Biochemical analysis of CaurSOD4, a potential therapeutic target for the emerging fungal pathogen Candida auris

Abstract

Candida auris is an emerging multidrug-resistant fungal pathogen. With high mortality rates, there is an urgent need for new antifungals to combat C. auris. Possible anti-fungal targets includes Cu-only superoxide dismutases (SODs), extracellular SODs that are unique to fungi and effectively combat the superoxide burst of host immunity. Cu-only SODs are essential for virulence of diverse fungal pathogens, however little was understood about these enzymes in C. auris. We show here that C. auris secretes an enzymatically active Cu-only SOD (CaurSOD4) when cells are starved for Fe, a condition mimicking host environments. Although predicted to attach to cell walls, CaurSOD4 is detected as a soluble extracellular enzyme and can act at a distance to remove superoxide. CaurSOD4 selectively binds Cu and not Zn, and Cu binding is labile compared to bimetallic Cu/Zn SODs. Moreover, CaurSOD4 is susceptible to inhibition by various metal binding drugs that are without effect on mammalian Cu/Zn SODs. Our studies highlight CaurSOD4 as a potential antifungal target worthy of consideration.

Graphical Abstract

The Cu-only superoxide dismutase SOD4 of the emerging fungal pathogen and superbug Candida auris is susceptible to inhibition by metal binding drugs and represents a promising target for new anti-fungals.

The prevalence of fungal infections is increasing world-wide, concurrent with the emergence of several drug resistant pathogens. One recently emerged multidrug resistant fungus is Candida auris. Clinical isolates of C. auris can be resistant to all three classes of major antifungals and morbidity can be as high as 60% among immunocompromised individuals ^1–5. With the recent rise in C. auris infection outbreaks in health care facilities ⁶, there is an urgent need for new antifungals for this and other drug-resistant fungi of public health importance.

As part of innate immunity, fungi and other pathogenic microbes are subject to chemical attacks by host phagocytes, including bursts of superoxide and other reactive oxygen species (ROS) designed to kill microbes through oxidative damage ⁷. Successful pathogens evade this attack through antioxidant defenses, such as superoxide dismutase (SOD) enzymes that use Cu, Fe or Mn to catalyze the disproportionation of superoxide ⁸. Superoxide does not cross biological membranes unless protonated, e.g., in the low pH of host cell phagolysosomes, and the superoxide generated by host cells can be removed by microbial extracellular SODs. Across numerous bacterial and fungal pathogens, extracellular SODs are essential for pathogenesis ⁸ and are believed to prevent superoxide damage to certain cell surface or extracellular targets, or block cell entry of protonated superoxide ^{9, 10}. Intracellular SODs are also important virulence factors for numerous pathogenic microbes ⁸. Here we focus on the extracellular SODs of the fungal kingdom, the family of so-called Cu-only SODs.

Fungal Cu-only SODs are similar to the Cu/Zn family, but lack a Zn binding site and also the loop VII structure or so-called electrostatic loop that covers the active site. As a consequence of no loop VII, the active site of fungal Cu-only SODs is open, compared to the closed Cu site of Cu/Zn bimetallic SODs ^11–14. Mycobacteria also express a Cu-only SOD, yet this SOD retains loop VII and a closed active site ¹⁵. It is possible that with an open active site, the fungal Cu-only SODs can react with substrates other than superoxide, but to date, none have been identified. Fungal Cu-only SODs are all extracellular, with the vast majority containing GPI anchorage sites for cell surface attachment ^{12, 13}. Cu-only SODs are necessary for pathogenesis of diverse fungi, across pathogens that infect plants, mammals and insects ^{14, 16–21}. Most importantly, fungal Cu-only SODs are structurally distinct from the bimetallic Cu/Zn SOD counterparts in the plant and animal hosts, making Cu-only SODs promising new therapeutic targets for fungal infections ¹⁴.

The bulk of what is currently understood about the biochemistry of Cu-only SODs has emerged from studies with the opportunistic fungal pathogen Candida albicans. C. albicans has two catalytically active Cu-only SOD4 and SOD5 enzymes, and a larger SOD6 whose function is not known ^{11, 12, 16, 22, 23}. SOD4 and SOD5 both localize to the cell wall through GPI anchors but are expressed under distinct conditions ²³. SOD5 is only produced during C. albicans morphogenesis to a hyphal filamentous state while SOD4 is only expressed under Fe starvation conditions ^{16, 23, 24}. The link between Cu-only SODs and Fe starvation is common among several pathogenic yeasts and may represent a stress response to the Fe-limited environment of the animal host ²³. C. albicans SOD4 and SOD5 lack both the Zn site and the electrostatic loop structure of bimetallic Cu/Zn SODs, and as a consequence, Cu binding to C. albicans SOD4 and SOD5 is less stable and of lower affinity compared to that of Cu/Zn SODs ^{14, 23}. With this relatively poor Cu binding, C. albicans SOD5 is susceptible to inhibition by an array of metal binding FDA approved drugs that have no impact on mammalian Cu/Zn SODs ¹⁴.

The emerging fungal pathogen C. auris has two Cu-only SOD-like genes: one similar in size to C. albicans SOD4 and SOD5 (denoted as C. auris SOD4 or herein as CaurSOD4), and a second larger one akin to C. albicans SOD6. We have previously noted that C. auris SOD4 mRNA is induced under Fe starvation conditions, as is the case with other Candida sp ²³. Moreover, heterologous expression of a recombinant CaurSOD4 in C. albicans shows CaurSOD4 is an active SOD enzyme ²³. However, expression of native SOD4 from live C. auris cultures has not been documented. Additionally, it was not known if the biochemical properties of Cu binding seen with C. albicans Cu-only SODs could be extrapolated to CaurSOD4. To date, only C. albicans SOD4 and SOD5 are known to possess a labile Cu site vulnerable to inhibition by metal binding drugs. Whether these properties can extend to Cu-only SODs of other fungi was not known.

