OUP user menu

Identification and substrate specificity of a ferrichrome-type siderophore transporter (Arn1p) in Saccharomyces cerevisiae

Petra Heymann, Joachim F. Ernst, Günther Winkelmann
DOI: http://dx.doi.org/10.1111/j.1574-6968.2000.tb09108.x 221-227 First published online: 1 May 2000

Abstract

Genes encoding transporters for heterologous siderophores have been identified in Saccharomyces cerevisiae, of which SIT1, TAF1, and ENB1 encode the transporters for ferrioxamines, ferric triacetylfusarinine C and ferric enterobactin, respectively. In the present communication we have shown that a further gene encoding a member of the major facilitator superfamily, ARN1 (YHL040c), is involved in the transport of a specific class of ferrichromes, possessing anhydromevalonyl residues linked to Nδ-ornithine (ARN). Ferrirubin and ferrirhodin, which both are produced by filamentous fungi, are the most common representatives of this class of ferrichromes. A strain possessing a disruption in the ARN1 gene was unable to transport ferrirubin, ferrirhodin and also ferrichrome A, indicating that the encoded transporter recognizes anhydromevalonyl and the structurally-related methylglutaconyl side-chains surrounding the iron center. Ferrichromes possessing short-chain ornithine-Nδ-acetyl residues such as ferrichrome, ferricrocin and ferrichrysin, were excluded by the Arn1 transporter. Substitution of the iron-surrounding N-acyl chains of ferrichromes by propionyl residues had no effect, whereas substitution by butyryl residues led to recognition by the Arn1 transporter. This would indicate that a chain length of four C-atoms is sufficient to allow binding. Using different asperchromes (B1, D1) we also found that a minimal number of two anhydromevalonyl residues is sufficient for recognition by Arn1p. Contrary to the iron-surrounding N-acyl residues, the peptide backbone of ferrichromes was not an important determinant for the Arn1 transporter.

Keywords
  • Major facilitator superfamily
  • Iron transport
  • Saccharomyces cerevisiae

1 Introduction

Ferrichromes are typical fungal siderophores designed to solubilize and transport iron into fungal cells. They are widespread among basidiomycetous (i.e. Ustilago, Tilletia) and ascomycetous fungi (i.e. Aspergillus, Penicillium, Neurospora, Botrytis) [1]. Ferrichromes represent cyclic hexapeptides or heptapeptides containing three hydroxamate groups (N5-hydroxy-N5-acyl residues) as iron binding bidentates and possess high iron-complex formation constants (K>1029) even in acidic environments. The cyclo-peptidic nature also prevents rapid degradation by proteases making these siderophores the most stable iron-transporting molecules among all siderophores.

Three groups of ferrichrome-type siderophores have been described so far (reviewed in [2]: (1) ferrichromes (i.e. ferrichrome, ferricrocin, ferrichrysin, tetraglycylferrichrome) which contain exclusively ornithine-N5-acetyl-hydroxamic acid residues; (2) ferrichromes that contain branched-chain N5-acyl residues, such as cis-anhydromevalonic acid (ferrirubin), trans-anhydromevalonic acid (ferrirhodin, des(diserylgylcyl)-ferrirhodin (DDF)), trans-β-methylglutaconic acid (ferrichrome A) or malonic acid (malonichrome) and (3) ferrichromes that contain a mixture of different ornithine-N5-acyl residues, as found in asperchromes produced by Aspergillus ochraceus[3]. In most fungi a mixture of different siderophores is produced which may vary depending on the cultivation conditions. As an example, in Aspergillus strains the production of ferricrocin is accompanied by ester-type siderophores (fusarinines and triacetyl-fusarinine C), while in Penicillium strains the production of ferrichrome is accompanied by coprogen. Also Ustilago sp. has been found to produce ferrichrome and ferrichrome A. Similar observations have been made with Neurospora, Gliocladium, Trichoderma and Agaricus bisporus (the common mushrooms) [1]. Dermatophytic fungi like Epidermophyton and Trichophyton have also been shown to synthesize ferrichromes [4]. Several mycorrhizal fungi of the ericoid type produce ferrichromes and fusarinines [5]. Thus, a broad range of fungi are able to produce a variety of siderophores among which the ferrichromes often predominate [6].

