells expressing yeast ScNpr1 compared to npr1Δ cells transformed with an empty vector. None of the Npr1-like proteins enabled stabilization of Gap1. The npr1Δ cells transformed with FfNPR1-1 showed a reduced Gap1 signal compared to the npr1Δ cells transformed with an empty vector suggesting that the heterologous FfNpr1 kinase could somehow favor the degradation of Gap1. This could explain the apparent loss of proline utilization observed upon FfNPR1-1 expression. FfNpr1-2 and FfNpr1-3 expression were accompanied with a similar effect on the Gap1 detected levels, although to a lesser extent.
Yeast Npr1 was recently shown to control the Mep2 ammonium transport activity by mediating phosphorylation of the permease [30]. Mep2 thereby appears as a double band on SDSpage, the slow-migrating form corresponding to the phosphorylated and active Mep2 permease. Our western blot analysis shows that in contrast to yeast Npr1, none of the three potential Npr1-like proteins from Fusarium restore the phosphorylation of Mep2. We thereby conclude that growth on ammonium of npr1Δ cells expressing FfNpr1-1 probably relies on activation of Mep1 and/or Mep3 but not on Mep2.
Taken together, at least two of the 3 Npr1-like F. fujikuroi kinases (FfNpr1-1 and FfNpr1-3) appear to have an impact on the functionality of different permeases of nitrogen sources in S. cerevisiae. The targets of the different FfNpr kinases seem to overlap in some cases. Our data suggest that the functions of yeast Npr1 could be split among different kinase proteins in F. fujikuroi, mainly FfNpr1-1 and FfNpr1-3. However, none of the tested Fusarium Npr1-like proteins is able to stabilize and restore yeast Gap1 function.

FfNpr kinases only slightly influence nitrogen utilization in F. fujikuroi
To elucidate a potential role of the three FfNpr1 kinases in F. fujikuroi, FfNPR1-1, FfNPR1-2 and FfNPR1-3 deletion mutants were generated (for details, see Materials and Methods section). Homologous in locus integration of the gene replacement fragments was confirmed by Southern Blot or diagnostic PCR (S2 Fig). Additionally, we generated ΔΔFfNPR1-1/FfNPR1-2 and ΔΔFfNPR1-1/FfNPR1-3 double deletion mutants. The growth of the wild-type, ΔFfGAP1, and the single and double NPR1 deletion strains was examined after 4 days incubation on minimal agar media with different nitrogen sources at two concentrations (1 and 10 mM). Each growth assay was done with two independent transformants for each deletion experiment. As they showed an identical growth behavior, plate assays are shown for one transformant each  Fig 7A). For comparison, ΔAREA and ΔAREB mutants which have lost one of the two global nitrogen regulators, the GATA transcription factors AreA or AreB [13], were included into these growth assays.
While the ΔAREA and ΔAREB mutants revealed the well known growth defects on some nitrogen sources [41], the FfNPR1 mutants did not show significant growth defects compared to the wild-type on all tested media. Surprisingly, the FfNPR1-3 mutant showed a significantly increased colony diameter compared to the wild-type on all nitrogen sources, including ammonium. The colonies appeared less dense and sparse aerial hyphae. In addition, the ΔFfNPR1-3 mutant produced a red pigment under low nitrogen conditions (Fig 7A). However, this distinctive phenotype was not shared with the FfNPR1-1/FfNPR1-3 double mutant, which showed a morphology similar to ΔFfNPR1-1 and the wild-type. Complementation of the ΔFfNPR1-3 mutant (ΔFfNPR1-3 C ) restored the wild-type growth on all tested amino acids, proving that the observed phenotype is caused by the NPR1-3 deletion (S3 Fig). Despite the different The General Amino Acid Permease of Fusarium fujikuroi The General Amino Acid Permease of Fusarium fujikuroi morphology of the ΔFfNPR1-3 mutant, the actual biomass produced on solid and liquid minimal media (1 mM glutamine) was the same for the wild-type and the ΔFfNPR1-3 and ΔΔFfNPR1-1/FfNPR1-3 mutants (data not shown).
