The Siderophore Transporter Sit1 Determines Susceptibility to the Antifungal VL-2397

VL-2397 (previously termed ASP2397) is an antifungal, aluminum-chelating cyclic hexapeptide with a structure analogous to that of ferrichrome-type siderophores, whereby replacement of aluminum by iron was shown to decrease the antifungal activity of this compound. Here, we found that inactivation of an importer for ferrichrome-type siderophores, termed Sit1, renders Aspergillus fumigatus resistant to VL-2397.

ABSTRACT

VL-2397 (previously termed ASP2397) is an antifungal, aluminum-chelating cyclic hexapeptide with a structure analogous to that of ferrichrome-type siderophores, whereby replacement of aluminum by iron was shown to decrease the antifungal activity of this compound. Here, we found that inactivation of an importer for ferrichrome-type siderophores, termed Sit1, renders Aspergillus fumigatus resistant to VL-2397. Moreover, expression of the endogenous sit1 gene under the control of a xylose-inducible promoter (to uncouple sit1 expression from iron repression) combined with C-terminal tagging with a fluorescent protein demonstrated localization of Sit1 in the plasma membrane and xylose-dependent VL-2397 susceptibility. This underlines that Sit1-mediated uptake is essential for VL-2397 susceptibility. Under xylose-induced sit1 expression, VL-2397 also retained antifungal activity after replacing aluminum with iron, which demonstrates that VL-2397 bears antifungal activity independent of cellular aluminum importation. Analysis of sit1 expression indicated that the reduced antifungal activity of the iron-chelated VL-2397 is caused by downregulation of sit1 expression by the imported iron. Furthermore, we demonstrate that defects in iron homeostatic mechanisms modulate the activity of VL-2397. In contrast to A. fumigatus and Candida glabrata, Saccharomyces cerevisiae displays intrinsic resistance to VL-2397 antifungal activity. However, expression of sit1 from A. fumigatus, or its homologue from C. glabrata, resulted in susceptibility to VL-2397, which suggests that the intrinsic resistance of S. cerevisiae is based on lack of uptake and that A. fumigatus, C. glabrata, and S. cerevisiae share an intracellular target for VL-2397.

INTRODUCTION

Invasive aspergillosis (IA) is a life-threatening fungal infection with increasing rates among immunocompromised patients, including patients with AIDS or leukemia as well as bone marrow or organ transplants. Although azoles and polyenes are available as primary treatments for IA, the 6-week all-cause mortality rate remains high (∼20%) for both drug classes (1), which highlights the importance for the development of safe and more effective alternatives (2). The major causal agent of aspergillosis is Aspergillus fumigatus (3).

Recently, VL-2397 (previously termed ASP2397), isolated from the fungus Acremonium persicinum MF-347833, was identified as an antifungal agent (4, 5). VL-2397 is a cyclic hexapeptide with the sequence cyclo{Asn-Leu-dPhe-[(N5-acetyl-N5-hydroxy-Orn)₃]-}Al(III) (C₄₀H₅₉AlN₁₀O₁₃; molecular mass of 915.4 Da; dPhe is d-phenylalanine, while the other amino acids are all L-stereoisomers). VL-2397 resembles a ferrichrome-type siderophore chelating aluminum; compared to the intracellular siderophore of A. fumigatus, ferricrocin, Asn-Leu-dPhe replaces Gly-Ser-Gly in the hexapeptide. The sequence Asn-Leu-dPhe is not found in any other characterized ferrichrome-type siderophores, and a VL-2397 derivative carrying a dPhe-to-Leu exchange did not show antifungal activity, underlining the importance of dPhe in this biological activity (4). Naturally, siderophores are produced to chelate iron for uptake, intracellular storage, or intracellular transport, whereby siderophore biosynthesis and uptake is transcriptionally upregulated during iron starvation (6). A. fumigatus takes up siderophores chelating not only iron but also other metals, such as gallium (7). A. fumigatus produces four siderophores: fusarinine C (FsC), triacetylfusarinine C (TAFC), ferricrocin (FC), and hydroxyferricrocin (HFC) (8), whereby FsC and TAFC are secreted for solubilization and uptake of iron. The iron chelates are taken up by specific siderophore iron transporters. FC is used for hyphal iron storage and distribution (9), while its derivative, HFC, is required for conidial iron storage (10). In addition to uptake of endogenously secreted siderophores, A. fumigatus is able to import and utilize xenosiderophores (siderophores produced by other microorganisms), such as ferrichromes or ferrioxamines (11). A. fumigatus possesses five potential siderophore transporters, which are repressed during iron sufficiency via the transcription factor SreA (12). Two of them, Sit1 and Sit2, were shown to transport ferrichrome (13).

VL-2397 is an aluminum chelate; exchange of aluminum by gallium preserved, while exchange by iron impaired, its antifungal activity (4, 5). These findings suggest that VL-2397-mediated uptake of gallium or aluminum contributes to the antifungal activity or, alternatively, that VL-2397 interferes with iron homeostasis of A. fumigatus. In this respect, it is important to note that inactivation of the two high-affinity iron acquisition systems, siderophore-mediated iron uptake and reductive iron assimilation, blocks growth of A. fumigatus under standard growth conditions. Moreover, chelates of the siderophore ferrioxamine with gallium or aluminum were shown to possess antimicrobial activity (14).