The biosynthesis of ferrichromes has been summarized in a recent review [2] showing that in Ustilago maydis two genes are involved, sid1 and sid2, of which sid1 represents the ornithine-N5-oxygenase and sid2 possibly has a function in the biosynthesis of the peptide ring. Furthermore, ferrichrome biosynthesis is regulated by a gene (URBS1) encoding a GATA-type transcription factor containing two putative zinc finger motifs [7]. Although a detailed picture of ferrichrome biosynthesis is still lacking, some important steps in the biosynthesis of ferrichromes are already known [2].

In contrast to many filamentous fungi and most yeasts, Saccharomyces cerevisiae does not synthesize siderophores, but still can utilize a variety of fungal and bacterial siderophores. While initial studies have not conclusively excluded the possibility of exogenous reductive removal of iron from siderophores by the membrane-bound reductases (Fre1p, Fre2p) and subsequent transport of ionic iron via the Fet3/Ftr transport system into cells of S. cerevisiae[8], it now seems clear that several fungal and bacterial siderophores can be transported intact via membrane transporters encoded by genes of the major facilitator superfamily [9]. Of the previously designated unknown major facilitator (UMF) superfamily genes, three have been recently been identified: YEL065w (SIT1) encodes a transporter for ferrioxamine B [10], YHL047c (TAF1) encodes a transporter for fusarinines [11] and YOL158c (ENB1) encodes a transporter for enterobactin [12]. Since most if not all of the previous UMF family genes have a function in siderophore transport they now represent the siderophore iron transporter (SIT) family.

In the present paper we demonstrate that a further gene, YHL040c, designated ARN1, can be assigned to the SIT family since its gene product is required to specifically recognize and transport ferrirubin and related ferrichromes (ferrirhodin and ferrichrome A) which possess branched-chain ornithine-N5-acyl residues.

2 Materials and methods

2.1 Strains and culture conditions

The strain possessing the disrupted gene YHL040c (BY 4742; MATα; his3Δ1; leu2Δ0; lys2Δ0; ura3Δ0; YHL040c::kanMX4) and the parental strain (BY 4742; MATα; his3Δ1; leu2Δ0; lys2Δ0; ura3Δ0) were from Euroscarf, Frankfurt, Germany and were maintained on YPD (1% yeast extract, 2% peptone, 2% glucose) agar and grown in liquid YPD medium at 30°C under aeration on a rotary shaker.

2.2 Siderophores

All siderophores used in the present investigation (Fig. 1) were from the stock of our siderophore collection or were gifts from collaborating laboratories. Ferrichrome and ferrichrome A were prepared from Ustilago sphaerogena. Ferrichrysin was isolated from Aspergillus melleus and ferricrocin was isolated from Aspergillus viridi-nutants[13]. Propionyl- and butyryl-derivatives had been previously synthesized by the Keller-Schierlein group at the ETH Zürich, Switzerland. Asperchrome B1, D1 and DDF [14] were kindly provided by D. van der Helm, Norman, OK, USA. Ferrirubin, asperchromes and DDF were isolated from A. ochraceus according to Jalal et al. [3] and ferrirhodin was isolated from Botrytis cinerea[15]. Hexahydroferrirhodin was obtained by hydrogenation of ferrirhodin using hydrogen and palladium as a catalyst. Enantio-ferrichrome had been earlier synthesized from D-ornithine by Naegeli and Keller-Schierlein [16]. Deferration of hydroxamate siderophores was according to Wiebe and Winkelmann [17] using 8-hydroxyquinoline.

Figure 1

Structural formula of ferrichrome-type siderophores analyzed in the present investigation. Siderophores recognized by the Arn1 transporter are in bold type.

