Novel Zinc-Attenuating Compounds as Potent Broad-Spectrum Antifungal Agents with In Vitro and In Vivo Efficacy

ABSTRACT An increase in the incidence of rare but hard-to-treat invasive fungal pathogens as well as resistance to the currently available antifungal drugs calls for new broad-spectrum antifungals with a novel mechanism of action. Here we report the identification and characterization of two novel zinc-attenuating compounds, ZAC307 and ZAC989, which exhibit broad-spectrum in vitro antifungal activity and in vivo efficacy in a fungal kidney burden candidiasis model. The compounds were identified serendipitously as part of a drug discovery process aimed at finding novel inhibitors of the fungal plasma membrane proton ATPase Pma1. Based on their structure, we hypothesized that they might act as zinc chelators. Indeed, both fluorescence-based affinity determination and potentiometric assays revealed these compounds, subsequently termed zinc-attenuating compounds (ZACs), to have strong affinity for zinc, and their growth inhibitory effects on Candida albicans and Aspergillus fumigatus could be inactivated by the addition of exogenous zinc to fungal growth media. We determined the ZACs to be fungistatic, with a low propensity for resistance development. Gene expression analysis suggested that the ZACs interfere negatively with the expression of genes encoding the major components of the A. fumigatus zinc uptake system, thus supporting perturbance of zinc homeostasis as the likely mode of action. With demonstrated in vitro and in vivo antifungal activity, low propensity for resistance development, and a novel mode of action, the ZACs represent a promising new class of antifungal compounds, and their advancement in a drug development program is therefore warranted.

T he success of pathogenic microorganisms hinges upon their ability to sequester essential nutrients from their host during infection. Through a process known as nutritional immunity, the host immune system sequesters metals that are necessary for microbial growth, resulting in an extremely nutrient-limited host environment (1). For example, vertebrates express a number of iron-binding molecules, e.g., the transferrin family, that ensure extremely low concentrations of free iron in the body (2). Additionally, neutrophils and other myeloid and nonmyeloid cells synthesize large amounts of the antimicrobial Zn 2ϩ /Mn 2ϩ -chelating protein calprotectin during infection, and the contribution of calprotectin to the innate immune response against yeast and filamentous fungal pathogens is well documented (3)(4)(5).
For fungal pathogens to grow and establish infection inside their hosts, they must be able to obtain iron, zinc, and other essential metals from the harsh environment imposed by nutritional immunity (6). Consequently, successful pathogens have evolved elegant mechanisms to sequester essential metals from their hosts during infection.
The mechanisms for iron sequestration are best described and include the expression of high-affinity iron transporters, iron-chelating siderophores, and iron-binding proteins (1, 7,8). Although iron acquisition is recognized as a virulence factor for many fungal pathogens (7), research in recent years has highlighted the important contribution that zinc sequestration makes to fungal pathogenesis and virulence (4,9). Indeed, fungal acquisition of zinc has been clearly demonstrated to be essential for fungal growth and pathogenicity, and zinc-depleting conditions are known to reduce fungal growth in vitro (3,10,11).
In all fungal species, the major zinc-binding proteins include Cu 2ϩ /Zn 2ϩ superoxide dismutases (SODs), alcohol dehydrogenase, and ribosomal proteins (12). SODs are key enzymes in fungal virulence and are necessary for the detoxification of reactive oxygen species generated by host cells during fungal infection (13). In Aspergillus fumigatus, zinc uptake is regulated by the transcriptional regulator ZafA, and deletion of zafA has been shown to not only impair germination and overall growth capacity of A. fumigatus in zinc-limiting media but also completely abrogate A. fumigatus virulence in a murine model of invasive aspergillosis (11). Thus, the control of access to zinc is one of the central battlefields on which the outcome of an infection is decided. In further support of this notion, calprotectin comprises ϳ40% of total protein content in the neutrophil cytoplasm during infection, and its antifungal effect can be reversed in vitro by micromolar quantities of zinc (3,4,9). Because of the great need for fungal zinc uptake during infection, it has been hypothesized that both chelation therapy and the modulation of zinc homeostasis and zinc acquisition are promising antifungal strategies (14)(15)(16)(17)(18).
We have previously reported the identification of novel antifungal compounds targeting the fungal plasma membrane H ϩ -ATPase (19,20). In the further optimization process a number of compounds were synthesized, and we found two of these compounds, ZAC307 and ZAC989, to be very potent inhibitors of Candida albicans growth, despite the fact that they lacked H ϩ -ATPase-inhibitory activity. Due to their characteristic arrangement of an aromatic structure with nitrogen bound in close proximity to a hydroxyl group, we speculated that ZAC307 and ZAC989 could act as metal chelators. Thus, the goals of this study were (i) to investigate the chelating properties of these compounds, (ii) to characterize the spectrum of antifungal activity of these compounds in vitro, (iii) to ascertain whether the compounds were fungistatic or fungicidal and the propensity of C. albicans to develop resistance against these compounds, (iv) to investigate whether the antifungal activity was caused by extracellular zinc sequestration or if the compounds were taken up by Candida albicans cells, (v) to assess whether these compounds influenced the expression level of genes encoding zinc transporters required for zinc uptake from zinc-limiting media and that of other genes regulated by ZafA, which is the master regulator of zinc homeostasis in Aspergillus fumigatus, and (vi) to test and evaluate the effects of these compounds against mammalian cells and their antifungal efficacy in vivo in a murine model of candidiasis.