VL-2397 successfully passed clinical phase I, while a phase II clinical trial evaluating VL-2397 as a potential first-line treatment for invasive aspergillosis was discontinued in 2019 for strategic reasons (www.vical.com/investors/news-releases). Regardless, VL-2397 represents a highly attractive compound in a new class of antifungal drugs with a potentially novel mode of action. VL-2397 was shown to exhibit activity against a broad range of clinically relevant azole-resistant A. fumigatus isolates (15) and is consequently promising in the context of increasing azole resistance (16). Interestingly, VL-2397 also displayed activity against Candida glabrata but not against Saccharomyces cerevisiae, a closely related yeast. The reason for this difference in susceptibility was enigmatic and might point to VL-2397’s mode of action.

In this study, we demonstrated that (i) uptake of VL-2397 in A. fumigatus depends on the siderophore transporter Sit1, (ii) the antifungal activity of VL-2397 is independent of cellular importation of aluminum, (iii) iron availability and defects in iron homeostatic mechanisms in A. fumigatus modulate the activity of VL-2397, and (iv) the intrinsic VL-2397 resistance of S. cerevisiae is only based on the lack of its uptake.

RESULTS

Inactivation of the siderophore-iron transporter Sit1 renders A. fumigatus resistant to VL-2397.

The ferrichrome-type structure suggested that VL-2397 could be transported into A. fumigatus cells via a siderophore transporter. A. fumigatus has been shown previously to take up ferrichrome-type siderophores (8, 11). Park et al. demonstrated that the siderophore-iron transporters 1 (Sit1; AFUA_7G06060) and 2 (Sit2; AFUA_7G04730) mediate uptake of ferrichrome in A. fumigatus (13). To analyze a potential function of Sit1 or Sit2 in uptake of VL-2397, we generated mutants lacking either the sit1 or the sit2 gene, here referred to as the Δsit1 or Δsit2 strain, respectively. Therefore, the coding regions of sit1 or sit2 were replaced in A. fumigatus strain AfS77 by hygromycin and pyrithiamine resistance cassettes, respectively, as described in Materials and Methods. AfS77 is a derivative of A. fumigatus ATCC 46645 largely lacking nonhomologous recombination (akuA::loxB), which facilitates genetic manipulation (17, 18).

VL-2397 in vitro activity was evaluated by determination of the MIC according to the recommendation of EUCAST (19). A. fumigatus AfS77 and ATCC 46645 displayed a MIC of 1.0 mg/liter for VL-2397 (Table 1), which is comparable to that of other A. fumigatus isolates (15). Inactivation of Sit1 (Δsit1 strain) increased the MIC to >16 mg/liter, i.e., it rendered A. fumigatus highly resistant to VL-2397, while inactivation of Sit2 did not affect VL-2397 susceptibility (Table 1). These results showed that susceptibility of A. fumigatus to VL-2397 requires uptake by the siderophore transporter Sit1, while Sit2 appears uninvolved.

TABLE 1.

VL-2397 susceptibility depends on Sit1 and is increased by defects in iron homeostatic mechanisms

Strain (background)	Description	MICa (mg/liter)
ATCC 46645	Wild-type isolate	1
AfS77 (ATCC 46645)	Derivative of wild-type lacking nonhomologous end joining	1
Δsit1 (AfS77)	Lacks the siderophore transporter Sit1	>16
Δsit2 (AfS77)	Lacks the siderophore transporter Sit2	1
ΔsidF (ATCC 46645)	Lacks extracellular siderophores	0.125
ΔsidC (ATCC 46645)	Lacks intracellular siderophores	0.25
ΔsreA (AfS77)	Lacks the transcription factor repressing iron uptake under iron sufficiency	0.125
ΔhapX (AfS77)	Lacks the major transcription factor mediating adaptation to iron starvation and iron excess	1
ΔsrbA (AfS77)	Lacks the transcription factor regulating adaptation to iron starvation, adaptation to hypoxia, and ergosterol biosynthesis	1
Δyap1 (ATCC 46645)	Lacks the major transcription factor mediating exogenous antioxidative defense	0.5

^aThe MIC was analyzed according to EUCAST recommendations.

VL-2397 susceptibility depends on iron availability.

In A. fumigatus, expression of sit1 is repressed by iron, which is mediated by the iron-regulatory transcription factor SreA (12), suggesting that VL-2397 uptake and, consequently, susceptibility are affected by iron availability. In agreement with this, an increase of the iron concentration by 0.03 mM increased the VL-2397 MIC for AfS77 from 1 to 2 mg/liter (Table 2), and, more impressive, the addition of bathophenanthroline disulfonate (BPS) to a final concentration of 0.2 mM decreased the MIC to 0.06 mg/liter. BPS is a ferrous iron-specific chelator that inactivates reductive iron assimilation and consequently activates the siderophore system, including Sit1, as this is the only high-affinity iron acquisition system remaining under this condition (20). Similarly, supplementation with BPS increased susceptibility and iron supplementation decreased susceptibility to VL-2397 when A. fumigatus was grown on minimal medium plates (Fig. 1). In accordance with the MIC analysis, the Δsit1 mutant displayed resistance against VL-2397 under all conditions, while the Δsit2 mutant showed only slightly increased resistance at 0.5 mg/liter VL-2397 in the presence of BPS. Taken together, these data demonstrated that VL-2397 susceptibility is affected by iron availability, which is in agreement with the transcriptional regulation of sit1 by iron.

TABLE 2.