As none of the single and double mutants revealed a significant growth defect on one of the tested nitrogen sources, we generated ΔΔΔFfNPR1-1/FfNPR1-2/FfNPR1-3 triple deletion mutants and performed plate assays. Although these triple mutants showed a slightly reduced growth rate on all tested media compared to the wild-type, they still were able to utilize all nitrogenous compounds for growth (Fig 7B) suggesting that it is not attributed to a specific defect in amino acid uptake.
Summarizing, the NPR1 deletion strains showed no significant growth defects on various amino acids, in contrast to the global nitrogen regulator mutants ΔAREA and ΔAREB. These data suggest that the tested FfNpr1 kinases have no major and global impact on the activity of amino acid permeases in F. fujikuroi.
FfNpr kinases of F. fujikuroi do not influence the nitrogen dependent sorting of FfGap1 As shown above, the sorting of FfGap1 depends on nitrogen availability. To assess whether any of FfNpr kinases affect this sorting process, we transformed the ΔNPR1-1 and ΔNPR1-3 mutant strains with the plasmid pFfGAP1-GFP. The generated transformants were selected for in locus integrations of the FfGAP1-GFP construct. The resulting ΔNPR1-1/FfGAP1-GFP and ΔNPR1-3/FfGAP1-GFP transformant strains and the wild type carrying the FfGAP1-GFP construct were grown for 1 day under nitrogen limiting conditions (3 mM glutamine), and the cultures were monitored by fluorescence microscopy before and 120 minutes after the addition of 12 mM glutamine (Fig 8).
As in the wild-type, both FfNPR1 mutants displayed an abundance of FfGap1-Gfp in the plasma membrane during starvation, while the fusion protein was completely sorted to the vacuoles 120 minutes after addition of glutamine, indicating that the single deletion of FfNpr1-1 and FfNpr1-3 does not affect nitrogen-dependent sorting of FfGap1.
Similarly, the addition of nitrate or ammonium resulted in continuous enrichment of the fluorescence signal in the vacuoles in all strains (S4A, S4B and S4C Fig). However, with nitrate it took longer to transport FfGap1 to the vacuoles in both the wild-type and the mutant backgrounds, compared to glutamine and ammonium. This delay is probably due to the fact that nitrate has to be converted first to ammonium and glutamine to be used by the fungus.
Summarizing, these microscopic observations indicate that intracellular sorting of FfGap1 in response to nitrogen does not exclusively depend on the activity of the tested Npr1-like protein kinases in F. fujikuroi. Nevertheless, sorting of FfGap1 depends on the nitrogen availability.

Sorting of FfGap1 depends on conserved lysine residues
In S. cerevisiae it has been shown that endocytosis and degradation of Gap1 upon addition of the preferred nitrogen source ammonium is triggered by ubiquitination of N-terminal lysine residues at positions 9 and 16. Npr1 prevents the ubiquitination of Gap1 at these lysine residues during growth on non-preferred nitrogen sources [26,27,55,56].