RESULTS
ZAC307 and ZAC989 have high binding affinity for zinc and copper but not for magnesium and calcium. ZAC307, ZAC989, ZAC623 (collectively referred to as ZACs), and the reference compounds EDTA and TPEN [N,N,N',N'-tetrakis(2-pyridylmethyl)ethylenediamine] (Fig. 1) were evaluated for their zinc binding properties. ZAC307 and ZAC989 have dissociation constants (K d ) in the low nanomolar range (13 to 71 nM), as determined by a fluorescence-based competition assay (Table 1). ZAC623 exhibited poor affinity for zinc, with a dissociation constant of Ͼ6 M. Dissociation constants for EDTA and TPEN could not be determined with this assay, as they were below the measurable range, but both have previously been reported to be very potent zinc chelators (21).
The Zn 2ϩ -binding properties of ZAC307 and ZAC989 were further evaluated using a potentiometric assay, where pH is measured as a function of base (NaOH) added to the compound in either the absence or presence of metal. Since potentiometric methods require millimolar concentrations, and ZAC307 and ZAC989 displayed poor solubility in water at such high concentrations, the measurements were performed in a mixture of dimethyl sulfoxide (DMSO) and water (70:30, vol/vol), as described  previously (22). To determine the deprotonation constant, a solution of 1 mM ZAC307 or ZAC989 was titrated with 0.3 M NaOH at constant ionic strength ( Fig. 2A and B). In a second run, the same titration was performed in the presence of 0.5 eq of Zn 2ϩ for ZAC307 and ZAC989. A shift in the pH curve in the presence of the metal (ZAC307, 0.5 eq of Zn 2ϩ [ Fig. 2A]) compared to the absence of the metal (ZAC307 [ Fig. 2A]) indicates binding of Zn 2ϩ to the compound. The measured pH data were analyzed with the Hyperquad program suite, taking into account all relevant equilibrium constants, including also the constants for metal hydroxylation. The analysis provides the pK a values and metal complex stability constants, as well as ligand-metal complex speciation calculation, indicating how many ZAC molecules are involved in coordinating the Zn 2ϩ ion at different pH values. Refinement of the measured pH data for ZAC307 provided a pK a value of 6.84 and formation constants log ␤ 1 , log ␤ 2 , and log ␤ 3 of 7.47, 13.27, and 18.14, corresponding to the formation of 1:1, 2:1, and 3:1 ligand-Zn(II) complexes, respectively. The corresponding species distribution diagram is displayed in (C, D, and E) Species distribution diagram as a function of pH for a system containing 0.5 mM Zn(II) and 1 mM ZAC307 (C), 0.5 mM Zn(II) and 1 mM ZAC989 (D), and 0.5 mM Cu(II) and 1 mM ZAC989 (E). Fig. 2C, and it shows that at neutral pH in a DMSO-water solvent mixture, the ligand-zinc stoichiometry is a mixture of 1:1, 2:1, and 3:1 binding, with 2:1 and 3:1 being the dominant stoichiometries. Similar potentiometric experiments were carried out with CaCl 2 and MgCl 2 in place of Zn(NO 3 ) 2 , and these revealed that Ca 2ϩ and Mg 2ϩ binding to ZAC307 is negligible ( Fig. 2A). Potentiometric titration of ZAC307 with CuSO 4 was not possible due to precipitation of the resulting complex, suggesting that ZAC307 also binds copper. Refinement of the measured pH data for ZAC989 provided a pK a value of 7.70, and in the presence of zinc we measured formation constants log ␤ 1 , log ␤ 2 , and log ␤ 3 of 7. 35, 14.30, and 19.71, corresponding to the formation of 1:1, 2:1, and 3:1 ligand-Zn(II) complexes, respectively (Fig. 2D). Potentiometric experiments were also carried out with ZAC989 and CuSO 4 and with CaCl 2 and MgCl 2 , and the data indicated that ZAC989 chelates copper, whereas binding of calcium and magnesium is negligible. ZAC989 bound copper with the formation constants log ␤ 1 , log ␤ 2 , and log ␤ 3 of 7.83, 13.37, and 20.62, respectively. The most dominant ZAC989-copper stoichiometry was a 3:1 stoichiometry (Fig. 2E).
ZACs are potent broad-spectrum fungistatic yeast inhibitors that work intracellularly and display low potential for resistance development. ZAC307 and ZAC989 exhibit antifungal activity and display potent growth inhibition in the low microgram-per-milliliter range (0.2 to 0.9 g/ml) against a number of pathogenic Candida species, including a Candida glabrata strain with increased efflux pump activity ( Table 2). The MIC was defined as the lowest compound concentration that resulted in at least 50% growth inhibition for yeasts, which corresponded to a prominent decrease in visible growth. For molds the MIC was defined as the lowest concentration of the compound that resulted in no visible growth. In Candida albicans this value was 0.6 g/ml for ZAC989 and 0.4 g/ml for ZAC989. ZAC623 did not display growth-inhibitory activity against C. albicans (Fig. 3A). The known potent metal chelators EDTA and TPEN both exhibited a MIC of ϳ0.05 g/ml, but TPEN led to a more complete growth inhibition than did EDTA (Fig. 3B). The antifungal effects of ZAC989 and ZAC307 were reversed by exogenous addition of zinc or copper to the growth medium in the presence of either ZAC989 or ZAC307 ( Fig. 3C and D). Zinc ions were most effective in reversing the growth-inhibitory effects of the ZACs, with restoration of fungal growth observed in the presence of 1 M Zn 2ϩ . Addition of iron (Fe 2ϩ ) had a modest effect on the antifungal activity of ZAC989 and ZAC307, with a fungal growth rate of approximately 50% in the presence of 100 M Fe 2ϩ compared to control cells. In accordance with the results obtained from potentiometric titration, the addition of magnesium or calcium had no effect on the antifungal activity of the ZACs (Fig. 3C and D).