VL-2397 susceptibility is affected by iron availability and is independent of cellular aluminum import^a

% xylose and strain	MICb (mg/liter) for:
	VL-2397			AS2488053
	RPMI + 0.2 mM BPS	RPMI	RPMI + 0.03 mM Fe	RPMI + 0.2 mM BPS	RPMI	RPMI + 0.03 mM Fe
0% xylose
AfS77	0.06	1	2	0.5	>16	>16
sit1xylP	>16	>16	>16	>16	>16	>16
sit1xylP-venus	>16	>16	>16	>16	>16	>16
1% xylose
AfS77	0.06	1	2	0.25	>16	>16
sit1xylP	0.03	0.125	0.125	0.06	0.125	0.125
sit1xylP-venus	0.03	0.06	0.06	0.03	0.06	0.06

^aThe MIC of A. fumigatus AfS77 and strains expressing sit1 under the control of xylose-inducible xylP promoter with (sit1^xylP-venus) and without (sit1^xylP) C-terminal GFP tagging under conditions of different iron availability in the absence (xylP-repressing) or presence (xylP-inducing) of 1% xylose. For iron supplementation, FeS0₄ was used.

^bThe MIC was analyzed according to EUCAST recommendations.

FIG 1.

VL-2397 susceptibility is affected by iron availability. A total of 10³ spores of AfS77, Δsit1, and Δsit2 strains were point-inoculated on minimal medium reflecting iron starvation (−Fe/BPS; containing 0.2 mM the iron-specific chelator bathophenanthroline disulfonate), iron starvation without iron chelator (−Fe), and iron sufficiency (+Fe; 0.03 mM FeSO₄) containing 0 to 1 mg/liter VL-2397. Photos were taken after incubation for 48 h at 37°C.

Defective intra- or extracellular siderophore biosynthesis as well as lack of SreA increases susceptibility to VL-2397.

Due to iron-dependent VL-2397 efficacy, we analyzed the susceptibility of previously described A. fumigatus mutant strains with defects in iron homeostatic mechanisms (Table 1). Blocking extracellular siderophore biosynthesis (ΔsidF strain; lacks the acyltransferase SidF, which is essential for biosynthesis of fusarinine C and TAFC [8]) or intracellular siderophore biosynthesis (ΔsidC strain; lacks the nonribosomal peptide synthetase SidC, which is essential for biosynthesis of ferricrocin and hydroxyferricrocin [8]) decreased the MIC to 0.125 and 0.25 mg/liter, respectively. Blocking of extra- and intracellular siderophore biosynthesis impairs iron acquisition and intracellular handling of iron, which causes increased iron starvation (8). Therefore, the most likely explanation for these results is the increased expression of sit1 and consequently increased VL-2397 uptake in these mutant strains. Lack of the iron-regulatory transcription factor SreA (ΔsreA strain [12]) decreased the MIC to 0.125 mg/liter. Lack of SreA causes transcriptional derepression of genes involved in iron uptake during iron sufficiency, including the Sit1-encoding gene (12). Again, the increased uptake of VL-2397 is the most likely reason for the hypersusceptibility of this mutant strain. The lack of the transcription factor HapX, which is crucial for adaptation to iron starvation (21), or SrbA, which regulates ergosterol biosynthesis and coordinates adaptation to iron starvation and hypoxia (22), did not affect VL-2397 susceptibility. The lack of the transcription factor Af-Yap1, which coordinates the response to oxidative stress (23), slightly decreased VL-2397 susceptibility (MIC of 0.5 mg/liter). As lack of Af-Yap1 causes hypersusceptibility to oxidative stress, these data indicate that VL-2397 toxicity includes oxidative stress. In conclusion, genetic defects resulting in increased iron starvation or derepression of the siderophore system result in increased VL-2397 susceptibility.

Aluminum is not essential for the antifungal activity of VL-2397.

Previously, exchange of aluminum by iron in VL-2397, leading to a compound termed AS2488053, impaired the antifungal activity (4, 5). In agreement with this, we did not observe antifungal activity by AS2488053 under standard conditions or in media supplemented with 0.03 mM iron (Table 2). However, in the presence of 0.2 mM BPS, AS2488053 had a MIC of 0.5 mg/liter (Table 2). This was still 8.3-fold higher than the MIC of VL-2397 under the same conditions, confirming the importance of aluminum for the antifungal activity of VL-2397. Alternatively, cellular uptake of iron via AS2488053 might be sufficient to repress the expression of the sit1 gene and consequently block further uptake in a negative feedback regulation, leading to protection against the drug.

To uncouple sit1 expression from iron repression, we replaced the sit1 promoter with the xylose-inducible xylP promoter (24) in the AfS77 background with (sit1^xylP-venus strain) or without (sit1^xylP strains) C-terminal tagging of Sit1 with the yellow fluorescence protein derivative Venus, as described in Materials and Methods.

To confirm the xylP-controlled expression of sit1 in the mutant strains, we conducted Northern blot analysis of sit1 transcript levels and, as internal controls, of sit2, mirB (AFUA_3G03640), and cycA (AFUA_2G13110) (Fig. 2). Like sit1, sit2 and mirB genes are repressed by iron, while the cytochrome c gene is induced by iron (12, 25, 26). As shown previously in A. fumigatus ATCC 46646, sit1 expression was repressed during iron sufficiency in AfS77. In contrast, expression of xylP promoter-driven sit1 was induced by xylose but became independent of iron availability (during iron starvation and sufficiency), while expression of sit2, mirB, and cycA showed a wild-type-like iron regulation in all strains.