Here we investigate the biochemical and cell biology properties of CaurSOD4. We show that live cultures of C. auris indeed produce enzymatically active CaurSOD4 during Fe starvation conditions, yet localization is not restricted to the cell wall as expected, and the protein is directly secreted in the extracellular milieu. CaurSOD4 binds Cu with high specificity, and Cu binding affinity is similar to that previously reported for C. albicans Cu-only SODs. Most importantly, CaurSOD4 is highly susceptible to inhibition by metal binding drugs that are without effect on mammalian Cu/Zn SOD. These findings are discussed in the context of Cu-only SODs as promising drug targets for treating multi-drug resistant pathogens such as C. auris.

Results and Discussion

The extracellular CaurSOD4 in cultures of C. auris

Previously we showed that the mRNA for SOD4 from C. auris is induced under Fe-starvation conditions, similar to SOD4 from C. albicans ²³. Here we sought to determine if C. auris produces an enzymatically active CaurSOD4 polypeptide under these Fe starvation conditions. In the experiment of Fig. 1A, a clinical isolate of C. auris was grown in the presence of 150 μM of the Fe chelator bathophenanthroline disulfonate (BPS), the same Fe starvation conditions known to induce CaurSOD4 mRNA ²³.

Fig. 1: Secretion of Candida auris SOD4 from fungal cultures.

A - C) The growth media from cultures of C. auris (A), C. albicans expressing recombinant CaurSOD4 (B),or the designated strains of C. albicans (C) was analyzed for Cu-only SODs by immunoblot (A top, B and C) or SOD activity (A bottom). Where indicated (+), cultures contained 150 μM of the Fe chelator BPS to induce expression of Fe-regulated CaurSOD4 and CalbSOD4. For immunoblots, samples were deglycosylated with PNGase F (A, C) or Endo H (B) and probed with either anti-C. albicans SOD5 to detect CaurSOD4 (A top, B) or with anti-C. albicans SOD4 to detect CalbSOD4 (C). Molecular weight markers are indicated on left. For SOD activity (A bottom), samples not deglycosylated were subjected to native gel electrophoresis and NBT staining. Strains utilized: (A) C. auris isolate AR0388; (B) C. albicans strain KC2 secreting recombinant CaurSOD4 lacking a GPI anchor site; (C) WT, C. albicans SC5314; sod4Δ/Δ, mutant SS101 lacking CalbSOD4; sod4/5Δ, mutant SS102 lacking both CalbSOD4 and CalbSOD5. (D) CaurSOD4 is aligned against C. albicans SOD4 and SOD5 showing Cu binding histidines and disulfide in aqua and pink. The predicted omega site for GPI anchorage is red, and red underline indicates the ser/thr rich region of CalbSOD5 and CalbSOD4 missing in CaurSOD4. Asterisks indicate amino acid identity and dots, similarity.

The spent growth media was concentrated, treated with the PNGase F N-glycosidase to remove sugars and analyzed by immunoblot for the presence of a Cu-only SOD. For these studies, an anti-C. albicans SOD5 antibody was used that cross reacts well with CaurSOD4 based on studies with recombinant protein. In BPS treated cultures, C. auris produced immunoreactive bands of ≈17 and ≈20 kDa that were absent under Fe-replete (-BPS) conditions (Fig. 1A top). A similar immunoreactive doublet is seen when recombinant CaurSOD4 is expressed heterologously in C. albicans and secreted (Fig. 1B, see also Fig. 2A). The faster migrating species on immunoblots appears to be a proteolytic product of the SOD (see ahead Fig. 2C and Methods), while the 20-21 kDa species following PNGase F treatment of C. auris samples is consistent with the predicted 20.7 kDa molecular weight of full length CaurSOD4. C. auris expresses one other Cu-only SOD, named SOD6, however the predicted molecular weight of SOD6 is 30 kDa and SOD6 is not induced by Fe starvation ²³. We therefore conclude that the immunoreactive bands produced specifically by Fe-starved C. auris cells represent extracellular CaurSOD4. The appearance of CaurSOD4 in the growth media of C. auris cells tracks well with SOD enzymatic activity as seen in a native gel assay for SOD activity (Fig. 1A bottom). These studies support the notion that under Fe starvation conditions, C. auris releases enzymatically active CaurSOD4 into the extracellular environment.

Fig. 2: Cu activation and oxidation of the disulfide in secreted CaurSOD4.

(A) The spent growth media from the indicated C. albicans strains expressing recombinant CaurSOD4 were assayed for SOD activity (top) and CaurSOD4 protein levels (bottom) as in Fig. 1A and 1B respectively. “Con”, control C. albicans not expressing CaurSOD4. (B) Redox western of mammalian SOD1 where samples were incubated with or without 100 mM DTT prior to SDS-gel electrophoresis. “Ox” and “Red” indicate positions of disulfide-oxidized and reduced SOD1, respectively. (C) Samples of CaurSOD4 secreted from C. albicans were treated with Endo H or PNGase F in the presence or absence of DTT prior to reducing and denaturing gel electrophoresis. The 16-17 kDa CaurSOD4 specie appears to be a proteolytic product generated under reducing conditions of sample processing (see Methods). In lane 4, the high MW bands and diminished levels of the ≈20 kDA band indicate incomplete deglycosylation by PNGase F under non-reducing conditions. (D - F) Samples containing CaurSOD4 secreted from C. albicans (D,E) or C. auris (F) were incubated with Cu where indicated (+Cu) and were either subject to SOD activity analysis by the native gel assay (E, F bottom) or treated with Endo H under non-reducing conditions and subject to redox western as in part B (D and F top).