To examine whether ubiquitination might be involved in FfGap1 sorting, we compared the sequence of the Gap1 proteins of S. cerevisiae and F. fujikuroi and found two potential ubiquitination targets at position K7 and K15 of FfGap1. Point mutations were introduced into the respective codons to exchange both lysine residues for alanine (K7A; K15A). The mutated and wild-type FfGAP1-GFP fusion constructs were transformed into the ΔFfGAP1 mutant. The resulting strains expressing FfGAP1(K7/15A)-GFP and FfGAP1-GFP (wild-type FfGAP1), respectively, were cultivated under nitrogen starvation conditions for 48 hours. The subcellular localization of FfGap1-Gfp and FfGap1(K7/15A)-Gfp was examined by fluorescence microscopy before and up to 5 hours after the addition of 12 mM glutamine. Gfp signals of both the wild-type and the mutated FfGap1 were detected at the plasma membrane and in intracellular structures under nitrogen starvation conditions (Fig 9A). Two hours after addition of glutamine, the wild-type FfGap1 was totally absent from the membranes and localized in the lumen of vacuoles, as shown before (Fig 4). In contrast, the mutated FfGap1(K7/15A) fusion protein was still localized at the membranes at this time point, and traces of the mutated protein were visible at the membrane even five hours after addition of glutamine. Longer incubation (6 hours and more) resulted in a complete sorting of the mutated Gap1-Gfp proteins to the vacuoles (data not shown). These data indicate that the substitution of the lysine residues at position 7 and 15 of FfGap1 has an impact on the duration of membrane localization after adding nitrogen. However, the general nitrogen-dependent sorting to the vacuoles was significantly delayed but not fully abolished.  To test whether the mutations have an effect on general protein stability and abundance, Western blot analyses were performed using an anti-GFP antibody. Strains carrying FfGAP1-GFP or the mutated FfGAP1(K7/15A)-GFP were cultivated with low nitrogen for 72 hours and harvested before and 2 hours after the addition of 12 mM glutamine (Fig 9B). The amount of the wild-type protein drastically decreased after addition of glutamine, indicating a nitrogen-dependent degradation of the protein. The amount of the mutated protein was significantly higher compared to the wild-type under both conditions what fits well to the observed delayed degradation process of FfGap1(K7/15)-Gfp after addition of glutamine. These data indicate that the lysine residues K7 and K15 are involved in nitrogen-dependent degradation, though additional factors might also intervene in the protein stability and sorting.

Discussion
Recently, it has been shown that in F. fujikuroi, most of the 45 SM gene clusters are regulated by nitrogen availability [7]. Some of them are targets of the GATA transcription factors AreA and/or AreB [5,6,41,49]. In addition to SM biosynthetic genes, a set of nitrogen permeaseencoding genes, such as the three ammonium transporter-encoding genes MEPA, MEPB, and MEPC or the peptide transporter-encoding gene MTD1 are subject to AreA-mediated nitrogen metabolite repression in F. fujikuroi [22,50]. Studies on S. cerevisiae reported that two nitrogen permeases, the ammonium transporter Mep2 [17,18,19] and the general ammonium permease Gap1 [16,57], also function as receptors for rapid activation of the protein kinase A (PKA) pathway upon addition of their substrate [58]. It is not known whether a similar transceptor function of Gap1 homologues and a functional link to the PKA pathway exist in filamentous fungi. In F. fujikuroi, a potential dual function as transporter and receptor has been shown so far only for the ammonium permease MepB [22]. Its deletion resulted in upregulation of otherwise nitrogen-repressed genes, e.g. the GA and bikaverin biosynthesis genes, under ammonium-sufficient conditions [22]. However, a potential dual role of Gap1 homologues as permeases and receptors have never been shown in any filamentous fungus. It is also not known whether the stability of Gap1 homologues is regulated by a Npr1-like protein kinase as it has been shown in yeast [36,46].
In this work, we identified and characterized a Gap1-homologue in F. fujikuroi and studied its regulation on transcriptional and protein levels as well as the impact of potential Npr1-like protein kinases on FfGap1 stability and subcellular localization.

Identification and function of FfGap1
The analysis of the recently sequenced genome of F. fujikuroi [7] revealed 99 potential AAP proteins with the characteristic AAP protein domains. This is a significantly higher number compared to 20, 27, and only 19 identified AAPs in the genomes of S. cerevisiae, Candida albicans and Aspergillus nidulans, respectively [59,60]. Among these genes, only FFUJ_09118 and FFUJ_05331 were highly expressed under nitrogen-limiting conditions in an AreA-dependent manner similar to the Gln3-dependent expression of the yeast GAP1 gene [40]. Yeast complementation experiments were consistent with FFUJ_09118 corresponding to a functional analogue of yeast Gap1. However, deletion of FfGAP1 in F. fujikuroi did not result in a distinct phenotype for a specific amino acid suggesting that the loss of FfGap1 function can be compensated by the presence of numerous additional transporters. In addition, we cannot exclude that among the 99 potential AAPs more Gap1-like proteins exist with a broad spectrum of transported amino acids similarly to Candida albicans [59].