FIG 2.

xylP promoter allows xylose-induced conditional expression of sit1 during both iron starvation and sufficiency. A. fumigatus AfS77, sit1^xylP-venus, and sit1^xylP strains were grown with or without 1% xylose in liquid minimal medium for 17 h at 37°C under iron starvation (−Fe) and sufficiency (+Fe; 0.03 mM FeSO₄) conditions. Total RNA was isolated and subjected to Northern blot analysis of the indicated genes. Ethidium bromide-stained RNA serves as a control for loading and quality of RNA. Exposure of films was the same for all genes except sit2. The sit2 film had to be exposed 4 times longer, indicating lower transcript levels.

Venus tagging of xylP promoter-controlled sit1 allowed monitoring of sit1 expression at the cellular level. In agreement with the transcriptional analysis (Fig. 2), epifluorescence microscopy demonstrated that the Sit1-Venus protein production is dependent on induction with xylose in the A. fumigatus sit1^xylP-venus strain (Fig. 3A). Moreover, these data also confirmed localization of Sit1 at the plasma membrane, as expected for a siderophore importer. Venus tagging of xylP promoter-controlled sit1 also allowed monitoring of sit1 expression at the protein level. Western blot analysis confirmed that the Sit1-Venus protein is produced during iron starvation and sufficiency in a xylose induction-dependent manner (Fig. 3B). Interestingly, the Sit1-Venus protein level was higher during iron starvation than during iron sufficiency. The latter indicates an influence of iron on the xylP promoter activity, sit1 transcript stability, or Sit1 protein stability.

FIG 3.

In the A. fumigatus sit1^xylP-venus strain, production and plasma membrane localization of the Sit1-Venus protein are dependent on xylose induction. (A) For epifluorescence microscopy, 2 × 10⁴ A. fumigatus sit1^xylP-venus conidia were grown in chamber slides containing 0.2 ml minimal medium for 16 h at 37°C, followed by incubation for another 1.5 h with or without 0.1% xylose. For iron starvation conditions (−Fe), iron-depleted minimal medium with 0.1 mM BPS was used. Differential interference contrast (DIC) and epifluorescence microscopy (Sit1-Venus) are shown. (B) Western blot analysis of cellular Sit1-Venus levels in the A. fumigatus sit1^xylP-venus strain during iron starvation (−Fe) and sufficiency (+Fe). The A. fumigatus sit1^xylP-venus strain was cultivated for 17 h at 37°C in liquid minimal medium under iron starvation (−Fe) and sufficiency (+Fe; 0.03 mM FeSO₄). Subsequently, cultures were incubated for another 0.5 h or 1.5 h with 0.1% xylose or 1.5 h without xylose (−). As a control for antibody specificity, AfS77 was cultivated under iron starvation (−Fe) for 17 h at 37°C in liquid minimal medium, followed by another 1.5 h with 0.1% xylose. Protein extraction and Western blot analysis were conducted as described in Materials and Methods. Sit1-Venus was detected with a monoclonal anti-GFP mouse antibody. The detected molecular mass (about 89.9 kDa) fits the expected molecular mass of a fusion protein of Sit1 (63.1 kDa) and Venus (26.8 kDa). Loading of equal protein amounts was confirmed by Coomassie staining.

In the next step, we analyzed the MIC of VL-2397 and AS2488053 for these strains under sit1 noninducing (without xylose) and inducing (with 1% xylose) conditions with different iron availabilities (Table 2). Notably, AfS77 displayed the same VL-2397 susceptibility with and without xylose. Without xylose, the sit1^xylP-venus and sit1^xylP mutant strains were resistant (MIC of >16 mg/liter) against both VL-2397 and AS2488053 under all conditions tested, which is in agreement with extremely low xylP promoter activity under noninducing conditions, as shown by Northern analysis (Fig. 2). In contrast, with xylose, the VL-2397 MIC for the sit1^xylP strain under standard conditions (RPMI) was even lower than that for AfS77 (0.125 versus 1 mg/liter), which correlated with higher sit1 expression, as shown by Northern analysis (Fig. 2). Iron supplementation did not affect the susceptibility, but iron chelation by BPS still decreased the VL-2397 MIC (0.03 mg/liter) for the sit1^xylP strain. Remarkably, Venus tagging of Sit1 (sit1^xylP-venus strain) additionally decreased the MIC under standard conditions and with iron supplementation (MIC of 0.125 mg/liter versus 0.06 mg/liter). These data indicate that Venus tagging increased the activity of Sit1. As Northern analysis indicated similar sit1 transcript levels in both mutant strains under inducing conditions, green fluorescent protein (GFP) tagging might increase Sit1 activity by raising the protein half-life and/or plasma membrane retention. Remarkably, under xylP-inducing conditions, AS2488053 had the same MIC as VL-2397 for sit1^xylPand sit1^xylP-venus strains. These data demonstrate that aluminum import is not crucial for the antifungal activity of VL-2397.

Exchange of aluminum by iron impairs the antifungal activity of VL-2397 via repression of uptake.