The presence of CaurSOD4 in the extracellular medium may seem at odds with its predicted attachment to the cell wall or plasma membrane through GPI anchors. However, soluble Cu-only SODs in the growth media have also been seen for H. capsulatum ¹⁷ and C. albicans ²³, and studies in C. albicans have shown that these SODs are derived from the cell wall, apparently released during cell wall remodeling ²³. C. albicans SOD4 (CalbSOD4) is induced under the same Fe starvation conditions as C. auris SOD4, and because it retains a beta glucan from the cell wall, CalbSOD4 in the growth media migrates as ≈35 kDa as opposed to the anticipated ≈20 kDa unmodified SOD (Fig. 1C lane 4) ²³. By sharp contrast, there is no similar evidence for beta-glucan modified CaurSOD4; the protein secreted from C. auris migrates in accordance with the predicted 20.7 kDa molecular weight of unmodified CaurSOD4 (Fig. 1A). Based on these findings, we conclude that C. auris secretes at least a portion of its SOD4 directly into the extracellular environment without prior attachment to the cell wall.

It is important to note that C. auris SOD4 does contain a predicted omega site for GPI anchorage (Fig. 1D) ²⁵. In yeast cells, sequences immediately upstream of the omega site can influence localization, and ser/thr rich regions promote attachment to the cell wall ²⁶. Cu-only SODs from several Candida sp, including C. albicans, Candida tropicalis and Candida dubliniensis all contain an extended ser/thr rich region ²³, an example of which is shown in Fig. 1D with C. albicans SOD4 and SOD5. However, CaurSOD4 is missing this extension (Fig. 1D). The lack of this ser/thr rich region for promoting cell wall attachment may explain how CaurSOD4 is directly exported from the secretory pathway. It is noteworthy that certain Aspergillus sp have evolved with Cu-only SODs that lack the GPI anchor site and are predicted to be directly secreted into the extracellular environment ²⁰. C. auris has been shown to be subject to ROS attack from neutrophils in vitro ²⁷ and the direct secretion of active SOD could prove advantageous to the fungus during such attack from host cells. This secretion may also be beneficial for therapeutic strategies targeting these extracellular enzymes, as they would be readily accessible to the drug.

Cu acquisition and disulfide oxidation of C. auris SOD4 in live fungal cultures

In eukaryotes, cell surface and secreted cuproproteins are known to acquire their Cu co-factor within the secretory pathway from a Golgi Cu ATPase, denotated as mammalian ATP7a/ATP7b or yeast CCC2 ^28–30. In accordance with this requirement for CCC2, C. albicans ccc2 mutants have defects in cell surface Cu-dependent ferroxidases ³⁰. Surprisingly, these same ccc2 mutants show no defect in activity of secreted SOD5; instead the Cu-only SOD appeared to be activated by Cu outside the cell ¹¹. We were curious as to whether the same holds true for C. auris SOD4, particularly since this SOD can directly emerge from the secretory pathway without prior cell wall attachment (Fig. 1A). The ccc2 mutant is currently not available for C. auris, and we therefore used our C. albicans expression system that faithfully secretes enzymatically active recombinant CaurSOD4. In the experiment of Fig. 2A, fungal cells expressing CaurSOD4 were harvested and incubated in fresh media for 1 hour to examine enzymatic activity of freshly secreted SOD. WT cells exhibit robust SOD activity in the growth media (Fig. 2A top, lane 1) that was absent in control C. albicans cells not expressing CaurSOD4 (lane 3). Activity of CaurSOD4 was not diminished by ccc2 null mutations (Fig. 2A, lane 2), demonstrating that the SOD acquired its Cu independent of CCC2. We conclude that as is the case with C. albicans SOD5, CaurSOD4 can acquire its Cu outside the cell, independent of the secretory pathway. However, we cannot exclude the possibility that the SOD attained its Cu in the secretory pathway but from a non-CCC2 source. We disfavor this possibility because other secreted cuproproteins are dependent on CCC2 ³⁰ and in C. albicans, the only other Cu-ATPase is found at the cell surface ³¹. Compared to our findings with fungal Cu-only SODs, the extracellular and secreted Cu/Zn SOD3 of mammals acquires Cu in the secretory pathway through ATP7a (equivalent to CCC2) ³². Bimetallic Cu/Zn SODs all contain a closed Cu binding site in contrast to the exposed Cu site of Cu-only SODs ¹¹. The unique open active site of Cu-only SODs may be ideally suited to capture Cu(II) outside the cell from a wide array of environmental sources of Cu(II).