Time-kill investigations revealed that the ZACs exhibited fungistatic activity against C. albicans, in contrast to amphotericin B (AMB), which exhibits fungicidal activity after 3 h of exposure (Fig. 4A). Both EDTA and TPEN exhibited a fungistatic effect within the first 24 h of exposure. Fungal growth recovery was evaluated after longer times of exposure of C. albicans cells to the ZACs. This revealed that the fungal cells were able to resume growth when moved to fresh growth media in the absence of ZACs (Fig. 4B). Fungal cells exposed to TPEN at concentrations above 2 g/ml showed poor recovery, and this may be explained by the strong chelating properties of TPEN that enable it to extract zinc from essential enzymes, leading to fungal cell death after prolonged exposure (Fig. 4C).
a This strain has mutations resulting in increased efflux pump activity compared to that of wild-type isolates (34). b ND, not determined.
In order to gain an understanding of the potential of ZACs to permeabilize into fungal cells, we monitored the intracellular zinc levels of C. albicans cells that were exposed to TPEN, EDTA or ZACs using the cell-permeable fluorescent probe Zinbo-5. The affinity constant of this probe for zinc is 2.2 nM (23), which is weaker than that of most zinc-binding proteins, and thus it reports only the free or weakly bound zinc ions. The probe localizes to the internal membrane system, including the endoplasmic reticulum (ER), in C. albicans (L. Kjellerup, A. L. Winther, D. Wilson, and A. T. Fuglsang, submitted for publication). A decreased Zinbo-5 fluorescent signal in the presence of fungal cells, compound, and zinc would indicate that intracellular Zinbo-5 is competing with the added compound for zinc ions. ZAC307 and ZAC989 were evaluated at a concentration of 25 M, equivalent to 7 g/ml and 9 g/ml, respectively. ZAC989 induced a time-dependent decrease in Zinbo-5 fluorescence, similarly to 1.3 g/ml of TPEN (Fig. 4D). ZAC307 decreased the Zinbo-5 fluorescence after only 1 h of incubation, and this decrease was greater than that observed for ZAC989 (Fig. 4D), despite ZAC307 having a 5-fold-lower affinity for zinc than ZAC989. These data suggested that ZAC307, ZAC989, and TPEN were cell permeative and bound intracellular zinc. In agreement with this, the extracellular chelator EDTA did not reduce the Zinbo-5 fluorescence under the same conditions.
We investigated the propensity for development of resistance to ZAC307 and ZAC989 by repeated exposure of C. albicans to ZACs in SDwoz medium (see Materials and Methods) over a 36-day period. We observed no change in the MIC for the ZACs after repeated ZAC exposure. In contrast, cells repeatedly exposed to fluconazole exhibited a significant increase in the MIC for fluconazole after 22 passages (Fig. 4E). Based on these results, it appeared that ZAC resistance was not easily induced in C. albicans.
ZACs efficiently inhibit the growth capacity of Aspergillus fumigatus under zinc-limiting conditions, and their inhibitory effects are inactivated by zinc. In albicans revealed that ZACs exhibit fungistatic activity. The following final concentrations of compound were applied: 0.5 g/ml of amphotericin B (AMB), 3.6 g/ml of ZAC989, 4.4 g/ml of EDTA, and 4.2 g/ml of TPEN. Data are means Ϯ SEMs for two biological replicates. (B and C) C. albicans cells exposed to ZAC989 and ZAC307 for 5 days resume normal growth (B), while cells exposed to TPEN for 5 days show poor recovery (C). C. albicans cells exposed to EDTA resumed visible growth after 48 h of compound incubation (concentration range, 0.22 g/ml to 22 g/ml) and were therefore not evaluated for MFC. (D) A decrease in intracellular zinc as evidenced by a decrease in Zinbo-5 fluorescence signal suggests that ZAC989 and ZAC307 act intracellularly in C. albicans. The following final concentrations of compound were applied: 9 g/ml of ZAC989, 7 g/ml of ZAC307, 15 g/ml of EDTA, and 1.3 g/ml of TPEN. Data are means Ϯ SEMs for two biological replicates. (E) Resistance induction study. There was no change in the MIC for the ZACs after repeated ZAC exposure, while a significant increase in the MIC was observed for C. albicans after 22 passages with repeated exposure to fluconazole (FLC) (1.0 g/ml).