The fact that AS2488053 was as active as VL-2397 in strains lacking repression of sit1 expression by iron indicated that iron import affects AS2488053 susceptibility. Therefore, we compared the short-term effects of VL-2397 and AS2488053 on expression of iron-regulated genes (sit1, sit2, mirB, and cycA; described above) under iron starvation (Fig. 4). In contrast to VL-2397, AS2488053 indeed repressed expression of sit1 mRNA (and other genes that are repressed by iron excess) but induced cycA expression, which is induced by iron. These data suggest that the iron import mediated by uptake of AS2488053 causes repression of sit1, which blocks further uptake of the drug. Therefore, the intracellular concentration reached by AS2488053 is expected to be significantly lower than that of VL-2397, which explains the differences in antifungal activity between these compounds.

FIG 4.

In contrast to VL-2397, AS2488053 represses transcription of iron-repressed genes, including sit1, and induces iron-induced genes. A. fumigatus AfS77 was grown in liquid minimal medium for 16 h at 37°C under iron starvation conditions. Subsequently, VL-2397 or AS2488053 was added to a final concentration of 4 mg/liter for 0.5 and 1.5 h, respectively, before harvesting the mycelia. Total RNA was isolated and subjected to Northern blot analysis of the indicated genes. Ethidium bromide-stained RNA serves as the control for loading and quality of RNA. Exposure of films was the same for all genes except sit2. The sit2 film had to be exposed 4 times longer, indicating lower transcript levels.

Expression of sit1 from either A. fumigatus or C. glabrata renders S. cerevisiae susceptible to VL-2397.

Interestingly, we found that S. cerevisiae, but not C. glabrata, displays resistance against VL-2397 (see below and data not shown) despite the fact that S. cerevisiae possesses four siderophore transporters, while C. glabrata has a single siderophore transporter (27, 28). To analyze if S. cerevisiae lacks VL-2397 uptake or the target for VL-2397 antifungal activity, we expressed sit1 from either A. fumigatus or C. glabrata with C-terminal GFP tagging under the control of the MUP1 promoter in S. cerevisiae on a centromeric low-copy-number plasmid, as described in Materials and Methods. The effect of VL-2397 was analyzed by measuring growth of four S. cerevisiae strains containing the same plasmid backbone (pRS415) in the presence of different VL-2397 concentrations (Fig. 5A). Growth of S. cerevisiae strains carrying the basic backbone plasmid or a plasmid mediating expression of the methionine permease-encoding gene MUP1 (C-terminally tagged with GFP under the control of the MUP1 promoter) was not affected by VL-2397, underlining intrinsic VL-2397 resistance of S. cerevisiae. In contrast, the plasmids mediating expression of sit1 from either A. fumigatus or C. glabrata caused similar VL-2397 concentration-dependent growth inhibition. Resistance and susceptibility of the four S. cerevisiae strains is also seen in agar disc diffusion assays (Fig. 5B).

FIG 5.

Expression of sit1 from either A. fumigatus or C. glabrata renders S. cerevisiae susceptible to VL-2397. (A) Growth curves for S. cerevisiae producing MUP1-GFP, C. glabrata producing Sit1-GFP, or A. fumigatus producing Sit1-GFP or containing the plasmid backbone (pRS415) with different VL-2397 concentrations. According to EUCAST recommendations, incubation was for 24 h at 35°C and the VL-2397 MIC for S. cerevisiae expressing Sit1-GFP from A. fumigatus or C. glabrata was 1 mg/liter. In contrast to EUCAST guidelines, the assay was conducted with YNB medium, as RPMI was incompatible with the yeast strains. (B) VL-2397 agar disc diffusion assay with yeast strains with and without Sit1 production. For this assay, 4 × 10⁵ CFU of the respective strain were plated on YNB agar and 3.2 μg VL-2397 was spotted on the disc. Pictures were taken after 48 h at 30°C.

These data demonstrate that the target for antifungal activity of VL-2397 is conserved in A. fumigatus, C. glabrata, and S. cerevisiae and that the intrinsic resistance of S. cerevisiae is due to lack of VL-2397 uptake.

DISCUSSION

VL-2397 is an antifungal drug with a novel mechanism of action with a ferrichrome-type structure with unusual amino acids in the siderophore backbone. Here, we demonstrate that cellular uptake of VL-2397 in A. fumigatus is mediated by the siderophore transporter Sit1. We have shown previously that A. fumigatus possesses five siderophore transporter-encoding genes, which are repressed by iron via the iron regulator SreA (12). Iron repression supports the role of the encoded transporters in iron acquisition. The substrate specificity of three of these transporters has been partially characterized. MirB transports the endogenous secreted siderophore TAFC (26, 29), Sit1 and Sit2 have been shown to transport ferrichrome, and Sit1 additionally mediates uptake of ferrioxamine B (13). Ferrichrome-type siderophores are produced exclusively by fungal species, while ferrioxamines are produced by actinomycetes. The transport of ferrichrome and VL-2397 by Sit1 demonstrates that this transporter accepts structurally quite diverse ferrichrome-type siderophores. The fact that inactivation of Sit1 rendered A. fumigatus resistant to VL-2397 suggested that Sit2 does not play a major role in VL-2397 uptake. In agreement, inactivation of Sit2 did not affect VL-2397 susceptibility. These data indicate that Sit1 and Sit2 have a certain substrate specificity among ferrichrome-type siderophores, as both transporters have been shown to transport ferrichrome. C. glabrata possesses a single siderophore transporter, termed Sit1, which was shown to transport ferrichrome (28). In contrast, S. cerevisiae employs four siderophore transporters, whereby two, Arn1 and Arn3/Sit1, were shown to transport ferrichrome-type siderophores (27). Nevertheless, we found that C. glabrata is susceptible to VL-2397 while S. cerevisiae is resistant, despite the fact that these yeast species are phylogenetically closely related. We found that expression of the Sit1-encoding gene from either A. fumigatus or C. glabrata renders S. cerevisiae susceptible to VL-2397, indicating that none of the four siderophore transporters of S. cerevisiae accepts VL-2397 as the substrate and that its intracellular target is conserved in this yeast. Again, these data demonstrate that there is a substrate specificity between ferrichrome-type siderophores and siderophore transporters.