As with all members of the Cu-containing SOD family including bimetallic Cu/Zn SODs and the fungal Cu-only SODs, CaurSOD4 contains two cysteines predicted to form a disulfide (Fig. 1D). In Cu/Zn SODs, this disulfide is important for protein stability and enzymatic activity ^{33 34}. We have previously shown that C. albicans SOD5 is secreted from fungal cells in the disulfide-oxidized form even as apo protein ¹¹ and we sought to test if the same were true for C. auris SOD4. Analysis of disulfides can be achieved through redox westerns, where in the absence of reducing agents (DTT), disulfide-oxidized SODs migrate slightly faster on denaturing gels, an example shown in Fig. 2B with Cu/Zn SOD1. To analyze the disulfide of secreted CaurSOD4, samples were deglycosylated with either PNGase F (broad specificity N-glycosidase) or Endo H (selective for high mannose chains) in the absence of the recommended DTT reducing agent. Unexpectedly, we observed that excluding DTT from the deglycosylation step minimized production of the ≈17 kDa product (Fig. 2C lane 2 and 4), indicating proteolysis of CaurSOD4 in DTT containing samples. The absence of DTT also somewhat inhibited deglycosylation by PNGase F, but not Endo H (Fig. 2C lanes 2 and 4), and therefore Endo H was chosen for redox western studies. When DTT was also eliminated during electrophoresis, Endo H-treated CaurSOD4 exhibited a downwards shift in mobility (Fig. 2D lane 2), indicative of a largely oxidized disulfide, similar to results with disulfide-oxidized SOD1 (Fig. 2B). It is important to note that CaurSOD4 secreted from C. albicans is not fully Cu-loaded as activity is increased somewhat by supplementing Cu(II) salts (Fig. 2E and Fig. 5A). Even with part apo protein in the cultures, the disulfide of CaurSOD4 is fully oxidized, and there was no change in the disulfide status with Cu(II) treatment (Fig. 2D lanes 2 and 4). We additionally examined the disulfide of native CaurSOD4 secreted from BPS-treated C. auris cultures. In this case, the enzyme appears fully loaded with Cu, since activity is not increased with Cu(II) supplements (Fig. 2F bottom). As with recombinant protein secreted from C. albicans (Fig. 2D), CaurSOD4 in C. auris cultures exhibited a downward shift in mobility without reducing agents and hence is largely disulfide-oxidized (Fig. 2F top). Thus, as is the case with C. albicans SOD5, CaurSOD4 from fungal cultures is disulfide-oxidized. The disulfide may be oxidized in the secretory pathway as has been described for mammalian extracellular Cu/Zn SOD3 ³⁵, or oxidized once Cu-only SODs are exposed to the aerobic extracellular milieu. This disulfide is expected to help stabilize the protein as has been described for Cu/Zn SODs^{33 34}. The precise cellular site of disulfide oxidation and role of the disulfide in Cu-only SOD stability are intriguing subjects of future study.

Fig. 5. CaurSOD4 activity dependent on Cu but not other metals.

(A,B) C. albicans expressing CaurSOD4 was cultured in media containing the indicated concentrations of CuCl₂ or ZnCl₂ (A) or in media containing 0.25 μM CuCl₂ supplemented where indicated with 10 μM FeCl₂ or MnCl₂. Samples were assayed for SOD activity (top) and CaurSOD4 protein (bottom) as in Fig. 2A. (C) Apo recombinant CaurSOD4 was dialyzed against the indicated metals as described in Methods and assayed for activity by the native gel assay.

The metal binding properties of C. auris SOD4

To investigate the biochemical properties of CaurSOD4 further, a recombinant version of the protein was produced in E. coli that lacked the C-terminal GPI anchor motif and N-terminal signal peptide, amino acids 30 to 173. CaurSOD4 was isolated from inclusion bodies using a purification scheme modified from that established for CalbSOD5 and CalbSOD4 ^{11, 12, 23}. In previous purification studies, C. albicans SOD4 and SOD5 differed in their ability to bind anion exchange resin due to differential surface charges ²³. We find that CaurSOD4 behaves more akin to C. albicans SOD5 and does not bind anion exchange resin (see Methods). Although by primary amino acid sequence, CaurSOD4 appears slightly more related to CalbSOD4 than CalbSOD5 (49.6 versus 44.6% identity across the core SOD region), using modeling programs, CaurSOD4 is predicted to have a surface charge distribution that is more similar to that of CalbSOD5 (Fig. 3A). CaurSOD4 purified to homogeneity as apo-protein contains a fully oxidized disulfide (Fig. 3B). This protein with an oxidized disulfide was readily reconstituted with Cu(II) to a 1:1 stoichiometry yielding enzymatically active SOD by the native gel assay (Fig. 3C). This result was corroborated using a more quantitative assay for SOD activity based on inhibition of the reduction the water-soluble tetrazolium salt WST-1 ^{14, 36, 37}. Purified recombinant CaurSOD4 was active in the WST-1 assay over a range of concentrations (100 ng – 500 ng) comparable to CalbSOD5 (Fig. 3D,E). Since recombinant CaurSOD4 is fully active upon reconstitution solely with Cu, CaurSOD4 is a Cu-only SOD enzyme.

Fig. 3. CaurSOD4 expression in E. coli.

(A) CaurSOD4 structural predictions (right) were made using MODELLER and compared to CalbSOD4 and CalbSOD5 PDB: 4N3T. Electrostatic surface charges are displayed with a scale ranging from −5 (red) to +5 (blue) in units of kBT/ec at pH 8.0. Orange arrow indicates C-terminus. (B) Recombinant CaurSOD4 purified to homogeneity from E. coli was analyzed by denaturing gel electrophoresis with or without DTT to analyze the disulfide as in Fig. 2, followed by Coomassie staining. (C-E) Recombinant CaurSOD4 and CalbSOD5 reconstituted with Cu were analyzed for enzymatic activity in the native gel assay as in Fig. 1A (C) or by analyzing the indicated amounts of protein in the WST1 assay (D,E). Controls include duplicate buffer alone and 150 ng SOD reaction without xanthine oxidase (-XO). The results of panel D are quantified in E; values shown are the averages of duplicate measurements with error ≤ 5%, representative of two experimental trials.

