Neurospora crassa strains used in this study and their sources are listed in Table 1. Designations of ergosterol biosynthesis genes followed a previous study in N. crassa, in which all of the genes were named with those of their orthologs in yeast (Sun et al., 2013). A suggested partial ergosterol biosynthesis pathway in N. crassa is shown in Figure 1. However, as different N. crassa designations have been used by various authors in the past, we have compiled a list of comparative erg gene designations for clarification (Supplementary Table 1). Strains were cultured on Vogel’s medium with 1.5% agar [Vogel’s minimum medium, supplemented with 2% (w/vol) sucrose for slants or glucose for plates] (Vogel, 1956). Liquid Vogel’s minimum medium supplemented with 2% glucose was used for preparing samples for RNA, protein and sterol extraction. Sorbose plates (Vogel’s minimum medium, supplemented with 20 g/L sorbose, 0.5 g/L fructose and 0.5 g/L glucose) were used for electroporation transformation (Davis and de Serres, 1970). All cultures were grown at 28°C.

Antifungal compounds were added as needed. The antifungal agents include ketoconazole (KTC, Sigma), amorolfine (AMOR, Tokyo Chemical Industry) and terbinafine (TERB, J&K), dissolved in DMSO, Methanol and Methanol, respectively. Media were amended with the drugs to a final concentration of 2 mg/L, 0.375 mg/L and 3 mg/L, respectively, which resulted in N. crassa growth inhibition of at least 50%. Hygromycin B (Amresco) at a final concentration of 150 mg/L was used for screening fungal transformants and added as required.

In order to generate a strain in which transcription of erg11 (NCU02624, encoding sterol 14α-demethylase, Figure 1) could be controlled, the native promoter of erg11 was replaced with the copper responsive promoter Ptcu-1, based on a previously described strategy (Lamb et al., 2013). The construct diagram was shown in Figure 2A. First, a dominant selectable marker cassette was fused to the 5′ terminus of the tcu-1 promoter. The appropriate fragments of the tcu-1 promoter (1716 bp) and the Hygromycin B resistance gene hph, flanked by the trpC promoter and terminator (2169 bp), were amplified using primers Ptcu-1F/R and HphR-F/R, respectively (Supplementary Table 2). A homologous recombination technology-based kit (ClonExpress MultiS One Step Cloning Kit; Vazyme, Nanjing, China) was used to insert the two fragments into pBluescript II SK, resulting in the formation of plasmid pSKII-hph-Ptcu-1. Second, a split marker method was used to construct the mutant and two rounds of PCR were performed to obtain the recombination fragments. In the first round, four fragments, comprised of the 565 bp upstream homologous region of erg11, 834 bp downstream homologous region of erg11 (erg11N), upstream half of hph-Ptcu-1 and downstream half of hph-Ptcu-1, were amplified using primer pairs exPerg-up-F/R, exPerg-down-F/R, hph-Ptcu-up-F/R and hph-Ptcu-down-F/R, respectively (Supplementary Table 2). In addition, the downstream homologous region of erg11 was amplified from a construct containing a 5 × myc-6 × his-tagged version of the gene (Hu et al., unpublished). In the second round of amplification, the two upstream fragments and the two downstream fragments were joined, by fusion PCR, using primer pairs exPerg-up-F, hph-Ptcu-up-R and hph-Ptcu-down-F, exPerg-down-R, respectively (Supplementary Table 2) and resulted in 5′ recombination fragment and 3′ recombination fragment (Figure 2A). The obtained two recombination fragments were purified and then directly co-transformed into an available strain that was Ku70 deficient in a Ras-1^bd background by electroporation (He et al., 2006). Transformants were screened on plates containing 50 mg/L bathocuproinedisulfonic acid (BCS, Sigma) and 150 mg/L Hygromycin B and verified by PCR using primer pair exPerg-vF/R (Supplementary Table 2). After grown for several generations on slants containing 150 mg/L Hygromycin B, homozygous strains were screened and verified by PCR and used for the following experiments.