A. fumigatus secretes neither ferrichrome-type siderophores nor ferrioxamine B, which are the known molecules transported by Sit1 and Sit2. Therefore, it is not expected that lack of Sit1 or Sit2 causes effects in the absence of such compounds. In agreement, deletion of the sit1 or sit2 gene in A. fumigatus was phenotypically inconspicuous in plate growth assays (see Fig. S2 in the supplemental material). Furthermore, deletion of sit1 or sit2 affected neither siderophore production (Fig. S3) nor susceptibility to voriconazole (MIC of 0.5 mg/liter) or amphotericin B (MIC of 1 mg/liter).

VL-2397 is an aluminum chelate, and previous studies demonstrated that exchange of aluminum for gallium preserved antifungal activity while exchange for iron impaired its antifungal activity (4, 5). Expression analysis revealed that uptake of AS2488053 (VL-2397 chelating iron instead of aluminum), but not VL-2397, leads to repression of sit1 expression. As uptake of AS2488053, but not VL-2397, leads to cellular import of iron, these data are in agreement with the previously reported iron repression of sit1 (12). In accordance with the iron-regulated uptake of VL-2397, we found that iron availability as well as defects in iron homeostatic mechanisms at the genetic level affect VL-2397 susceptibility.

Conditional expression via a xylose-inducible promoter, which allowed uncoupling of sit1 expression from iron regulation, resulted in similar antifungal activity of VL-2397 and AS2488053, demonstrating that cellular import of aluminum does not play a crucial role in the antifungal activity of VL-2397.

In healthy individuals, plasma iron is bound to transferrin with an iron saturation of transferrin of 15 to 45%, which limits pathogens’ access to iron. However, under pathological conditions, non-transferrin-bound iron develops once the binding capacity of transferrin is exceeded (30). For example, iron-overloaded patients with myelodysplastic syndrome or acute myeloid leukemia undergoing hematopoietic stem cell transplantation are at risk of developing non-transferrin-bound iron (30). The underlying mechanisms include chemotherapy-induced shutdown of erythropoiesis decreasing iron consumption, tissue damage causing iron release, as well as frequent red blood cell transfusion therapy for anemia. Several studies have proposed iron overload as a risk factor for the occurrence of infections, including invasive aspergillosis after transplantation (30–32). In agreement, non-transferrin-bound iron was recently shown to promote growth of A. fumigatus serum in vitro (33). As iron overload is a frequent problem in invasive aspergillosis patients and as iron decreases the activity of VL-2397 in vitro, it might be interesting to consider a combination of VL-2397 with iron chelation therapy.

Taken together, in this study we characterized (i) the cellular uptake of VL-2397, (ii) the role of iron regulation in VL-2397 activity, (iii) the role of aluminum in the antifungal activity of VL-2397, and (iv) the reason for intrinsic resistance of S. cerevisiae against VL-2397. Further studies will be required to elucidate the mode of action of the antifungal activity.

MATERIALS AND METHODS

A. fumigatus strains and growth media.

A. fumigatus strains used in this study are summarized in Table S1 in the supplemental material. For spore production, plate growth assay, and Northern analysis, Aspergillus fumigatus strains were cultivated at 37°C on Aspergillus minimal medium, according to reference 34, containing 1% glucose as a carbon source, 20 mM glutamine as a nitrogen source, and 0.03 mM FeSO₄ as an iron source. To generate iron starvation, addition of iron was omitted. Complex medium used for plate assays contained 20 g/liter glucose, 2 g/liter peptone, 1 g/liter yeast extract, 0.2 g/liter Casamino Acids, salts, and trace elements (34).

Inoculum preparation and antifungal susceptibility testing for A. fumigatus.

A. fumigatus strains were inoculated on solid minimal medium agar plates and incubated for 72 h at 37°C. Fungal spores were collected from agar plates by adding spore buffer containing 0.9% saline and 0.01% Tween 20, followed by conidial filtration using a 40-μm nylon cell strainer. Antifungal susceptibility testing of VL-2397 and AS2488053, the use of stock solutions, range of concentrations tested, and the quality control procedures are in accordance with the guidelines of the European Committee on Antimicrobial Susceptibility Testing (EUCAST) (19, 35, 36). In short, RPMI medium (with glutamine, without bicarbonate, and with pH indicator; R6504; Sigma-Aldrich) supplemented with 2% glucose was used as an assay medium. Conidium counting was performed in a hemocytometer, and an inoculum size of 5 × 10⁴ CFU/ml was used. The MIC was visually read after 48 h of incubation at 35°C in biological triplicates.