2.3 High-performance liquid chromatography (HPLC) separation

The purity of siderophores was checked by HPLC on a reversed phase column (Nucleosil C18, 5 μm, 4×250 mm, Grom, Herrenberg, Germany) using an acetonitrile/water gradient (6–40%) with 0.1% trifluoroacetic acid (TFA) added to both solvents. HPLC separation was run on a HPLC (LC-10AT pumps, equipped with gradient controler and automatic sampler, Shimadzu, Duisburg, Germany). The detector wavelength was set at 435 or 220 nm, which allowed the detection of hydroxamate siderophores or desferri-siderophores, respectively.

2.4 Growth promotion assays

Growth promotion tests were done as described earlier [10]. SD-soft agar plates (0.67% yeast nitrogen base (Difco, Detroit, MI, USA), 2% glucose, 0,5% agar) were prepared containing 400 μM bathophenanthroline disulfonic acid (BPDS) and inoculated with an overnight-grown yeast culture. Siderophores (10 μM) were pipetted on sterile filter disks (10 μl disk−1), dried in a microwave oven and placed on agar plates. Growth zones were determined after 24 and 48 h of incubation at 30°C.

2.5 Radioactive iron uptake

Strains were grown overnight in chemically-defined SD medium containing 0.67% yeast nitrogen base (Difco, Detroit, MI, USA) and 2% glucose. Transport kinetics were performed in SD medium containing 400 μM BPDS and 3.4 μM 55Fe-siderophores. Desferri-siderophores were labeled with 55FeCl3 (Amersham, UK, 55Fe in 0.1 M HCl). For time-dependent uptake studies cells (1 ml) were harvested at intervals by filtering through nitrocellulose filters and washed with 10 ml cold saline. Siderophore-bound iron taken up by the cells was measured by liquid scintillation counting. The amount of radioactive iron taken up was calculated as pmol per mg dry weight.

3 Results

3.1 Defective use of ferrichrome-type siderophores by the Arn1 mutant

As shown by growth promotion assays (Table 1) the parental strain (BY 4742) utilized a broad collection of siderophores, with the exception of enantio-ferrichrome, which is synthesized from D-ornithine, resulting in a Δ-cis configuration about the iron center. The disrupted strain (BY4742) containing a deletion of the YHL040c ORF (ARN1), failed to utilize ferrirubin, ferrirhodin and ferrichrome A, all of which contain branched-chain ornithine-N5-acyl residues (Fig. 1). Moreover, hexahydroferrirhodin, asperchrome B1 and butyryl-ferrichrysin were excluded by the Arn1 transporter. Thus, disruption of the ARN1 gene, in S. cerevisiae resulted in a strain that was unable to utilize only those ferrichromes that contain branched-chain ornithine-N5-acyl residues, such as ferrirubin and some structurally-related siderophores. Controls with ferrous iron and ferric citrate showed that siderophore-independent mechanisms of iron uptake are still active in both strains, as in these strains the reductases (Fre1, Fre2) and the transporters for ionic iron (Fet3/Ftr and Fet4) are still functional [2,8].

View this table:
Table 1

Growth promotion tests with the parental strain S. cerevisiae BY 4742 and the disrupted strain BY 4742 Δyhl040c (Δarn1)

SiderophoresBY 4742BY4742 Δyhl040c (Δarn1)
Ferrioxamine B++
Ferrioxamine E++
Coprogen++
Triacetylfusarinine C++
Enterobactin++
Ferrichrome++
Ferricrocin++
Propionylferrichrome++
Butyrylferrichrome++/−
Ferrichrysin++
Propionylferrichrysin++
Butyrylferrichrysin++/−
Asperchrome D1++
Asperchrome B1+
Ferrirubin+
Ferrichrome A++/−
Ferrirhodin+
Hexahydroferrirhodin+
Des(diserylglycyl)-ferrirhodin+
Enantio-ferrichrome
Fe-citrate (20× citrate surplus)++
FeSO4 (aerobic FeII/FeIII-mixture)++
H2O
Siderophores affected by the disruption of the ARN1 transporter are in bold type. Growth was tested on agar plates (iron-deficient SD-medium, containing 500 μM BPDS) using filter disks containing 10 μl siderophores (10 μM).
Symbols: +, strong growth; +/−, faint growth; −, no growth.