addition to potent antifungal effects on the five Candida species tested, ZAC307 and ZAC989 also potently inhibited the mold Aspergillus fumigatus and other Aspergillus species, as well a number of rare but very hard-to-treat members of the Mucorales order, including Rhizopus oryzae, Rhizopus microsporus, and Mucor indicus. The ZACs inhibited these molds and mucorales isolates in a range from 0.4 g/ml to 5.4 g/ml (Table 3). To assess the capacity of the ZACs to inhibit A. fumigatus growth in the presence of zinc, 1-ml aliquots of sRPMI zinc-limiting medium (RPMI 1640 medium [R8755; Sigma] supplemented with 10 M FeSO 4 ·7H 2 O, 1 M CuSO 4 ·5H 2 O, and 1 M MnCl 2 ·H 2 O) or this medium supplemented with 2, 5, or 50 M zinc were inoculated with 10 5 conidia of a wild-type A. fumigatus strain (AF14), dispensed in 24-well culture plates, and incubated in the presence of either ZAC307 or ZAC989 at a final concentration 21 g/ml or 27 g/ml (equivalent to 75 M), respectively (Fig. 5A). Graphical representation and quantification of the fungal growth in 24-well culture plates in the presence of ZACs (Fig. 5A) revealed that the growth capacity of a wild-type A. fumigatus strain was reduced under zinc-limiting conditions but increased gradually when the growth medium was supplemented with increasing amounts of zinc, until fungal growth was fully restored when cultured in medium supplemented with 50 M Zn (i.e., under zinc-replete conditions). Hence, the inhibitory effects of the ZACs against A. fumigatus were completely counteracted by simultaneous addition of zinc, similar to our observations for C. albicans ( Fig. 3C and D).
The zinc transporter ZrfC plays an important role in regulating fungal sensitivity to the ZACs. To ascertain whether ZAC307 and ZAC989 interfered with zinc uptake from the sRPMI zinc-limiting medium, we analyzed their effects on the growth capacity of the mutant strains AF48 (ΔzrfA ΔzrfB), AF721 (ΔzrfA ΔzrfB ΔzrfC), AF731 (ΔzrfA ΔzrfB ΔzrfC [zrfC]), which is an AF721 derivative strain that carries the zrfC gene reintroduced at the pyrG locus as described previously (24), and AF54 (ΔzrfC) (Fig. 5B to E). The overall effect of the ZACs on the growth capacity of the AF48 strain was similar to that of the wild-type strain AF14 (compare Fig. 5A and B). In the absence of ZACs, both strains exhibited a reduced growth capacity, from 100% to 70%, when cultured in media supplemented with 50 M compared to media supplemented with 2 M zinc. This corresponded to a 1.4-fold reduction in growth capacity. However, in the presence of 2 M zinc plus 75 M ZAC307, the growth capacity of AF14 and AF48 was reduced, respectively, 2.9-and 3.9-fold compared to that in the presence of 50 M zinc. Similarly, 2 M zinc plus 75 M ZAC989 reduced the growth capacity of AF14 and AF48 3.9-and 4.5-fold, respectively. The AF721 strain did not grow under zinc-limiting conditions, and hence the effect of these compounds on this strain could not be tested (Fig. 5C). The reintroduction of zrfC in a strain with a ΔzrfA ΔzrfB ΔzrfC genetic background restored the fungal growth capacity in the presence of the ZACs (Fig. 5D) at the same level as that of the wild-type and AF48 strains ( Fig. 5A and B). Finally, in the presence of 2 M zinc either with or without simultaneous exposure to ZAC307 or ZAC989, the growth capacity of the AF54 strain was reduced an average of 24.6-fold compared to its growth capacity in the presence of 50 M zinc (Fig. 5E); i.e., the ZACs were between 5-and 8-fold more efficient as inhibitors of the growth capacity of a ΔzrfC strain than of a wild-type or ΔzrfA ΔzrfB strain, which suggested that the effect of ZACs could be counteracted to a certain extent by the function of ZrfC.
ZAC307 and ZAC989 inhibit the transcription of genes regulated by ZafA under zinc-limiting conditions. The major regulator of the A. fumigatus zinc homeostatic response under zinc-limiting conditions is the transcription factor ZafA (11), which is a  zinc-responsive factor that senses the intracellular concentration of zinc in a way similar to that of its orthologue Zap1 in the yeast Saccharomyces cerevisiae (25). Thus, when the cytoplasmic zinc content is high enough, ZafA becomes saturated with Zn 2ϩ ions and adopts a transcriptionally inactive conformation. In contrast, when the intracellular concentration drops below a certain threshold, ZafA begins to release Zn 2ϩ ions and gradually adopts a transcriptionally active conformation, whereby it is able to induce the expression of zrfA and zrfB in acidic zinc-limiting media and of zrfC in alkaline zinc-limiting media (11,24). We employed quantitative reverse transcription-PCR (RT-qPCR) to assess whether exposure to ZAC307, ZAC989, EDTA, or TPEN influenced the expression of several ZafA target genes and other genes not regulated directly by ZafA (as controls) ( Table 4). The set of genes investigated was selected based on a genome-wide transcription analysis of A. fumigatus grown under zinc-limiting conditions that had been performed previously in our laboratory (J. A. Calera, unpublished data), using the primers listed in Table  5. All selected ZafA target genes were induced by ZafA under zinc-limiting conditions with the exception of the putative zinc storage vacuole transporter zrcA, which was repressed by ZafA under zinc-limiting conditions (Calera, unpublished). As expected, the relative expression levels of all ZafA target genes induced by ZafA under zinclimiting conditions were dramatically reduced, to almost undetectable levels, upon the addition of Zn 2ϩ (Fig. 6). In contrast, the expression level of the zrcA gene, which was repressed by ZafA, and most of the genes not regulated by ZafA increased under zinc-replete conditions to different extents, with the exception of actA, whose expression level remained