S. cerevisiae strains and growth media.

All experiments were performed with S. cerevisiae SEY6210 (MATα leu2-3,112 ura3-52 his3-200 trp1-901 lys2-801 suc2-9); strains used in this study are summarized in Table S2 in the supplemental material. For growth under standard conditions, S. cerevisiae cells were incubated in yeast nitrogen base (YNB) synthetic medium (6.7 g/liter YNB with 20 g/liter glucose and ammonium sulfate) and supplemented with amino acids and nucleobases: 20 mg/liter arginine, 50 mg/liter lysine, 200 mg/liter threonine, 30 mg/liter tyrosine, 20 mg/liter histidine, 20 mg/liter tryptophan, 60 mg/liter leucine, 20 mg/liter adenine hemisulfate, and 20 mg/liter uracil. For plasmid-containing yeast strains, leucine was omitted from the YNB growth media because plasmid maintenance requires the absence of leucine, as it is based on the selection marker LEU2.

Inoculum preparation and antifungal susceptibility testing for S. cerevisiae.

For yeast suspensions, S. cerevisiae strains were inoculated on solid YNB agar plates and incubated for 48 h at 30°C. Cells were scraped from the agar plates and dissolved in sterile water.

VL-2397 susceptibility testing of S. cerevisiae strains was carried out in YNB synthetic medium lacking leucine and methionine. The standard medium for antifungal susceptibility testing, RPMI, could not be used, as it contains leucine and methionine, which was incompatible with the yeast strains used in this study, because plasmid maintenance requires the absence of leucine (described above) and the MUP1 promoter used for expression of sit1 genes is repressed by methionine (37). Yeast counting was performed in a hemocytometer, and an inoculum size of 1 × 10⁵ CFU/ml was used for susceptibility testing. Otherwise, MIC testing was performed according to EUCAST guidelines in biological triplicates (38).

VL-2397 preparation and drug dilution method.

VL-2397 was dissolved in 100% dimethyl sulfoxide (DMSO) to obtain a concentration of 1,600 mg/liter and was stored at 4°C. The solution was diluted 50-fold in RPMI medium to prepare a drug working solution of 32 mg/liter and was stored at –20°C. This drug working solution was used for generating 2-fold serial dilutions in 96-well plates from 16 to 0.03 mg/liter, according to EUCAST guidelines. The maximum concentration of DMSO in the susceptibility tests was 1% at 16 mg/liter VL-2397. VL-2397 and its metal-free derivative were supplied by Vical Inc.

Preparation of AS2488053 (VL-2397 iron form).

A total of 5.0 mg of apoVL-2397 (5.60 μmol) was dissolved in 50% dimethyl sulfoxide–50% water to obtain a concentration of 1 mg/ml. Six milligrams (36.9 μmol) of FeCl₃ was dissolved in 0.1 ml HCl (0.1 M) and added to the apoVL-2397 solution, followed by addition of 20 μl of a sodium-acetate trihydrate solution (1.14 M) for pH adjustment. The mixture was incubated for 15 min at room temperature and analyzed by reversed-phase high performance liquid chromatography (RP-HPLC). Here, the iron-containing product was purified by preparative RP-HPLC to give a dark red-colored solid after freeze drying. Analytical data were the following: [Fe]VL-2397, 4.56 mg (4.82 μmol); matrix-assisted laser desorption ionization–time of flight mass spectrometry, m/z [M⁺H⁺] = 944.16 (C₆₉H₉₂FeN₈O₁₅; exact mass, 944.3 [calculated]).

Determination of siderophore production.

Production of extracellular (TAFC) and intracellular (FC) siderophores was measured as described previously (39).

Generation of A. fumigatus Δsit1, sit1^xylP, sit1^xylP-venus, and Δsit2 mutant strains.

For generation of plasmid pMMHL70 containing the Δsit1 deletion cassette, PCR amplifications with oligonucleotides MM322 and MM323 (termed MM322/323), MM324/332, and MM330/333 were performed to amplify a 5′-homologous region of sit1 (AFUA_7G06060), a hygromycin resistance cassette, and the 3′-homologous region of sit1, respectively. The resulting fragments were assembled by a NEBuilder reaction (NEBuilder HiFi DNA assembly; New England Biolabs) in a pUC19 (Thermo Fisher) backbone linearized with primers MM124/125. For transformation of A. fumigatus, a fragment amplified from plasmid pMMHL70 with primers MM322/331 was used.

For generation of plasmid pMMHL69, used for insertion of the xylP promoter and C-terminal tagging of sit1, PCR amplifications with oligonucleotides MM322/323, MM324/325, MM326/327, MM328/329, and MM330/331 were performed to amplify a 5′-homologous region of sit1, a hygromycin resistance/xylanase promoter cassette containing the hygromycin resistance cassette from plasmid pAN7-1 (40), the xylP promoter from Penicillium chrysogenum (24), the protein-encoding sequence of sit1, the protein-encoding sequence for the GFP derivative Venus-encoding gene (41), and the 3′-homologous region of sit1, respectively. The resulting fragments were assembled by a NEBuilder reaction (NEBuilder HiFi DNA assembly; New England Biolabs) in a pUC19 (Thermo Fisher) backbone, which was PCR amplified with primers MM124/125. For transformation of A. fumigatus, a fragment amplified from plasmid pMMHL69 with primers MM322/331 was used.

For generation of plasmid pMA01 containing the Δsit2 deletion cassette, PCR amplifications with oligonucleotides MA01/02, MA03/04, and MA05/06 were performed to amplify a 5′-homologous region of sit2 (AFU_7G04730), a pyrithiamine resistance cassette (ptrA), and the 3′-homologous region of sit2, respectively. The resulting fragments were assembled by a NEBuilder reaction (NEBuilder HiFi DNA assembly; New England Biolabs) in a pUC19 (Thermo Fisher) backbone linearized with primer MM124/125. For transformation of A. fumigatus, a fragment amplified from plasmid pMA01 with primers MM322/331 was used.