The erg11 inactivation strain was complemented by replacing the tcu-1 promoter with the native promoter of erg11. Briefly, three fragments, comprised of the terminator region of trpC, native promoter of erg11 and 5 × myc-6 × his tagged coding region of erg11, were amplified using primers TtrpC-F/R, Perg11F/R, and exPerg-down-F’/R’, respectively (Supplementary Table 2). The three fragments were then joined, by fusion PCR, using primer pairs TtrpC-F and exPerg-down-R’ (Supplementary Table 2). The obtained fragment was purified and then directly transformed into the erg11 inactivation strain Ptcu-1::erg11. Since the Ptcu-1::erg11 strain cannot grow on media containing Cu²⁺⁺, Cu²⁺ were used to screen the transformants. The genetic alterations introduced to the transformants were then verified by PCR.

The erg2 gene (NCU04156, encoding C-8 sterol isomerase, Figure 1) was deleted as described (Colot et al., 2006). Briefly, a 1153 bp promoter region and 1304 bp terminator region of erg2, as well as a 1416 bp fragment containing the Hygromycin B resistant gene, were amplified using primer pairs Perg2-up-F/R, Perg2-down-F/R and hph-F/R, respectively (Supplementary Table 2). The three fragments were subjected to fusion PCR using primers Perg2-up-F and Perg2-down-R (Supplementary Table 2). The obtained fragment was purified and then transformed into the wild-type strain FGSC #4200 by electroporation transformation as previously reported (Wang et al., 2015). Transformants were screened on plates containing 150 mg/L Hygromycin B and the gene replacement was verified by PCR.

For macroscopic and microscopic analysis of mutants, Vogel’s minimal medium dishes were inoculated with 2 μL aliquot of conidial suspension (2 × 10⁶ /mL) and the strains were cultured at 28°C. Growth media either contained or lacked CuSO₄ or was amended with BCS supplements. The final concentrations of CuSO₄ and BCS were 100 and 50 μM, respectively. Growth was documented after growing for at least 28 h and colony edges were subjected to microscopic examination. Spore suspensions were also inoculated onto Vogel’s slants with or without CuSO₄ or BCS supplements and incubated at 28°C. After 5–7 days, aerial hyphal formation and conidiation were analyzed.

Calcofluor white (0.5 g/L, Fluka, Sigma) and DAPI (Sango Biotech) were used to visualize cell wall and nucleus, respectively. Briefly, conidia of each strain were inoculated into liquid media to a final concentration of 2 × 10⁶ conidia per mL and allowed to germinate and grow for 10 h. The cultures were centrifuged and the pellets were fixed in 10% formalin solution for 30 min. The fixed mycelia were then washed with phosphate-buffered saline (PBS) for 2–3 times. After that, 5 μL of concentrated culture was mixed with calcofluor white and DAPI to a final concentration of 500 μg/L and 3 μg/mL, respectively, and stained for 10 min. Fluorescent microscopic observation was performed with a Zeiss ImagerA2 microscope equipped with a digital camera and DAPI Zeiss filter set was used for visualizing.

Samples for RNA extraction were prepared as previously described (Zhang et al., 2012) with some modifications. Briefly, conidia were first inoculated on plates containing Vogel’s liquid medium and grew for at least 24 h. The mycelial mat was then cut into fragments of about 2–4 mm². About 10–12 such fragments were used to inoculate 150 ml flasks containing 100 ml liquid Vogel’s medium. The cultures were incubated at 28°C at 200 rpm. After growing for 13.5 h, antifungal agents were added and the cultures were further grown for 22 h. Since erg11 is essential, the inactivation strain and its parental strain were grown in liquid Vogel’s medium containing 50 μM BCS during the first 13.5 h and then transferred to two new 150 ml flasks which were either unamended (control) or supplemented 200 μM CuSO₄, respectively. The control flasks, but not the 200 μM CuSO₄ flask also contained 50 μM BCS to support growth.

After 22 h of drug treatment, mycelial samples were harvested, flash frozen in liquid nitrogen, and ground into fine powder. Total RNA extraction was performed according to the standard TRIzol protocol (Invitrogen, Carlsbad, CA, United States). The cDNAs were synthesized from 3 μg total RNA using a cDNA synthesis kit (FastQuant RT Kit (with gDNase); TIANGEN, China) according to the manufacturer’s protocol.