Unlike Cu-only SODs with just a single metal site, Cu/Zn SODs contain two metal binding sites and the Cu and Zn co-factors can bind the wrong site under certain conditions. For example, Zn can occupy the Cu site of Cu/Zn SOD1 in vitro ^{38, 39} and in cells in vivo ⁴⁰, and Cu can migrate to an open Zn site in Cu/Zn SOD1 ⁴¹. In contrast, our previous studies showed that only Cu and not Zn will bind the active site of C. albicans SOD5 ¹¹. We tested whether this high selectivity for Cu over Zn could be extended to C. auris SOD4. In the experiments of Fig. 4, samples of apo SOD were incubated with Cu(II) or Zn salts, or with equimolar amounts of both metals, or without metals as a control. Unbound metals were removed by extensive dialysis and metal content analyzed by inductively coupled plasma mass spectrometry (ICP-MS). SOD activity was also assayed using the native gel assay. As seen in Fig. 4A, only Cu(II) stably bound to CaurSOD4, not Zn, and the presence of Zn did not interfere with Cu(II) binding to the SOD. Similar results were obtained with CalbSOD5 examined in parallel (Fig. 4B). Notably, the Cu-bound proteins resulting from reconstitution with either Cu(II) or with Cu(II) and Zn were active in the native gel assay (Fig. 4A and 4B bottom). These results demonstrate that like C. albicans SOD5, C. auris SOD4 has a high selectivity for binding Cu(II) over Zn.

Fig. 4: Cu versus Zn binding to CaurSOD4.

Cu(II) and Zn binding of recombinant CaurSOD4 (A, C) and CalbSOD5 (B, D) was assessed as described in Methods. Metal binding and subsequent dialysis to remove excess metals was carried out at pH 4.9 (A,B) or 7.4 (C,D). Graphs indicate the stoichiometry of Cu(II) (grey bars) and Zn (white bars) binding per mole SOD protein following incubation with CuCl₂ (“+Cu”), ZnCl₂ (“+Zn”), both metals (“+Cu +Zn”) or with no metals (“-“). Below each graph is SOD activity of the corresponding sample by the native gel assay. Results are representative of three replicates. The ability of Cu to inhibit Zn binding at pH 7.4 is statistically significant as determine by unpaired T-test, **P ≤0.008.

The aforementioned metal binding experiments were conducted at an acidic (4.9) pH, similar to previous metal reconstitution studies with C. albicans SOD5 (pH 5.5) ^{11, 12}. Candida species can thrive in host environments that range from acidic to alkaline pH ⁴² and since these SODs are in direct contact with the extracellular environment, we additionally examined metal binding at a higher pH (7.4). C. auris SOD4 stably binds Cu(II) at this higher pH, yet surprisingly, we also observed binding of Zn at this pH (Fig. 4C). This binding of Zn appears to occur in the active site, because Cu(II) readily competed for Zn binding, yielding enzymatically active C. auris SOD4 (Fig. 4C top and bottom). Similar results were obtained with C. albicans SOD5 (Fig. 4D). The binding of Zn to the Cu site at pH 7.4 may reflect de-protonation of the active site histidines, as predicted for Zn-histidine interactions ⁴³. Even though Zn has the capacity to bind the SODs at higher pH, this should not preclude Cu(II) from activating the enzyme in vivo, as Cu(II) effectively competes with Zn when both metals are present (Fig. 4C, D).

We also examined Cu and Zn interactions with CaurSOD4 in live fungal cultures where we know the SOD is activated by extracellular Cu(II). The growth media for these studies typically has 0.25 μM Cu(II) and 2.5 μM Zn, yet even with this 10 fold molar excess of Zn, the secreted CaurSOD4 is active (Fig. 5A, lane 1). Activity was not diminished upon increasing Zn to a 50-fold excess over Cu(II) (Fig. 5A, lane 3). Cu(II) supplements did increase activity to a certain degree as expected (Fig. 5A, lane 2) but Zn had no impact on this Cu(II) activation of CaurSOD4 (Fig. 5A, lane 4). We also find that supplementing excess levels of other metals including Mn(II) and Fe(III) salts did not inhibit activity of C. auris SOD4 (Fig. 5B). Mn and Fe are co-factors for the unrelated Mn- and Fe-containing family of SODs, but due to metal-specific redox properties, Mn and Fe cannot serve as co-factors for the Cu-SOD family ⁴⁴. To confirm that Mn and Fe cannot activate CaurSOD4, the apo recombinant protein was subjected to dialysis against the various metals as was done in Fig. 4, and we observe that activity is only achieved in the presence of Cu(II) (Fig. 5C). Together, the studies of Figs. 4 and 5 show that C. auris SOD4 has a strong preference for binding Cu(II) in the catalytic site and we predict that during infection, the enzyme will retain activity in the mixed metal environment of the animal host.

Our findings here with CaurSOD4 are vastly different from what has been reported for the bimetallic intracellular Cu/Zn SOD of fungi and animals. Even at low pH, Zn can bind to the Cu site of Cu/Zn SOD1 in vitro ³⁸. However in live cells, only Cu and not Zn binds the Cu site of intracellular SOD1, and this specificity is achieved through a partner protein, the CCS copper chaperone ^44–46. There are no known Cu chaperones for fungal Cu-only SODs and to prevent mis-metallation of these enzymes, the SODs have evolved with an intrinsic ability to bind Cu in the active site over other metals. This innate ability to exclude Zn from the active site should prove advantageous to a SOD that acquires Cu outside the cell where metal homeostasis is poorly controlled.

The Cu binding affinity of CaurSOD4 and its susceptibility to inhibition by metal chelators

We previously demonstrated that the Cu co-factor of CalbSOD5 is more labile than that of mammalian bimetallic Cu/Zn SOD1 and Cu(II) binds the active site with lower affinity ¹⁴. To understand if these properties can be extended to CaurSOD4, we determined the log K (stability constant) of Cu(II) binding using a previously described method based on PAR-Cu complex formation where measurements were conducted in phosphate buffer, pH 7.4 ^{14, 47, 48}. Using this method, we observed a mean log K value of 16.5 for Cu(II) binding to CaurSOD4 (Fig. 6A). C. albicans SOD5 examined in parallel showed a comparable Cu(II) binding constant of 15.6 (Fig. 6A), similar to values of 15.5 previously published for C. albicans SOD4 and SOD5 ^{14, 23}. The marginal increase in Cu(II) binding observed with CaurSOD4 was determined to not be statistically significant (P value, 0.29; Fig. 6A). We conclude that C. auris SOD4 has an inherent Cu(II) binding affinity that is similar to that of C. albicans Cu-only SODs.