Transformation of A. fumigatus AfS77 was performed according to Tilburn et al. (42). For selection of transformants on minimal medium plates, 0.1 mg/ml hygromycin B (Calbiochem) or 0.1 mg/ml pyrithiamine (Sigma-Aldrich) was used. Plasmid pMMHL69 comprises three sequences that are identical to the A. fumigatus sit1 locus: (i) the sit1 5′-untranslated region (UTR), (ii) the sit1 coding sequence, and (iii) the sit1 3′-UTR (Fig. S1A). Therefore, two different integration modes were possible: homologous recombination via the sit1 5′-UTR and the sit1 coding sequence led to a strain expressing sit1 without a C-terminal Venus tag under the control of the xylP promoter, while homologous recombination via the sit1 5′-UTR and the sit1 3′-UTR led to a strain expressing sit1 with a C-terminal Venus tag under the control of the xylP promoter. Correct genetic manipulations were proven by Southern blotting (Fig. S1C). Primers used for cloning and amplification of hybridization probes are listed in Tables S3 and S4, respectively.

Generation of Saccharomyces cerevisiae strains.

For generation of plasmid pMMHL73, the coding sequence of C. glabrata sit1 was amplified from genomic DNA with primer MM340/341. The resulting fragment was fused to the methionine-repressible MUP1 promoter (43) and a GFP in a pRS415-derived plasmid (44, 45) by NEBuilder fusion. The plasmid backbone was PCR amplified with primers MM336/337.

For generation of plasmid pMMHL75, the coding sequence of A. fumigatus sit1 was amplified from cDNA. The resulting fragment was fused to the methionine-repressible MUP1 promoter and a GFP in a pRS415-derived plasmid (44) by NEBuilder fusion. The plasmid backbone was PCR amplified with primers MM336/337.

Yeast transformation with plasmid DNA was performed by a standard procedure described previously (46), with the following modifications. Yeast was grown in YPD medium (1% yeast extract, 2% peptone, 2% glucose) for 24 h, and 5 OD₆₀₀ (optical density at 600 nm) units of cells were collected by centrifugation at 5,000 × g for 3 min. Cells were resuspended in 0.3 ml of transformation buffer (40% polyethylene glycol 400 in 0.1 M lithium acetate, 1 mM EDTA, and 20 mM Tris). One hundred nanograms of plasmid DNA was added together with 0.015 ml of salmon sperm DNA. The mixture was incubated with rotating for 1 h at room temperature and then was plated on YNB agar plates lacking leucine. Plates were incubated at 26°C for 3 days until visible CFU appeared. Each single CFU was then cultivated in liquid media with the same composition as that of agar plates.

Northern blot analyses.

Total RNA was isolated according to the TRI Reagent (Sigma-Aldrich) method using peqGOLD phase trap reaction tubes (PEQLAB). Formaldehyde-containing agarose gels were used to separate 10 μg of total RNA before being blotted onto Hybond-N+ membranes (Amersham Biosciences) and hybridized with digoxigenin-labeled probes. Primers used for amplification of hybridization probes are listed in Table S4.

Western blot analysis.

For protein extraction, collected mycelia were freeze-dried and homogenized, and 5 mg of mycelium powder was resuspended in 1 ml 0.2 M NaOH–0.2% β-mercaptoethanol. After addition of 0.075 ml trichloroacetic acid (100%, wt/vol), samples were incubated on ice for 10 min and centrifuged at 4°C for 5 min at 14,000 rpm. The pellet was discarded; 0.1 ml of 1 M Tris and 0.1 ml of 2× SDS protein sample buffer (8 M urea, 10% SDS, 12% glycerol, 6% β-mercaptoethanol, 0.004% bromphenol blue, and 0.125 M Tris HCl, pH 6.8) was added to the supernatant. The samples were heated for 5 min at 45°C and vortexed until the pellet was fully dissolved; 0.015-ml aliquots were resolved on a 10% SDS-polyacrylamide gel and transferred onto a polyvinylidene difluoride (PVDF) membrane for Western blot assay using a Trans-Blot turbo transfer system (Bio-Rad). Sit1-Venus was detected using a mouse monoclonal anti-GFP antibody (1:10,000; 11814460001; Roche) as the primary antibody, followed by a goat anti-mouse IgG coupled to horseradish peroxidase (1:10,000; A4416; Sigma-Aldrich) as a secondary antibody. Proteins were detected with enhanced chemiluminescence (ECL; GE Healthcare).

Fluorescence microscopy.

For microscopy, strains were grown in 8-well chamber slides (Ibidi) with 2 × 10⁴ spores/well (final concentration of 10⁵/ml) for 16 h with 0.1% (wt/vol) xylose under iron starvation (−Fe; 0.1 mM BPS) or iron sufficiency (+Fe; 0.03 mM FeSO₄). Mycelia were examined with a spinning-disc confocal microscopic system (Ultra VIEW VoX; PerkinElmer, Waltham, MA) that was connected to a Zeiss AxioObserver Z1 inverted microscope (Zeiss, Oberkochen, Germany). Images were acquired with Volocity software (PerkinElmer) with a 63× oil immersion objective with a numerical aperture of 1.4. The laser wavelength used for excitation of Venus was 488 nm.