3.2 Siderophore uptake measurements

The loss of function in the disrupted strain was also demonstrated by transport experiments using 55Fe-labeled siderophores (Fig. 2). Growth promotion tests yield an iron-dependent growth zone, visualized by colony growth after 48 h on an agar medium. Radiolabeled siderophore uptake in liquid medium represents a short time response which directly shows transfer of the labeled molecule into the cells. As shown in Fig. 2, ferrirubin transport was observed in the wild type but was insignificant in the disruptant. A certain amount of 55Fe-ferrirubin (5–10 pmol) was always found associated with cells of both strains when transport assays were performed. However, a significant increase with time was only seen in the wild type whereas the mutant was clearly defect in 55Fe-ferrirubin uptake. As controls, dead cells or cells in the presence of respiratory inhibitors, like sodium azide, showed no uptake, although the non-specific absorption of 55Fe-siderophores (5–10 pmol) was still detected (data not shown). Transport studies with 55Fe-ferrirhodin and 55Fe-ferrichrome A showed similar transport characteristics, confirming the results obtained in growth promotion assays (Table 1). The transport kinetics with 55Fe-ferrichrome A were not as clear as those with 55Fe-ferrirubin and 55Fe-ferrirhodin, which seems to be due to the fact that ferrichrome A tends to bind more strongly to the cells (due to the presence of three carboxyl groups). This was evident with both strains, although ferrichrome A was not used as an iron source by the disrupted strain as shown by growth promotion tests (Table 1). Uptake of 55Fe-ferrichrysin was studied in the disruptant and the parental strain, which both showed nearly indistinguishable uptake rates, confirming that the transport for ferrichrysin was not affected by gene disruption. As ferrichrysin has the same backbone as ferrirubin, ferrirhodin or ferrichrome A, this can be taken as a proof that the ser-ser-gly cyclic peptide backbone present in all these ferrichromes is not an important determinant for the Arn1 transporter.

Figure 2

Transport of 55Fe-labeled ferrirubin, ferrirhodin, ferrichrome A and ferrichrysin. Cells were incubated in SD medium in the presence of labeled siderophores. Samples were taken at intervals, filtered, washed and the radioactivity of the cells was counted in a liquid scintillation counter as described in Section 2. Closed symbols represent uptake by the parental strain and open symbols represent uptake by the disruptant.

55Fe-ferrirubin uptake and competition with ferrirhodin confirmed the specificity of the Arn1-transporter for ferrirubin and ferrirhodin, which is further evidence that structurally-related siderophores are recognized by the same transporter protein (Arn1p). Interestingly, competitive inhibition of 55Fe-ferrirubin uptake by ferrichrysin could be observed, although ferrichrysin is taken up via another (unidentified) ferrichrome transporter (data not shown).

3.3 Importance of the siderophore side-chain structure for Arn1p-mediated uptake

The results with different ferrichrome-type siderophores indicated that only those ferrichromes are transported by Arn1p, which contained branched-chain N-acyl residues. Whereas all ferrichromes possessing short-chain N-acyl residues, such as ferrichrome, ferricrocin and ferrichrysin were excluded. Thus, the main determinant for siderophore recognition by Arn1p is a branched-chain ornithine-N5-acyl residue of the anhydromevalonyl or β-methyl-glutaconyl type. The configuration cis/trans seems to be unimportant, as ferrirubin and ferrirhodin, which possess opposite configurations of the anhydromevalonyl residue, are equally well transported. This was corroborated by the finding that hexahydro-ferrirhodin, lacking the double bond, was transported as well (Table 1). Moreover, alcohol- or acid-functional end-groups of the branched-chain ferrichrome-type siderophores had no influence on transport as indicated by their equal acceptance by the Arn1 transporter.