similar to that observed before the zinc shock. Interestingly, exposure to either ZAC307 or ZAC989 inhibited the expression of the ZafA target genes similarly. We also observed reduced expression levels for most of the investigated genes that were not regulated by ZafA following ZAC307 exposure. In particular, the expression level of pmaA, which encodes the orthologue of the Pma1 H ϩ -ATPase from S. cerevisiae, and that of the gdpA and tubB1 genes were reduced at levels similar to that of the ZafA target genes (Fig. 6). In contrast, treatment with ZAC989 did not have a noticeable effect on the expression levels of the genes not regulated by ZafA, which remained similar to that observed under zinc-limiting conditions (Fig. 6). This finding suggested that although the overall outcomes of the treatment with ZAC307 and ZAC989 on ZafA-regulated genes were quite similar, the precise mode of action of  ZAC307 on gene expression was different from that of ZAC989, which appeared to inhibit the ZafA regulated genes more specifically than ZAC307. Finally, we anticipated that chelation of extracellular zinc upon addition of a relatively high concentration of EDTA to the culture media for a short period (2 h) would result in a transient hyperactivation of ZafA and concomitant upregulation of the most direct ZafA target genes, including those encoding zinc transporters. We expected to observe the same effect with TPEN treatment, although in this case chelation of intracellular zinc should exacerbate the zinc starvation status of the fungal cells, leading to a more extended hyperactivation of ZafA and higher expression of the ZafA target genes than attained with EDTA after the same incubation period. The expression profile for the ZafA-regulated genes observed in EDTA-or TPEN-treated cultures reflected precisely what we predicted (Fig. 6).
In summary, these results suggested that the antifungal effects of ZAC307 and ZAC989 were most likely mediated through a mechanism that ultimately results in the inhibition of the transcriptional activation activity of ZafA.
Cytotoxicity and off-target activity studies. ZAC307, ZAC989, EDTA, and TPEN were evaluated for mammalian cytotoxicity in a standard hepatocyte proliferation assay, where the mammalian HepG2 cell line was exposed to the compounds for either 24 h or 72 h. After 24 h of exposure to ZAC307 and ZAC989, the half-maximal effective concentration (EC 50 ) was Ͼ28 g/ml, while after 72 h of exposure, the EC 50 s were 13.2 g/ml and 6.9 g/ml, respectively (Table 6). With antifungal activity against yeast and the Mucorales isolates in the 0.2-to 1.7-g/ml range, the compounds exhibited a reasonable selectivity index toward mammalian cells. However, the selectivity index between the Aspergillus species and mammalian cells was limited. The nonpermeative chelator EDTA did not affect the proliferation of HepG2 cells, while the potent zinc chelator TPEN had an EC 50 of 1.6 g/ml after 24 h of exposure.
ZACs exhibited in vivo efficacy in a murine fungal kidney burden candidiasis model. ZAC989 and ZAC307 were investigated for in vivo efficacy in a fungal kidney burden model (Fig. 7; Table 7). In this model, BALB/c mice were infected intraperito- neally (i.p.) on day 0. Initially, administration of ZACs included a pretreatment 24 h prior (day Ϫ1) to infection (day 0) by the i.p. route. The mice were then treated with either ZAC for 4 days (day Ϫ1 to day 2) to maximize the likelihood of observing in vivo efficacy. The endpoint was mean log CFU in the kidneys of treated animals compared to the value for untreated animals. Fluconazole was chosen as a comparator compound and dosed per os (p.o.), and treatment with fluconazole resulted in a significant reduction in kidney burden, 2.78 log CFU/kidney (Fig. 7). The in vivo studies also revealed that i.p. dosing of ZAC989 at 60 mg/kg of body weight resulted in a statistically significant reduction, 1.71 log CFU/kidney, while ZAC307 administration at 60 mg/kg led to a significant reduction, 1.06 log CFU/kidney. ZAC307 yielded equal in vivo efficacies with and without pretreatment ( Fig. 7; Table 7). We observed no adverse effects following dosing of 60 mg/kg of ZAC307, but for ZAC989, we observed lethargy lasting for 5 to 15 min after dosing.

DISCUSSION
During this study, we identified a new series of zinc-attenuating compounds with broad-spectrum antifungal activity in vitro and in vivo activity in a candidiasis fungal kidney burden model. The compounds ZAC307 and ZAC989 possess a characteristic arrangement of an aromatic structure with nitrogen bound in close proximity to a hydroxyl group. This structural arrangement led us to speculate whether ZAC307 and ZAC989 were metal chelators. To address this hypothesis, we synthesized ZAC623 as a  control compound where the hydroxyl group is replaced with an amino group, and as expected, this compound lacked metal-chelating and antifungal properties ( Fig. 1 and  3A; Table 1). The metal-chelating compounds EDTA and TPEN have previously been described as antifungal compounds (16,17,26); thus, they were selected as comparators in this study.
Our findings indicate that ZAC307 and ZAC989 chelate both zinc and copper, with zinc ions being most effective in reversing the growth-inhibitory effects of the ZACs (Fig. 3C and D). The ZACs have lower affinity for iron and negligible affinity for magnesium and calcium. The ZACs are less potent zinc chelators than the known zinc, copper, and iron chelators TPEN and EDTA, but they inhibit fungal growth more effectively than EDTA (Fig. 3A and B). The inhibitory effects of the ZACs can be inactivated by the addition of excess zinc in both Candida albicans and Aspergillus fumigatus, which indicates that these compounds interfere either directly or indirectly with fungal zinc homeostasis.