Deletion of FfGAP1 did also not reveal any alterations in expression pattern of nitrogenrepressed SM-genes suggesting that FfGap1 does not possess an additional regulatory role as a nitrogen-sensing transceptor, which was described for Gap1 and several more nutrient transporters in S. cerevisiae [16,57,61] or the ammonium permease MepB in F. fujikuroi [22].

FfGap1-sorting depends on nitrogen availability
After demonstrating that the expression of FfGAP1 is regulated in an NCR-sensitive and AreA-dependent manner, we were intrigued by the possibility that sorting of FfGap1 to the plasma membrane might, like in yeast, depend on the nitrogen status of the cell. Fluorescence microscopy revealed abundance of FfGap1-Gfp at the plasma membrane during nitrogen starvation and migration of the fusion protein to the lumen of vacuoles. The active translocation of FfGap1 to the organelles has been observed in the time course as moving, fluorescence emitting structures.
The transport of FfGap1-Gfp to the vacuoles upon glutamine addition is similar to the findings in yeast [14,62]. However, in contrast to yeast, addition of proline also led to rapid sorting of FfGap1-Gfp to the vacuoles while nitrate-mediated delocalization of FfGap1-Gfp took longer. This delayed sorting could be explained by the time needed for converting nitrate to glutamine through nitrite and ammonium. Previously we have shown that repression of NCR target genes, e.g. the GA and bikaverin biosynthesis genes, by nitrate takes also much longer than with glutamine and ammonium [41].
Ammonium ions also have a strong negative impact on the plasma membrane localization of FfGap1. This is in contrast to the studies of Roberg and coworkers [63], who found a stimulatory effect of ammonium on Gap1 activity in S. cerevisiae. Contradictory to the findings by Roberg, other authors described the inactivation of Gap1 upon addition of ammonium to poor nitrogen sources, data that supports our present findings [64,65,66]. The apparent differences might be elucidated by different growth conditions and strain backgrounds used by the different researchers.
It is remarkable that cytoplasmic FfGap1-Gfp was detectable at all times, indicating constant sorting of FfGap1. The visible smaller intracellular structures carrying FfGap1-Gfp during nitrogen starvation may be part of the trans-Golgi network or the multivesicular endosome, which both participate in ubiquitination-dependent Gap1 sorting in yeast [27,61,67]. The dynamic regulation of Gap1 sorting in S. cerevisiae allows the cell to rapidly enhance the import of amino acids when internal amino acid levels decrease. On the other hand, strict regulation of Gap1 sorting avoids excessive accumulation of amino acids that were found to be harmful [63,64]. A similar process might be important in F. fujikuroi to respond quickly to changing nitrogen conditions.

Intracellular sorting and stability of FfGap1 is influenced by conserved lysine residues
In yeast it was shown that a set of enzymes facilitate ubiquitination of specific lysine residues at position 9 and 16 of ScGap1, which then leads to endocytosis and subsequent degradation in the vacuole. Mutation of these residues resulted in a stable Gap1 protein at the plasma membrane, even when favored nitrogen sources like ammonium are available [27].
In order to investigate a similar mechanism in F. fujikuroi, we substituted the conserved lysine residues at position 7 and 15 of FfGap1 by alanine. Localization experiments revealed that the mutated protein FfGap1(K7/15A) remains much longer at the plasma membrane after addition of glutamine. In addition, these point mutations resulted in overall increased stability compared to the wild type-protein. According to the yeast-model, it is most likely that FfGap1 is, in a similar way, regulated on a post-translational level by ubiquitination of the lysine residues. However, the actual involvement of ubiquitination has yet to be shown experimentally by monitoring the grade of ubiquitination of FfGap1 and identification of potential proteins that enable ubiquitination. Potential candidates for such proteins could be homologues of the S. cerevisiae ubiquitin-protein ligase Rsp5 or the Bul1 and Bul2 adaptor proteins, which are the key components of Gap1-ubiquitination in yeast [28,29].