Fig. 6: Cu binding affinity and chemical inhibition of CaurSOD4.

(A) The log K stability constants for Cu binding to CaurSOD4 in comparison to CalbSOD5 was carried out by a PAR metal binding assay as described in Methods. Results are mean of six samples over two experimental trials, with standard deviation (SD) in parenthesis. By unpaired T-test there is no statistically significant difference between the log K measurements for CaurSOD4 and CalbSOD5 (P = 0.289). (B-E) The indicated metal binding compounds with structures shown were tested for dose-dependent inhibition of CaurSOD4 in comparison to CalbSOD5 and mammalian Cu/Zn SOD1 using the WST-1 assay. Results are representative of two experimental trials with four samples, error bars are shown. (F) The extracellular growth medium from C. albicans secreting C. auris SOD4 was treated with the indicated concentrations of PZ, or DMSO-only control (0 μM PZ) for 1 hour prior to analysis. The secreted CaurSOD4 was tested for SOD activity (top) and CaurSOD4 protein (bottom) as in Fig 5A,B.

Compared to Cu-only SODs, measurements of the log K for Cu binding to bimetallic Cu/Zn SOD1 have been difficult to obtain due to extraordinary affinity of SOD1 for Cu. Measurements have been obtained by partially destabilizing the enzyme through chemical denaturants ⁴⁹, by reducing the disulfide ⁵⁰ or by removing the Zn co-factor ⁴⁷. Interestingly, the log K for Cu(II) binding to Zn-free SOD1 (log K ~15.6 – 16.1) ⁴⁷ is very similar to that of Cu-only SODs, indicating a role for the missing Zn in the lower Cu binding affinity of fungal Cu-only SODs.

As a consequence of its relatively labile Cu binding site, several small molecule metal chelators were previously identified as chemical inhibitors of CalbSOD5 activity ¹⁴. To determine if CaurSOD4 was similarly susceptible to inhibition by metal binding compounds, we tested select small molecules previously shown to selectively inhibit CalbSOD5 and not mammalian Cu/Zn SOD1. The WST-1 SOD activity assay was used as a robust and reliable reporter for SOD inhibition ^{14, 36, 37}. Four compounds were chosen for study. These include the Cu- and Zn-binding pyrithione zinc (PZ) ⁵¹, the 8-hydroxyquinoline Cu-binding chloroxine ⁵² and clioquinol ⁵³, and the hydroxamic acid metal binding ciclopirox ⁵⁴. As seen in Fig. 6B-E, all four compounds were capable of inhibiting recombinant CaurSOD4 in a manner virtually identical to that of recombinant CalbSOD5. Importantly, even at the highest concentration of inhibitor tested, mammalian Cu/Zn SOD1 remained fully active (Fig. 6B-E).

The most effective inhibition was seen with PZ and we tested whether this drug could also inhibit CaurSOD4 secreted from live fungal cultures in its glycosylated form. As seen in Fig. 6F, CaurSOD4 in fungal cultures is susceptible to PZ inhibition, with 5 μM PZ being sufficient to abolish SOD activity. Secreted glycoslyated CaurSOD4 is susceptible to specific inhibition by small molecular chelators, similar to previously published findings with CalbSOD5 ¹⁴.

Based on its Cu(II) binding properties (relatively low affinity, susceptibility to metal binding inhibitors) CaurSOD4 could very well serve as a new therapeutic anti-fungal target, assuming that CaurSOD4 is indeed a virulence factor for the fungus as expected. Across a diverse range of fungi that have been tested, elimination of a single Cu-only SOD results in a drastic loss of virulence. In addition to C. albicans ^{14, 16}, Cu-only SODs are essential for pathogenesis of the human pulmonary and systemic mycosis pathogens Histoplasma capsulatum ¹⁷, and Paracoccidioides sp ¹⁸, the insect pathogenic fungus Beauveria bassiana ¹⁹, and the plant fungal pathogens Fusarium oxysporum ²⁰ and Puccinia striiformis ²¹. Regardless of the host environments (plants, insects, animals) the extracellular Cu-only SODs act in front-line defense to effectively combat the oxidative insults of host immunity ^{14, 16–22, 55, 56}. The same is expected for C. auris when virulence can be examined using appropriate animal models of C. auris infections.

Conclusions

With the rapid rising incidence of C. auris infections and the remarkable resistance of this microbe to all classes of known anti-fungals, there is an urgent need for new therapeutic strategies. Based on the biochemical studies presented here, C. auris SOD4 is an attractive target. In a wide array of other fungi, this enzyme is essential for pathogenesis and in C. auris, the Cu-only SOD4 has a labile Cu binding site that is highly susceptible to inhibition by metal binding drugs and chelators. Moreover, the exclusive extracellular existence of the Cu-only SOD should make this enzyme vulnerable to chemical inhibitors. Our studies here can pave the way for future analyses on the role of Cu-only SOD4 in C. auris pathogenesis and as an innovative target for developing antifungals.

Methods

Candida strains and growth conditions

The C. albicans strains used in this study were derived from clinical isolate SC5314 or from KC2 (ura3::imm434/ura3::imm434). The KC68 ccc2Δ/Δ strain (ccc2b::hisG/ccc2a:: hisG ura3::imm434/ura3::imm434) was engineered using the KC2 parent as described ³⁰ and was a gift from Daniel Kornitzer, IIT. KC2 and KC68 strains expressing recombinant CaurSOD4 were obtained by transformation with the pMT10 plasmid for expressing CaurSOD4 sequences +1 to +173 (lacking the predicted site for GPI anchorage) under the MET3 promoter ²³. The C. albicans SS101 sod4Δ/Δ and SS102 sod4Δ/Δ sod5Δ/Δ strains were derived from SC5314 as described ²³. The C. auris strain AR0388 is a clinical isolate obtained from CDC and FDA antibiotic resistance isolate bank and it originated from Southeast Asia.