The first challenge that any microorganism faces in the homeostatic response to zinc deficiency is to obtain zinc from the surrounding environment. The major components of the zinc uptake system in A. fumigatus that facilitates zinc uptake from zinc-limiting media are the ZIP plasma membrane transporters ZrfA, ZrfB, and ZrfC. The ZrfA and ZrfB transporters operate mainly under acidic zinc-limiting conditions (27), although they also contribute to zinc uptake from alkaline zinc-limiting media along with ZrfC (3), which is expressed exclusively in alkaline media (24). Therefore, we reasoned that if the ZACs inhibited the intake of zinc mediated by these transporters, the growth capacity of a wild-type strain in the presence of ZACs should be reduced to the same level as that of fungal mutant strains lacking the acidic (ZrfA and ZrfB) and/or the alkaline (ZrfC) zinc transporter. The investigation of the growth capacity of a ΔzrfC mutant in the presence of the ZACs suggested that ZrfC plays an important role in overcoming ZAC inhibition, since deletion of zrfC increased the sensitivity of A. fumigatus to these compounds (AF14 versus AF54 [ Fig. 5A and E]). Hence, it could be possible that the  actual effect of ZACs on the ΔzrfA ΔzrfB strain was masked to some extent by the ZrfC function. As expected, the AF721 strain lacking zrfC did not grow under zinc-limiting conditions, and the effects of these compounds on this strain could therefore not be tested (Fig. 5C). However, we observed that ZAC989 noticeably inhibited the growth capacity of AF721 in the presence of 50 M zinc. Thus, it is plausible that ZAC989 interfered with a zinc homeostatic process other than zinc uptake. Furthermore, the higher growth capacity of the ΔzrfC strain in the absence of ZACs than of the wild-type, AF48, or AF731 strain suggested that the lack of zrfC in AF54 may have been compensated for by the overexpression of zrfA and zrfB, as reported previously (3). On the other hand, the stronger growth inhibition of the growth capacity of the AF54 strain in the presence of ZAC307 and ZAC989 than of the wild-type, AF48, or AF731 strain suggested that the levels of expression of the zrfA and zrfB genes in AF54 were insufficient to counteract the effects of ZAC307 and ZAC989. Taken together, these results suggest that both ZAC307 and ZAC989 interfered negatively with the expression of the genes encoding zinc transporters rather than with their zinc uptake function. Indeed, gene expression analysis by RT-qPCR suggested that the antifungal effects of ZAC307 and ZAC989 were mediated through a mechanism that ultimately results in inhibition of the transcriptional activation activity of ZafA. In addition, and since ZafA activity was inactivated by zinc under physiological conditions, the effects of ZAC307 and ZAC989 on the expression of ZafA-regulated genes could be exerted either directly, upon their binding to ZafA, or indirectly by increasing the cytosolic concentration of Zn 2ϩ ions that would bind to and inactivate ZafA. Although direct binding of ZAC307 or ZAC989 to ZafA was an attractive possibility, we consider it more likely that the ZACs triggered a transient rise in the cytosolic concentration of Zn 2ϩ ions by favoring their releasing from cytosolic zinc ligands and/or zinc storage compartments (e.g., the vacuole). For the data presented in Fig. 4D, we observed that the ZACs promote a decrease in the available zinc levels within the internal membrane system, including the ER, in C. albicans. This decrease can be explained either by a direct competition between Zinbo-5 and ZACs within these compartments or by promotion by the ZACs of the release of zinc from the ER into the cytosol. The latter scenario could induce a higher concentration of zinc in the cytosol, and if the same situation occurs in A. fumigatus, this could explain the decrease in transcription of the ZafA-induced genes, similar to when zinc is added extracellularly. Moreover, since most intracellular eukaryotic zinc proteins bind zinc with a higher affinity than ZACs (K d values for most zinc-binding proteins are between 0.1 and 1.0 pM) (28), it is more likely that these compounds interfere with the regulation of zinc homeostasis by promoting the release of Zn 2ϩ ions from intracellular storage compartments. Nevertheless, since acidification of cytoplasm can also promote zinc release from cytoplasmic zinc ligands (29), it could also be possible that the ZACs disturbed the activity of proteins involved in maintaining the pH homeostasis of the cytoplasm. In this regard, we observed that the expression level of pmaA was higher under zinc-replete than under zinc-limiting conditions. This finding suggests a putative interplay between the function of PmaA and the Zn 2ϩ transport capacity of the fungal zinc transporters, provided that the expression level of the pmaA gene correlated with the H ϩ -ATPase (PmaA) activity. It is known that the accumulation of zinc into the vacuole by CDF transporters (e.g., ZrcA) is mediated through a Zn 2ϩ /H ϩ antiporter mechanism and relies on the proton gradient generated by the V-ATPase (30). In contrast, and although it is not completely known how ZIP proteins transport zinc across the plasma membrane, it seems that it is not dependent on the proton gradient generated by the plasma membrane H ϩ -ATPase but is pH dependent, such that intracellular acidification increases zinc transport, whereas extracellular acidification decreases zinc transport (31)(32)(33). In this regard, it is plausible that under the alkaline zinc-limiting conditions provided by sRPMI medium, a low PmaA activity would increase intracellular acidification to favor zinc uptake by ZIP transporters located in the plasma membrane. In contrast, a high PmaA activity under zinc-replete conditions would reduce intracellular acidification and increase