Gene-specific primers were designed using the online tools PrimerQuest or Primer 6 and are listed in Supplementary Table 3. qRT-PCR was performed on a CFX96 multicolor real-time PCR detection system (Bio-Rad, Hercules, CA, United States) with SYBR green detection (KAPA SYBR^® FAST qPCR Kits; KAPA BIOSYSTEMS, Boston, MA, United States) according to the manufacturer’s instructions. Three independent experiments were carried out and each cDNA sample was analyzed at least in duplicate. The average threshold cycle (CT) values were used to calculate relative expression levels normalized to β-tubulin using the 2^-ΔΔCT method (Livak and Schmittgen, 2001). The significance of the differences between each two samples were evaluated by an unpaired two-tailed Student’s t-test using GraphPad Prism 7.0 and a p-value of less than 0.05 was considered significant.

For protein extraction, the collected mycelia were immediately frozen and then ground into fine powder for extraction. Proteins were extracted as previously reported (He et al., 2005) and quantified with the 2-D Quant Kit (GE, United States). Forty microgram protein of each sample was loaded for electrophoresis with SDS-PAGE and then transferred to a PVDF membrane (Merck Millipore, Germany). The tagged ERG11 protein level was detected using anti-myc antibodies (Abbkine, Redlands, CA, United States) using a standard protocol (Garceau et al., 1997).

For analyzing intracellular sterols, mycelial samples were heat-dried at 80°C and then ground into fine powder for extraction. Extraction and analysis were performed as previously described (Sun et al., 2013) using ergosterol (J&K Scientific Ltd.) and KTC standards.

For analyzing extracellular sterols, the cultured medium was extracted by chloroform after lyophilization. The extractions were then analyzed by HPLC-MS as described (Sun et al., 2013).

The significance of the differences between each two samples were evaluated by an unpaired two-tailed Student’s t-test using GraphPad Prism 7.0 and a p-value of less than 0.05 was considered as significant.

In order to test whether transcriptional responses to azoles can be induced by disruption of ergosterol biosynthesis, we first used a genetic approach to impair the function of an essential enzyme in the pathway –sterol 14α-demethylase. To do so, we generated a strain, Ptcu-1::erg11, in which the native N. crassa erg11 promoter was replaced by the promoter of a copper responsive gene tcu-1 (NCU00830, encoding the Ctr-domain containing copper transporter). When controlled by this promoter, expression of the target gene is repressed by high copper concentration while addition of the copper chelator BCS can release the suppression by copper (Ouyang et al., 2015).

The Ptcu-1::erg11 phenotype was characterized and compared to its parental bd ku70^RIP strain. As shown in Figure 2B, on Vogel’s plates containing 50 μM BCS or 100 μM CuSO₄, the growth of the parental strain was not affected, similar to that on normal Vogel’s plates. The growth of Ptcu-1::erg11 was severely inhibited on normal Vogel’s plate which containing 1 μM Cu²⁺, and only a small visible colony was formed (Figure 2B). The formed colonies of Ptcu-1::erg11 on the Vogel’s plates were also examined microscopically and clustered hyphae were observed (Supplementary Figure 1A). When 100 μM CuSO₄ were added to Vogel’s plates, no visible growth of Ptcu-1::erg11 was detected and only short germlings were observed under microscope (Figure 2B and Supplementary Figure 1A). When copper chelator BCS were added to Vogel’s plates, the growth of Ptcu-1::erg11 was indistinguishable from that of the parental strain. Similar results were observed on Vogel’s slants amended with 100 μM BCS or 250 μM CuSO₄ (Figure 2B).

When cultured in liquid medium containing 100 μM CuSO₄, conidia of Ptcu-1::erg11 geminated normally (Figure 2E). However, after about 2 rounds of nuclear division, hyphal tips started to swell and grew no further. Nonetheless, nuclear division was not arrested (Figure 2E). As a result, multiple nuclei clustered together at the swollen hyphal tips. In the medium amended with the Cu²⁺ chelator BCS, hyphal growth of Ptcu-1::erg11 was not arrested and nuclei were distributed evenly along the germ tubes (Figure 2E), as observed in the parental strain (Figure 2E). This result indicates that erg11 is required for hyphal elongation but not germination.