All yeast strains were maintained by growth at 30°C in YPD media with 1% yeast extract, 2% peptone, and 2% glucose. For analysis of native endogenous Cu-only SODs secreted from fungi, C. auris or the SC5314 based C. albicans strains were grown in YPD media that was pre-filtered through a 10,000 Da molecular weight cut off centrifugal filter (MilliporeSigma). To induce native CaurSOD4 and CalbSOD4 expression, cultures were Fe-starved by supplementing this media with 150 μM of the Fe chelator bathophenanthrolinedisulfonic acid (BPS). 50 ml cultures were inoculated with 500 μl of a 1 x 10⁶ CFU/ml C. auris suspension or seeded at an OD₆₀₀ of 0.02 in the case of C. albicans, and all cultures grown at 30°C for 24 hrs. For the expression of recombinant CaurSOD4 lacking the GPI anchor, KC2 and KC68 C. albicans strains transformed with pMT10 were grown in a synthetic complete media formulated on yeast nitrogen base and lacking methionine and cysteine (SC-Cys-Met) as described ^{11, 12, 23}. Cells were grown in 15 or 50 ml cultures at 30°C with aeration to late log or stationary phase.

Biochemical analysis of SODs secreted in fungal cultures

To analyze SODs secreted from fungi, the aforementioned spent growth media samples were filtered through 0.2 micron filters and concentrated 100-fold using 10,000 Da molecular weight cut off filters (Vivaspin20, MilliporeSigma). For immunoblot analysis, the media concentrates were treated with either Endo H or PNGase F (New England Biolabs) to deglycosylate the secreted SODs as previously described ^{11, 12, 23} . In most cases, Endo H or PNGase F reactions were carried out in the presence of 40 mM DTT per manufacturers recommendations, except in studies of the SOD disulfide by redox western where DTT was eliminated (see below). Endo H and PNGase F gave similar results except the products of Endo H processing migrated slightly slower on denaturing gels, the equivalent of 1-2 kDa as seen in Fig. 2C. PNGase F has a broader specificity for N-linked carbohydrates (New England Biolabs), and some sugars may remain following Endo H treatment. Samples equivalent to 5-20 OD₆₀₀ cell units (for recombinant CaurSOD4) or equivalent to one eighth of the total 50 ml culture (for native CaurSOD4 and CalbSOD4) were subjected to either SDS reducing gel electrophoresis with 4-12% Bis-Tris acrylamide gels (Thermo Fisher Scientific) or in the case of redox westerns, SDS non-reducing gel electrophoresis where samples not incubated with 100 mM DTT retained an oxidized disulfide and migrated more rapidly during electrophoresis. Following electrophoresis with a 10-250 kDa protein ladder (BioRad) and transfer to PVDF membranes, membranes were incubated with anti-C. albicans SOD5 antibody ^{12, 23} at a 1:2500 dilution to detect C. auris SOD4, or with anti-C. albicans SOD4 antibody ²³ at 1:5000 to detect C. albicans SOD4. In the case of redox westerns involving SOD1 (as in Fig. 2B), 100 ng of bovine Cu/Zn SOD1 (Sigma) was used, and was probed with anti-SOD1 antibody ⁵⁷ at 1: 5000 dilution. A secondary goat anti-rabbit IgG Alexa Fluor 680 antibody (Thermo Fisher Scientific) was used at 1:10,000 dilution. Immunoblots were imaged using Odyssey software at the 700 nm channel. CaurSOD4 secreted from fungal cultures typically appears on immunoblots as a doublet of ≈17 and ≈20 kDa. The smaller species is greatly diminished when the spent growth media is processed without the DTT prescribed for Endo H or PNGase F. Elimination of DTT from gel electrophoresis also helps prevent CaurSOD4 proteolysis in C. auris samples. The spent growth media is expected to contain numerous peptidases secreted by the fungi, and under reducing and denaturing conditions that may impact the disulfide, CaurSOD4 appears susceptible to these or other proteases.

Enzymatic activity of secreted SODs was analyzed by a native gel assay for SOD activity. Concentrated spent medium samples without deglycosylation treatment were subjected to native gel electrophoresis on precast 10% Tris glycine gels for 35-45 min at 50 mA and 4°C. The equivalent to 10 – 25 OD₆₀₀ cell units were analyzed in the case of recombinant CaurSOD4, and with native CaurSOD4 and CalbSOD4, the equivalent to 2.5% of the total culture was used. Following electrophoresis, the gel was stained with NBT as described ^{11, 58} and SOD activity detected by achromatic bands.

Expression and Purification of C. auris SOD4 using an E. coli system

C. auris SOD4 was purified to homogeneity using an E. coli expression system similar to that developed for C. albicans SOD4 and SOD5 ^{11, 23}. C. auris SOD4 sequences encoding amino acids 30-173 were inserted into the backbone expression vector pAG10H-SOD5 ^{11, 12} lacking SOD5 sequences using Gibson Assembly Master Mix (New England Biolabs). SOD4 sequences were codon optimized for expression in E. coli by site directed mutagenesis. The resulting expression vector has C. albicans SOD4 preceded by a 10X- His tag and a TEV cleavage site. Following TEV cleavage, residues G-A-M-V precede residue 30 of the C. auris SOD4.