Despite these similarities, the loss of the conserved residues K7 and K15 did not result in complete membrane-stabilization of FfGap1 under nitrogen surplus as it has been shown for S. cerevisiae. Furthermore, although the overall stability of the mutated protein was greatly increased in comparison to the wild-type, it was still subject to a nitrogen-induced degradation when glutamine was added, though at a much slower rate. These data suggest that there are either more lysine residues which are targets of ubiquitination, and/or that additional factors might be involved in nitrogen-induced destabilization of the permease. Finding the right target for the precisely regulated ubiquitination will be an interesting aim of future studies.
F. fujikuroi Npr1 homologues slightly influence nitrogen utilization but have no apparent influence on FfGap1 sorting In yeast, the serine/threonine protein kinase ScNpr1 positively influences the stable localization of Gap1 in the membrane during poor nitrogen supply [27]. The genome of F. fujikuroi contains three genes coding for Npr1-like proteins, and complementation of the ScNPR1 yeast mutant with these three candidate genes overcome some, but not all, of the ΔScnpr1 growth defects. FfNpr1-1 provided more Npr1-like effects than the other two FfNpr kinases, especially compared to FfNpr1-2. The protein sequence of FfNpr1-2 is significantly shorter than the yeast Npr1 and the other two F. fujikuroi homologues and contains less serine residues at the N-terminus (S1 Fig), which could explain the low functional conservation of both proteins. The surprising and complete inability of the FfNPR1-1-complemented yeast ΔScNPR1 mutant to grow on proline indicates that FfNpr1-1 seems to inhibit the uptake of proline, which is mediated by the two nitrogen-regulated permeases Gap1 and Put4 as well as two other AAPs, Gnp1 and Agp1 [68,69].
Furthermore, absence of growth of FfNpr1-expressing cells on citrulline and tryptophane reveals that FfNpr1 did neither restore yeast Gap1 functions nor stabilize the protein, indicating significant differences between ScNpr1 and FfNpr1-1. The results of Western analysis confirmed our assumption that FfNpr1-1 is not able to reconstitute Gap1 activity. Only yeast cells transformed with the native ScNpr1 gene were able to prevent degradation of Gap1 in the S. cerevisiae npr1Δ mutant when incubated with urea. Interestingly, activity of FfNpr1-1 in the ΔScNPR1 background seems to enhance the degradation process of Gap1 even more than in the deletion mutant. This could explain the complete inability of the FfNpr1-1-complemented yeast strain to grow on proline, assuming that a similar effect acts on Put4. FfNpr1-2-and FfNpr1-3-complemented yeast strains also displayed a weaker abundance of Gap1. Nevertheless, the decline of Gap1 is much more severe in the FfNpr1 strain, suggesting major regulatory defects caused by FfNpr1-1.
Single, double and triple deletions of FfNPR1 genes in F. fujikuroi revealed no major defects in utilization of certain nitrogen sources. Deletion of FfNPR1-3 resulted in the most severe effect. The growth morphology of the mutant with an enhanced colony diameter but reduced aerial mycelium is reminiscent of the growth of the wild-type on poor nitrogen sources or under overall low nutrient availability. This indicates that the utilization of all tested nitrogen sources might be slightly impaired in the case of ΔFfNPR1-3. Complementation of the mutant with the FfNPR1-3 gene copy fully restored the wild type phenotype. However, we have no explanation why this phenotype is not shared by the ΔΔFfNPR1-1/FfNPR1-3 double or ΔΔΔFfNPR1-1/FfNPR1-2/FfNPR1-3 triple mutant. Currently, the role of all three Npr1 homologues in F. fujkuroi is unknown and needs further investigation. It is also not known by which mechanism FfNpr1-3 affects utilization of nitrogen sources in F. fujikuroi and upon heterologous expression in S. cerevisiae. However, we were able to demonstrate that, unlike in yeast, the tested Npr1 homologues in F. fujikuroi are not exclusively involved in regulating the sorting and stability of FfGap1. These results are in accordance with the inability of the FfNpr kinases to restore Gap1 activity and stability in the npr1Δ yeast mutant.