Since only the hyphal elongation was found to be affected in Ptcu-1::erg11 treated with Cu²⁺, the growth of the strains was examined by shifting the mycelial plugs from BCS to CuSO₄ plates. As shown in Supplementary Figure 1B, after shifting to Vogel’s plates containing 100 μM CuSO₄, the mycelium of Ptcu-1::erg11 strain grew for several hours and then ceased further colony growth, while the growth of parental strain was not affected. This result further verified that erg11 is required for hyphal elongation, although the mycelium could grow for a while, possibly due to the presence of sufficient quantities of erg11 mRNA or sterols within the mycelium to support growth.

To verify that sterol 14α-demethylase ERG11 is affected in the Ptcu-1::erg11 strain treated with CuSO₄, the sterol profile of this strain treated with CuSO₄ was compared to that of the strain treated with BCS. As shown in Figure 2D, CuSO₄ treatment resulted in the reduction of ergosterol to a level of approximately 56% of that treated with BCS along with the accumulation of lanosterol, eburicol and 14α-methyl-3,6-diol, similar to that found in N. crassa treated with azoles (Chen et al., 2016). The transcription and protein levels of erg11 and ERG11 in Ptcu-1::erg11 in the presence of BCS and CuSO₄, were also determined. In the presence of BCS, erg11 expression was as high as 50 folds as that of the parental strain and the protein was easily detected (Figure 3). Mycelium pre-grown in the medium with BCS was transferred to the medium with CuSO₄. The growth of the Ptcu-1::erg11 strain became slower than the parental strain. After 22 h at the time for RNA analysis, its growth was less than 50% that of the parental strain. Its erg11 expression was twofold lower than the parental strain and the protein was undetectable (Figure 3).

All the phenotypic features described above indicate that erg11 is an essential gene in N. crassa. To further verify that the observed effects were all due to the introduction of the inducible erg11 cassette, Ptcu-1::erg11 was complemented by replacing the Ptcu-1 promoter with the native erg11 promoter. Since Ptcu-1::erg11 could be inhibited by excess Cu²⁺, the transformants were screened on the medium containing 250 μM CuSO₄ and verified by PCR. The complemented strains restored wild-type characteristics to the Ptcu-1::erg11 strain (Figures 2B,C and Supplementary Figure 1), verifying the essential roles of erg11 for fungal growth.

To determine the relationship between the transcriptional response by the efflux-pump CDR4 encoding gene and ergosterol biosynthesis inhibition, we examined changes in cdr4 expression when erg11 was inactivated. In addition, we monitored the transcriptional changes of several N. crassa genes for ergosterol biosynthesis whose responses to azoles have been previously described (Sun et al., 2013). We compared the transcriptional responses between the Ptcu-1::erg11 strain and its parental counterpart when cultured in Vogel’s medium with BCS or CuSO₄. When compared to the parental strain, BCS amendment did not affect most of the tested genes except for erg11 in the Ptcu-1::erg11 strain. The presence of CuSO₄ did not alter the abundance of any of the tested gene in the parental strain. However, down-regulation was observed in erg11 expression measured in the Ptcu-1::erg11 strain cultured in the presence of the CuSO₄ amendment, indicating that suppression of transcription of the gene regulated by the tcu-1 promoter was functional (Figure 3B, left panel). However, a marked difference in transcript abundance of other tested genes was observed between the two strains grown in the presence of CuSO₄. Transcript levels of cdr4 as well as those of ergosterol biosynthesis genes were all significantly higher in the Ptcu-1::erg11 strain, in spite of the lack of KTC in the medium. While the growth of Ptcu-1::erg11 was inhibited after CuSO₄ was added to the medium, the mRNA levels of cdr4, erg5 (NCU05278, encoding sterol C-22 desaturase), erg24 (NCU08762, encoding sterol C-14 reductase) and erg6 (NCU03006, encoding sterol C-24 methyltransferase) were dramatically induced, by at least 28, 5, 6.5, and 33-fold, respectively, in a manner similar to the levels of induction imposed by the KTC treatment (Figure 3B). Hence, the presence of the drug was not a prerequisite for the transcriptional response observed.