When ferrichromes possessing unbranched, linear ornithine-N5-acyl residues were compared in growth promotion studies (Table 1), we observed that uptake of propionyl derivatives, such as propionylferrichrome and propionylferrichrysin, were not affected, while the corresponding butyryl derivatives (butyrylferrichrome and butyrylferrichrysin) showed a significantly reduced uptake in the disruptant. The branched-chain siderophores and the butyryl derivatives have a similar 4-C-atomic distance of the methyl group in common, which seems to be the putative determinant for recognition and transport by Arn1p (Fig. 3).

Figure 3

Proposed model for recognition of ferrichromes possessing branched-chain N-acyl residues by the Arn1 transporter.

3.4 Siderophore side-chain numbers affect Arn1-mediated transport

Using the two asperchromes which had been previously isolated from A. ochraceus by van der Helm et al. [3] we were able to demonstrate that the number of branched-chain trans-anhydromevalonyl residues is also an important factor for recognition and transport by the Arn1p. Compared to ferrirubin, possessing three trans-anhydromevalonyl (AMV) residues, asperchrome B1 contains two AMV and one acetyl residue, while asperchrome D1 contains one AMV and two acetyl residues as shown in Fig. 1. When ferrirubin, asperchrome B1 and asperchrome D1 were compared in growth promotion tests using wild type and disruptant, it became evident that ferrirubin and asperchrome B1 were well recognized, whereas asperchrome D1 was not. This result can be taken as evidence that a tripodal iron-surrounding arrangement of N-acyl residues is optimal for recognition, but that two residues are still sufficient to allow recognition and transport by the Arn1 transporter. A significant dependence on the kind and number of N-acyl residues has been reported earlier in a study with the fungus Neurospora crassa, where increasing numbers of branched-chain residues decreased the affinity to the ferrichrome transport system [2,18]. The differences observed with short-chain and branched-chain ferrichrome-type siderophores now is reflected by the existence of two different transporters as shown here for S. cerevisiae, Arn1p recognizing ferrirubin, ferrirhodin and ferrichrome A and other short-chain ferrichromes by a still unidentified ferrichrome transporter.

4 Discussion

The gene, YHL040c (ARN1), previously assigned to the UMF family genes, has been shown in the present investigation to encode a membrane transporter (Arn1p) for a certain group of ferrichromes (ferrirubin, ferrirhodin, ferrichrome A) possessing branched-chain ornithine-N5-acyl residues surrounding the iron center. While some of the ferrichromes, such as ferrichrome, ferricrocin and ferrichrysin contain short-chain acetyl hydroxamic acid residues, the branched-chain ferrichromes contain either trans-5-hydroxy-3-methyl-2-pentenoyl (ferrirubin), cis-5-hydroxy-3-methyl-2-pentenoyl (ferrirhodin) or trans-4-carboxy-3-methyl-2-butenoyl=trans-β-methyl-glutaconyl (ferrichrome A) residues [19,20].

There is no evidence that the peptide backbone is involved in recognition by the Arn1 transporter. While ferrichrysin shares the same ser-ser-gly cyclic peptide sequence with ferrirubin, ferrirhodin and ferrichrome A, it is still excluded by the Arn1 transporter. Additional proof of this observation comes from studies with DDF, a ferrirhodin analog lacking a cyclic backbone which has been found as a natural byproduct in siderophores from A. ochraceus[3]. Although ferric DDF was not well transported by the wild type, it was excluded by the disruptant, confirming that branched-chain N-acyl residues need to interact with the Arn1 transporter even in the absence of any peptide backbone.

Another important aspect of siderophore transport is the stereospecificity of recognition by the SIT. Previous studies with enantio-ferrichrome in the fungus N. crassa had already shown that reversal of the natural Λ-cis configuration of ferrichrome to a Δ-cis configuration in synthetic enantio-ferrichrome, resulted in a complete loss of recognition [21,22]. When enantio-ferrichrome was tested in growth promotion tests with a wild-type strain of S. cerevisiae and its corresponding disruptant no growth was observed, confirming the requirement of a correct configuration of iron-surrounding N-acyl residues for recognition by the (unidentified) ferrichrome transporter.