In summary, we identified a functional homologue of the S. cerevisae Gap1 among the 99 potential AAPs in F. fujikuroi. FfGap1 is able to restore growth defects of the yeast gap1 mutant. Subcellular localization of FfGap1 strictly depends on nitrogen availability but not on the presence of Npr1-like proteins in F. fujikuroi. Furthermore, two conserved lysine residues (K9, K15) are probably involved in ubiquitination-mediated degradation of FfGap1.
For submerse culture experiments F. fujikuroi strains were first grown for 72 h at 28°C in 300 ml Erlenmeyer flasks with 100 ml Darken medium [70] on a rotary shaker. 500 μl of this culture were then used to inoculate 100 ml of ICI (Imperial Chemical Industries, UK) media [71] containing either 6 mM glutamine or 60 mM glutamine. Growth proceeded for 1-3 days on a rotary shaker at 28°C in the dark. For DNA isolation and protoplasting, F. fujikuroi strains were incubated in 100 ml modified ICI medium (Imperial Chemical Industries Ltd., UK) [71] containing 10 g/l fructose instead of glucose and 0,5 g/l (NH 4 ) 2 SO 4 as nitrogen source at 28°C on a rotary shaker at 200 rpm for 3 days and 18 h, respectively. For RNA isolation, the fungal strains were grown in synthetic ICI medium with 6 mM glutamine, 30 mM glutamine or 12 mM NaNO 3 for 3 days. After this time, the mycelia were harvested, washed with deionized water, and flash frozen with liquid nitrogen prior to lyophilization. For fluorescence microscopy the fungal strains were grown in synthetic ICI medium with 6 mM glutamine for 3 days or with 3 mM glutamine for 1 day. After this time the strains were examined via microscopy before and after addition of glutamine, (NH 4 ) 2 SO 4 or NaNO 3 up to a concentration of 12 mM or 24 mM. To compare the growth of the NPR1 deletion mutants with that of the F. fujikuroi wild-type, the strains were grown on solidified (15 g/l agar) ICI-media containing different nitrogen sources as specified in the text. Plates were incubated at 28°C in the dark and pictures of the plates were taken at different time points.
The following S. cerevisiae strains were employed for complementation tests with F. fujikuroi AAP genes: gap1Δdip5Δ mutant strain M4276 (MATa ura3 ΔGAP1; isogenic to strain S288C) [44], and gap1Δssy1Δ mutant strain M4238 (MATa ura3 ΔGAP1 ΔSSY1; isogenic to strain S288C) [10]. For complementation with F. fujikuroi NPR genes, the S. cerevisiae wildtype strain 23344c and the ΔScNPR1 strain 30788a [27] have been used. Yeast cells of strain M4276 and M4238 were grown either in minimal medium without ammonium sulfate (10 g/l succinic acid, 6 g/l NaOH, 1.7 g/l yeast nitrogen base (Difco, Lawrence, USA) and 20 g/l D-glucose) or with addition of nitrogen sources as indicated in the text. Yeast cells of strain 23344c and 30788a were grown in minimal buffered (pH 6.1) medium with 3% glucose as the carbon source [72]. To this medium, nitrogen sources were added as required by the experiment and as specified in the text.

Yeast growth assays
The growth of S. cerevisiae strains transformed with Yep352 (empty plasmid), pScGAP1 (positive control) and plasmids carrying one of the tested F. fujikuroi AAP-encoding genes was tested on plates with minimal medium supplemented with ammonium sulfate or amino acids as described [59]. For growth assays, classical drop tests were performed. Fresh cells of all tested strains were resuspended in water and adjusted to an optical density [OD 600 ] of 1.0. A series of 10-fold dilutions were prepared from each suspension, and 3 μl aliquots were spotted on agar plates as described [59].