extracellular acidification (e.g., to counteract the Zn 2ϩ -induced dissipation of the electrochemical gradient that is essen-tial for fungal survival). The same reasoning could be applied to zinc transport by the ZrfF ZIP transporter located in the vacuolar membrane, such that the intravacuolar pH, which is kept lower than the cytosolic pH under normal conditions via V-ATPase activity, would favor the exit of Zn 2ϩ ions into the cytosol. However, the unexpected finding that ZAC307 reduces pmaA expression suggested that it might interfere with the putative mechanism that links the regulation of zinc homeostasis with the function of PmaA. The ZACs display broad-spectrum fungistatic activity and exhibit a low propensity for acquired resistance development compared to that of fluconazole. Additionally, they are superior antifungal agents to the nonpermeative chelator EDTA, and our RT-qPCR data suggest that the ZACs affect fungal zinc homeostasis differently from the very potent chelator TPEN. Therefore, the ZACs act distinctly from either EDTA or TPEN, both of which have previously been investigated as antifungal agents. EDTA has been evaluated as a combination treatment together with amphotericin B lipid complex (ABLC) in an invasive pulmonary aspergillosis model in immunosuppressed rats. The combination of EDTA with ABLC led to improved survival times and a lower tissue burden of A. fumigatus than with either agent alone (26). Furthermore, TPEN has been shown to significantly increase survival after 7 days compared to that with vehicle treatment in a murine model of invasive pulmonary aspergillosis (16). Moreover, administration of either of the two zinc-chelating agents phenanthroline and TPEN has been shown to lead to significant improvements in survival with concomitant reduction in fungal burden in immunosuppressed mice intranasally infected with A. fumigatus. Finally, it was shown that TPEN given in combination with caspofungin significantly increased survival times in murine models of invasive aspergillosis compared to that with either drug alone (17).
There is an unmet need for novel antifungal agents with broad-spectrum antifungal activity and low potential for resistance development for the treatment of invasive fungal infections. With potent antifungal activity both in vitro and in vivo, the ZACs fulfill these criteria, and their advancement in a drug development program is therefore warranted. Interestingly, the ZACs and EDTA did not inhibit mammalian cell proliferation considerably within the first 24 h of exposure (EC 50 Ͼ 28 g/ml), in contrast to TPEN (EC 50 ϭ 1.6 g/ml), and even after 72 h of ZAC exposure, we observed a Ͼ12-fold selectivity index in growth inhibition between C. albicans and HepG2 cells; the index between A. fumigatus and HepG2 cells was limited. However, it should be taken into account that metal ions are crucial in every cellular system, including the host. Therefore, any intervention aiming to treat an infection through ion sequestration must deal with the delicate balance between positive and negative effects in both the pathogen and the host. The therapeutic safety window as well as whether the ZACs can induce zinc deficiency in the host still needs to be addressed. In summary, interfering with fungal zinc-dependent processes represents a promising new approach to antifungal therapy, and this series of zinc-attenuating compounds represents a potentially new class of antifungal agents. specified compound dissolved in 80% ethanol), and 10 5 conidia was dispensed to achieve final concentrations of 0.7ϫ, 0.05% (vol/vol), 1.2% (vol/vol), and 10 5 conidia/ml, respectively. Plates were incubated for 44 h at 37°C in a humid atmosphere. To quantitate mycelial growth, each plate was scanned in an Agfa SnapScan 1236s scanner, and the intensity of the wells was quantified using the open-source image processing program Image J2. The data were represented and analyzed with Prism software 7.0.
Resistance study. The propensity for resistance development was investigated as also described previously (41), but with the following modifications. C. albicans was repeatedly exposed to either ZAC989 or ZAC307 in 1-ml cultures in SDwoz medium with a starting inoculum that had an optical density at 600 nm (OD 600 ) of 0.007. A compound concentration that resulted in ϳ90% growth inhibition was selected for these experiments (3.6 g/ml for ZAC989, 2.8 g/ml and 5.6 g/ml for ZAC307, and 0.5 g/ml and 1.0 g/ml for fluconazole). Over a 36-day period, culture aliquots of 100 l were periodically (every 1 to 2 days) transferred (passaged) to new culture tubes with 900 l of fresh medium and fresh compound. The cells were incubated at 30°C with gentle agitation (150 rpm), and the OD 600 of cultures was monitored throughout the period to ensure that the numbers of cells exposed to compounds were comparable across treatments for each passage. Cells were passaged a total of 22 times, and cultures were periodically tested for antifungal susceptibility by following the protocol for antifungal susceptibility testing, as described above.
Zinbo-5 assay. C. albicans BWP17 was grown overnight in YPD medium at 30°C and 150 rpm. The cells were pelleted, washed in PBS buffer (D8537; Sigma) 3 times, and resuspended to an OD 600 of 2.0. Two microliters of compound in DMSO was mixed with 100 l of cell suspension and incubated statically for 1, 8, or 24 h at 30°C in a 96-well black plate. Thirty minutes before the end of the incubation period, 100 l of 10 M Zinbo-5 (sc-222425; Santa Cruz Biotechnology) in PBS buffer was added. The affinity constant of this probe for zinc is 2.2 nM (23). The plate was then read on a plate reader (FLUOstar Optima; BMG Lab Technologies) with excitation at 355 nm and emission at 485 nm. The decrease in Zinbo-5 fluorescence was calculated relative to the untreated DMSO control.