Although a transporter for ferrichrome, ferricrocin and ferrichrysin could not be identified in the present investigation, we succeeded in assigning the gene YHL040c (ARN1) a function in the transport of ferrirubin and some structurally-related branched-chain ferrichromes, such as ferrirhodin and ferrichrome A. While ferrichromes (narrow sense) possessing short-chain N-acyl residues were not recognized by the Arn1 transporter, elongation of the N-acyl chains from acetyl to butyryl restored recognition, suggesting that the chain length of the N-acyl residues is an important determinant for recognition and transport by the Arn1 transporter in S. cerevisiae.

Thus, the Nδ-acyl residue seems to be an important structural element for recognition by the Arn1 transporter. Other structural differences in the ferrichromes seem to be of minor importance for recognition and transport. The fact that ferrichrysin, although possessing the same peptidic backbone (cycl-ser-ser-gly-orn1-orn2-orn3-) like all branched-chain ferrichromes, is not recognized by Arn1p, underlines the importance of the iron-surrounding residues. Despite the inability to be recognized by the Arn1 transporter, competition can be observed between structurally-unrelated siderophores such as ferrirubin and ferrichrysin when tested in the parental strain. We have made similar observations when the simultaneous uptake of coprogen and ferrichromes was studied in N. crassa, suggesting that the loaded transporters compete with one another during passage through the membrane [23,24]. As the siderophore iron transporters of the SIT family turn out to be very specific with regard to the substrates transported, the competition of structurally-unrelated siderophores may also be interpreted as a downstream effect, which may function inside the cells after transport across the membrane has been completed [10]. As previous studies from our laboratory have shown that the transport of hydroxamate siderophores in N. crassa functions as a proton symport mechanism [26], we would therefore suggest that the transport of hydroxamates in yeasts functions analogous to that of the filamentous fungi.

The proteins of the SIT family are fairly uniform in size (606–637 residues), suggesting similar functions. Three of them (YEL065w (SIT1), YHL047c (TAF1) and YOL158c (ENB1)) have now been assigned a function in siderophore transport. YEL065w and YOL158c are the most divergent members of this gene family and consequently recognize two structurally very different siderophores, the hydroxamate-type ferrioxamines and the catecholate-type enterobactin, respectively. The two closely-related paralogs, YHL047c and YHL040c, which may have arisen from a recent gene duplication event, could now be shown to encode transporters for structurally-related siderophores, the fusarinines and the branched-chain ferrichromes, both having anhydromevalonyl residues linked to N5-ornithine (ARN) as a common structural element. However, some structural variations of the ARN moiety seem to be allowed to be recognized by the Arn1 transporter. The ARN residue may be replaced by β-methyl-glutaconic acid residues as shown in ferrichrome A. Also the trans-configuration is not an absolute structural requirement, as the cis-configuration in ferrirhodin is also accepted by Arn1p.

The MFS–MDR proteins have been grouped into three clusters (I, II, III), of which the SIT proteins belong to cluster III possessing 14 putative transmembrane spanning domains. Some of the MFS–MDR families, as described by Goffeau et al. [25] are mainly associated with specific transport systems that catalyze the efflux of a variety of hydrophobic drugs by substrate-nonspecific mechanisms. It now becomes clear that the SIT family within the MFS superfamily are designed for the import of molecules in S. cerevisiae and that this import is highly specific for certain siderophore classes.

While this work was in progress we learned that Dr. Caroline Philpott (NIDDK, NHI) has also identified YHL040c using cDNA microarrays as an iron- and AFT1-regulated gene (personal communication). To avoid name conflict and to continue with our intention to give names based on transport substrates we have reached a consensus by using ARN1 as an agreed name although its derivation is different.

Acknowledgements

We thank Dick van der Helm for providing siderophores and Nadja Fahrbach for skillful technical assistance.

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
View Abstract