For complementing the S. cerevisiae mutant strain M4276 with GAP1 homologous genes from F. fujikuroi, full length cDNA fragments of the AAP-encoding genes FFUJ_09118, FFUJ_11370, FFUJ_05331 and FFUJ_01136 were amplified with the primers FFUJ_09118-PstI/ FFUJ_09118-Not, FFUJ_11370-Sal/FFUJ_11370-Not, FFUJ_05331-Sal/FFUJ_05331-Not, and FFUJ_01136-Pst/FFUJ_01136-Not, respectively, and cloned into the accordingly restricted vector yEX-C [39]. The empty vector was used as a negative control in transformations of S. cerevisiae M4276, whereas yEXPcGAP1 [39] was used as a positive control. pAS103 [33] containing the S. cerevisiae HA-tagged NPR1 gene under the control of the endogenous promoter was used as a positive control for complementation of the S. cerevisiae NPR1 mutant 30788a. For cloning the F. fujikuroi NPR1 homologous genes behind the yeast NPR1 promoter, vector pHA-NPR1 [33] carrying the yeast NPR1 gene with its native promoter, but without the NPR1 terminator, was used as basis vector. To remove the yeast NPR1 gene, the vector was digested with XbaI and religated, resulting in plasmid YEplacRL-Xba. The NPR1 terminator was amplified from S. cerevisiae genomic DNA with primers YNPR1-Term-Xba-F/YNPR1-Term-Pst1-R and cloned into the XbaI/PstI-digested vector YEplacRL-Xba. The resulting YEplac195-npr1-Term was restricted with XbaI and used for cloning the F. fujikuroi NPR1 homologous genes by yeast recombination cloning. For this cloning approach, the full length fragments of FfNPR1-1, FfNPR1-2 or FfNPR1-3 were amplified with the primer pairs NPR1Y-for/NPR1Y-rev, NPR2Y-for/NPR2Y-rev or NPR3Y-for/NPR3Y-rev, respectively, and cloned between the yeast ScNPR1 promotor and the yeast ScNPR1 terminator into the vector YEplac195-NPR1-Term.
For generating the pFfGAP1-GFP fusion vectors, a full-length cDNA clone of FfGAP1 was amplified using the primers FfGAP1-GFP-F-WT and FfGAP1-GFP-R that contain overlapping sequences homologous to the vector pNAN-OGG [75]. This vector contains a hygromycin resistance cassette, and a codon-optimized eGFP [77,80] under control of the Aspergillus nidulans oliC promoter and the gluc terminator. The FfGAP1 PCR product and the NcoI-digested plasmid pNAH-OGG were co-transformed into S. cerevisiae yielding pFfGAP1-GFP. The point-mutated FfGAP1 gene copy (lysine 7 and lysine 15 were exchanged for alanine) was generated by the same approach, but by using primer FfGAP1-GFP-F-Mut in combination with FfGAP1-GFP-R for amplification of the mutated gene allele. DNA of pooled yeast colonies was isolated as described above and transformed into E. coli. Plasmid DNA from single colonies was isolated and sequenced.
Both vectors were transformed into the ΔFfGAP1 mutant. Integration of the whole FfGAP1-GFP fusion constructs was controlled by diagnostic PCR.

Nucleic acid isolation and analysis
Lyophilized mycelium was ground into a fine powder and dispersed (in the case of DNA for use in PCR) in extraction buffer as described by Cenis [81]. DNA for Southern hybridization experiments was prepared following the protocol of Doyle and Doyle [82]. Plasmid DNA was extracted using the Genomed plasmid extraction kit (Genomed, Löhne, Germany). Total F. fujikuroi RNA was isolated using the RNAgents total RNA isolation kit (Promega, Mannheim, Germany). For Southern analysis, genomic DNA was digested to completion with appropriate restriction enzymes, fractionated in 1.0% (w/v) agarose gels, and transferred to nylon membranes (N+, Amersham). DNA probes were randomly labelled and hybridizations were carried out overnight at 65°C.