C. auris SOD4 protein was purified from E. coli using slight modifications of protocols published for C. albicans SOD4 and SOD5. E. coli inclusion bodies containing recombinant CaurSOD4 were solubilized in 8 M urea, and these denatured samples in 8 M urea, Tris (pH 8.0) were applied to a GE Healthcare Life Science His-Trap nickel column. CaurSOD4 was isolated using a linear imidazole gradient (0.1-0.5M). Fractions containing purified, unfolded protein (elution fractions) were pooled and 1.5 mM reduced glutathione was added. CaurSOD4 was then refolded by dialysis overnight against 50 mM Tris (pH 8.0) and 0.25 mM oxidized glutathione. CaurSOD4 His tag cleavage by His-tagged TEV protease and chromatographic isolation of tag-free CaurSOD4 was carried out as previously described ^{11, 12}. Refolded un-tagged CaurSOD4 was then loaded on a Mono-Q column anionic exchange column and the flow-through containing non-aggregated CaurSOD4 was collected. Pooled fractions of CaurSOD4 were dialyzed against 25 mM Tris (pH 8.0) and the purified refolded C. auris SOD4 was stripped of any residual metals by dialysis against 25 mM acetate pH 3.8, 10 mM EDTA, followed by sequential dialysis steps against 25 mM acetate pH 3. 8 and Bis-Tris pH 6.9. The apo protein at pH 6.9 was either stored at −80°C or subjected to metal reconstitution. Purification of apo C. albicans SOD5 from E. coli inclusion bodies followed published protocols ^{11, 12}.

Metal binding studies of purified CaurSOD4

Apo C. auris SOD4 and C. albicans SOD5 purified from E. coli were reconstituted with metals by dialysis similar to previously published procedures ^{11, 12}. With large scale protein preparations, metal binding and removal of unbound metals involved a series of 4 liter dialysis steps at 4°C. Dialysis with and without metals was carried out in 25 mM Bis-Tris pH 6.9 buffer in the case of C. auris SOD4, and in 25 mM acetate pH 5.5 buffer for C. albicans SOD5. Following final dialysis, samples were analyzed for Cu to protein stoichiometry by atomic absorption spectroscopy (AAS) using an AAnalyst600 graphite furnace atomic absorption spectrometer (PerkinElmer).

In smaller scale studies of Zn versus Cu binding as a function of pH, aliquots of apo CalbSOD5 and CaurSOD4 were first equilibrated in either acetate pH 4.9 or Tris pH 7.4 buffers by dialysis at 4°C overnight. For metal loading, 500 μL of 0.385 mg/mL (25 μM) apo SOD in either acetate pH 4.9 or Tris pH 7.4 buffer was dialyzed against 45 mls of the same buffer containing 250 μM of either CuSO₄, ZnCl₂ or both metals or no metals as control using a Thermo Fisher Scientific Slide-a-lyzer device. Samples were incubated with gentle rotating at 4°C for 24 hrs. Metals not stably bound were removed by additional dialysis steps without metals until Cu to protein equivalents in the CuSO₄-only samples reached 1:1 stoichiometry as measured by AAS. At that point, all samples were collected and concentrated using 3K molecular weight cut off filters (MilliporeSigma) to a final volume of 100 μL. The levels of metals stably bound to SOD protein were measured by inductively coupled plasma mass spectrometry (ICP-MS) in quantitative mode (Agilent 7700, University of Maryland, School of Pharmacy, Mass Spectrometry Center). Tests for CaurSOD4 activity following incubation with Mn and Fe salts (Fig. 5C) used a similar dialysis procedure carried out in 25 mM acetate pH 6.9 with 250 μM of either CuCl₂, ZnCl₂, MnCl₂ or Fe(ClO₄)₂.

To estimate the K_D for Cu binding to CaurSOD4 and CalbSOD5, Cu(II) release from Cu-bound SOD proteins was monitored by PAR-Cu(II) complex formation in phosphate buffer, pH 7.4 through absorbance at 500 nm as previously described ^{14, 59}.

SOD activity assays and inhibitor analyses of purified CaurSOD4

Activity of purified CaurSOD4 or CalbSOD5 was monitored either by native gel electrophoresis and NBT staining as described above using 10 μg of protein, or by an assay to monitor reduction of WST-1 (2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfo-phenyl)-2H-tetrazolium, monosodium salt) as previously described ^{14, 36, 37}. For the latter assay, indicated amounts of Cu bound CalbSOD5 or CaurSOD4 in 50 mM KPO₄ pH 7.8 were assayed in 200 μL reactions in triplicate in a 96-well plate format. Each reaction contained 0.1 mM xanthine, 0.2 mM EDTA, 0.16 milliunits xanthine oxidase and 0.3 mM WST-1 (Dojindo Molecular Technologies). Samples were incubated at 37°C for 45 minutes and SOD inhibition of WST-1 reduction by superoxide was measured by absorbance at 450 nm.

Small molecule inhibition of purified CaurSOD4 or CalbSOD5 or bovine Cu/Zn SOD1 (Sigma) was conducted using the WST-1 assay precisely as previously described ¹³. Percent inhibition of SOD activity with 300 ng SOD by the tested compounds was calculated by subtracting the A450 of the DMSO control from the sample A450 and dividing by the no SOD control. Inhibition values were plotted using Prism 9 GraphPad Software.

Abbreviations:

SOD

superoxide dismutase

CaurSOD4

C. auris SOD4

CalbSOD4

C. albicans SOD4

CalbSOD5

C. albicans SOD5

ROS

reactive oxygen species

BPS

bathophenanthrolinedisulfonic acid

AAS

atomic absorption spectroscopy

WST-1

(2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfo-phenyl)-2H-tetrazolium, monosodium salt)

ICP-MS

inductively coupled plasma mass spectrometry

PZ

pyrithione zinc