RNA isolation from Aspergillus fumigatus. A total of 1.5 ϫ 10 6 conidia of the wild-type strain (AF14) were inoculated into 20 ml of 0.7ϫ sRPMI, 0.05% Tween 20, and 1.2% ethanol dispensed into 100-ml culture flasks pretreated with an overnight wash in 2 mM EDTA (pH 8.0) to minimize the presence of metal traces, followed by a thorough washing with ultrapure Milli-Q water. The flasks were subsequently sterilized in an oven at 180°C. The cultures were incubated for 20 h at 37°C and 200 rpm before the following were added: (i) pure ethanol to a final concentration of 1.2% (vol/vol) (as a reference for the transcription profiles of all genes under zinc-limiting conditions), (ii) 1.2% ethanol plus a solution of ZnSO 4 to a final concentration 20 M zinc (as a reference for the transcription profiles of all genes under zinc-replete conditions following the zinc shift), (iii) 1.2% ethanol plus a solution of EDTA to a final concentration of 500 M, (iv) 1.2% ethanol plus a solution of TPEN to a final concentration of 10.6 g/ml, and (v) a 5 mM solution of each test compound (ZAC307 or ZAC989) dissolved in 80% ethanol to a final concentration of 75 M and 1.2% ethanol (which correspond to 21 g/ml and 27 g/ml, respectively). After compound addition, cultures were incubated for 2 h at 37°C and 200 rpm and mycelia were harvested by filtration through filter paper, washed twice with sterile waterm and snap-frozen in liquid nitrogen. After grinding of the mycelia in the presence of liquid nitrogen, total RNA was extracted using an RNeasy plant minikit (74904; Qiagen) according to the manufacturer's instructions. RNA was eluted in 50 l of RNase-free water. RNA integrity was verified on 0.8% agarose gels stained with ethidium bromide. RNA was stored at Ϫ80°C until use.
RT-qPCR. Total RNA concentration and quality were determined by UV spectrometry (NanoDrop ND1000 spectrophotometer; Thermo Fisher Scientific), and all samples were brought to a final concentration of 150 ng/l. Total RNA (1.5 g) was treated with RQ1 DNase I (M610; Promega) and subsequently assessed by conventional PCR for the complete absence of genomic DNA (gDNA). Subsequently, 1 g of DNase-treated RNA was reversed transcribed using SuperScript II reverse transcriptase (18064-014; Invitrogen, Thermo Fisher Scientific), with random hexamers (11034731001; Roche Diagnostics) as primers. Prior to qPCRs, cDNA samples were diluted 1:3 in water, except in the case of reactions against the 18S rRNA, for which samples were diluted 1:1,200 in water. qPCRs were performed on a Bio-Rad CFX96 instrument. A typical qPCR mixture (10 l) contained 13.5 ng of cDNA (32 pg when the qPCR was for 18S rRNA), a specific pair of primers (150 nM final concentration), and the SYBR premix Ex Taq (RR420A; TaKaRa). Primers used for qPCR are listed in Table 5. For all qPCRs, 40 cycles were performed using the following cycling conditions: denaturation at 95°C for 10 s, annealing at 59°C for 20 s, and extension at 72°C for 20 s. The relative expression ratio (rER) was calculated using the threshold cycle (2 ϪΔΔCT ) method (42) using the expression level of the 18S rRNA as an internal reference.
In vivo fungal kidney burden candidiasis model. A murine model of systemic candidiasis was established according to a previously described method (44). BALB/c mice were infected with a 0.1-ml inoculum (1 ϫ 10 5 to 5 ϫ 10 5 CFU) of Candida albicans SC5314 cells by the intravenous route on day 0. Compounds were tested at doses of 60 mg/kg. Administration of compound was initiated 24 h prior (day Ϫ1) to infection (day 0) by the intraperitoneal route, dosing twice a day for 4 days (day Ϫ1 to day 2). ZAC307 was also evaluated with administration of the compounds given after the infection at day 0 (no pretreatment) and following dosing twice a day for 3 days (days 0 to 2). Fluconazole was used as a comparator drug. Six mice were used for each group, and the untreated control group was exposed to the vehicle alone. ZAC989 and ZAC307 were formulated by taking 60 mg of compound and adding this to 2 ml and 4 ml of 0.1 N NaOH, respectively. After mixing and sonication of the resulting solutions, 4 ml of purified water was added. The pH was adjusted to 9.0 with 0.1 N HCl solution, followed by the addition of 90 mg of NaCl. The solutions were then diluted to 10 ml and filtered through a 0.22-m polyvinylidene difluoride (PVDF) filter. Sample collection and processing were performed as follows. Twelve hours after the last dose, all treated and untreated animals were sacrificed by cervical dislocation and kidneys were collected in 3 ml of sterile normal saline. The samples were homogenized, serially diluted, and plated on Sabouraud dextrose agar (SDA). SDA plates were incubated for 24 to 48 h at 35°C, and CFU were enumerated and reported as log CFU/kidney. The endpoint was mean log CFU of fungi in kidneys of treated animals compared to that of untreated animals. The study was conducted in conformance with an application submitted to the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA), New Delhi, India, after approval from the Institutional Animal Ethics Committee (IAEC).