PCR and RT-PCR
PCR reactions contained 25 ng DNA, 5 pmol of each primer, 200 μM desoxynucleotide triphosphates, 1 unit BioTherm DNA polymerase and 1x concentration of BioTherm buffer (Genecraft GmbH, Lüdinghausen, Germany). The reactions were started with 4 min at 94°C, followed by 35 cycles of 1 min per kb of the product at 94°C, 1 min at 56°to 65°C, 1 min at 70°C, and a final 10 min at 70°C. PCR products were cloned into pCR2.1-TOPO (Invitrogen). Resistance cassettes and eGFP for yeast recombination were amplified with the proofreading Phusion DNA polymerase (Finnzymes, Vantaa, Finland). These reactions contained 25 ng DNA, 5 pmol of each primer, 200 μM desoxynucleotide triphosphates, 1 unit Phusion polymerase and 1x concentration of HF-buffer. The reactions were started with 5 min at 95°C, followed by 35 cycles of 1 min at 94°C, 1 min at 56°to 65°C, 1 min per kb of the product at 72°C, and a final 10 min at 72°C. Primers are listed in S2 Table. Site-directed mutagenesis Site-directed mutagenesis was carried out as described by the manufacturer using the Quik-Change II Site-Directed Mutagenesis Kit (Agilent Technologies). To generate an FfGAP1 template vector for site directed mutagenesis, the full length FfGAP1 cDNA fragment from F. fujikuroi was amplified from cDNA by using the RT-primers FfGAP1-WT-F1 and FfGAP1-WT-R1 (S2 Table), and the PCR fragment was cloned into vector pCR2.1-TOPO (Invitrogen) resulting in vector pGAP1-cDNA. Primers for introducing specific point mutations into the wild-type FfGAP1 cDNA sequence are listed in S1 Table. The mutated gene copy was then cloned into the destination vector containing the eGFP gene and subsequently transformed into the wild-type and ΔNPR1 mutants.

Fungal transformations
Preparation of protoplasts of F. fujikuroi was carried out as described [83]. For deletion of FfGAP1 and single or double deletion of NPR1-homologous genes, strain IMI58289 or single NPR1 deletion strains were transformed with 10 μg of the FfGAP1 or the FfNPR1-1, FfNPR1-2 and FfNPR1-3 replacement fragments of the pΔGAP1 or the pΔNPR1 replacement vectors, respectively. Transformed protoplasts were regenerated at 28°C in a complete regeneration agar (0.7 M sucrose, 0.5 g/l yeast extract) with 100 μg/ml nourseothricin (Werner Agents, Jena, Germany), 100 μg/ml hygromycin or 100 μg/ml geneticin (Sigma-Aldrich, Taufkirchen, Germany) for 6 to 7 days as specified above. Single spore cultures for purification of the heterokaryons were established from the transformants with homologous integration of the replacement cassettes and used for DNA isolation and subsequent PCR and Southern blot analysis.
For FfGap1 localization studies, 107 protoplasts of strain IMI58289, ΔFfNPR1-1 or ΔFfNPR1-3 were transformed with 10 μg of vector pGAP1-Gfp carrying the FfGAP1-GFP construct or pGAP1K7/15A-GFP carrying the mutated FfGAP1K7/K15-GFP construct. The transformed protoplasts were regenerated and selected as described above. For yeast transformation, yeast cells were treated with lithium acetate [84] and transformed according to the method of Sherman et al. [85].

Western blot analysis
Membrane-enriched cell extracts of yeast were prepared as previously described [18]. For blot analysis, equal protein amounts were loaded onto an 8% SDS-polyacrylamide gel in a Tricine system [86].
For detection of Gap1, Mep2 and Pma1, polyclonal anti-Gap1, anti-Mep2 and anti-Pma1 antibodies raised in rabbits were used at 1:5000 dilution [27,87]. To detect the primary antibodies HRP (Horse radish peroxidase) conjugated anti-rabbit secondary antibodies were used in a 1:10000 dilution, followed by visualization of the occurring chemoluminescence.

In silico identification of potential AAPs
The putative AAP proteins were identified in the genome database of the F. fujikuroi wild-type strain IMI58289 [7] by a BlastP search [35] using the protein sequences of S. cerevisiae Gap1 and by searching for following InterPro domains: IPR004841 (amino acid permease domain); IPR004840 (conserved amino acid permease site); IPR004762 (fungal amino acid permease); and IPR013057 (transmembrane amino acid transporter).