In order to determine whether the transcriptional responses are caused by ergosterol depletion in N. crassa, similar to that of other fungi, we took advantage of the available Δerg3 (NCU06207, encoding sterol C-5 desaturase), Δerg4 [NCU01333, encoding sterol C-24(28) reductase] and Δerg5 strains and produced another Δerg2 strain (Table 1). These genes encode enzymes involved in last few steps of ergosterol biosynthesis, and deletion or mutation in any of these genes results in ergosterol depletion (Grindle and Farrow, 1978; Sun et al., 2013; Weichert et al., 2016). All the mentioned gene deletion mutants are viable (Figure 4A), similarly to that described in S. cerevisiae (Lees et al., 1995). We first determined the extent of ergosterol deficiency in these mutants by sterol profiling. Similar as wild type, ergosterol was easily detected in Δerg3 strain, possibly due to the present of a second copy of erg3 (Weichert et al., 2016). While in the Δerg2, Δerg4 and Δerg5 strains, no ergosterol were detected (Figure 4B). Rather, ergosta-5,7,22,24(28)-tetraenol was accumulated in the Δerg4 strain and ergosta-5,7,24(28)-trienol and ergosta-5,7-dienol were accumulated in the Δerg5 strain when compared to wild type (Figure 4B and Supplementary Table 4). In the Δerg2 strain, ergosta-5,8,22-trienol and fecosterol were accumulated (Supplementary Table 4).

When cultured in slants, a marked reduction in conidial abundance was observed in the Δerg2 strain. Furthermore, aerial hyphae grew higher along the slant sides and in a more compact manner (Figure 4A). The Δerg2 strain grew extremely slow on Vogel’s plates, exhibiting a growth of about 25% of wild type (Figure 4C). A few more branches at the hyphal tips of this mutant were also observed (Figure 4C). Deletion of erg4 also resulted in slow hyphal growth, albeit in a less severe manner than Δerg2, as this strain grew approximately 80% of wild type, yet its colony morphology was more compact and fluffy in appearance (Figure 4C). Thus, deletions in two of the analyzed genes encoding enzymes catalyzing last steps in ergosterol biosynthesis resulted in growth defects, while deletion of the others (erg3 and erg5) did not.

The expression of azole responsive genes including cdr4, erg11, erg2, erg5, erg6 and erg24 were then examined in these mutants by qRT-PCR. Interestingly, in a Δerg2 background, expression of cdr4 and erg24 were increased by 7 and 10-folds, respectively, while other genes tested were not affected (Figure 4D). This provides additional evidence that cdr4 responds to ergosterol biosynthesis inhibition. However, no change in gene expression was evident in the mutants of erg3, erg4 and erg5, which are downstream of erg2 (Figure 4D). Thus, the stress responses to azoles by cdr4 and ergosterol biosynthesis genes are not activated by ergosterol depletion.

Above results indicate that inhibition at different steps in ergosterol biosynthesis, such as deletion of erg2 or inactivation of erg11, could cause transcriptional responses by different genes. To further see transcriptional responses upon inhibition at more steps in ergosterol biosynthesis, we examined the transcriptional response of several genes to three different sterol biosynthesis inhibitors, KTC, AMOR and TERB. Different from azoles, TERB targets squalene epoxidase ERG1 (NCU08280), an enzyme upstream of sterol 14α-demethylase ERG11, while AMOR targets two enzymes downstream of ERG11 in the ergosterol biosynthetic pathway, sterol C8-C7 isomerase ERG2 and C-14 sterol reductase ERG24 (Lupetti et al., 2002) (Figure 1).

As cdr4 expression can be induced by inactivation of erg11 or deletion of erg2, it was not surprising that AMOR or KTC treatments led to an almost 5- and 13-fold increase in cdr4 expression, respectively (Figure 5), verifying that sterol inhibition does, in fact, induce cdr4 transcription. TERB had no effect on cdr4 expression (Supplementary Figure 2), suggesting that this pump may not be involved in expulsion of the compound from the cell, or, the inhibition of ergosterol biosynthesis by TERB cannot induce the expression of this pump.

We also analyzed the expression of several genes in ergosterol biosynthesis pathway, as affected by these antifungal drugs. As demonstrated above, transcription of these genes was induced by KTC (Figure 5). Similar to that observed in the Δerg2 strain, AMOR treatment only induced erg24 by 6 folds but not other erg genes (Figure 5). AMOR treatment also resulted in 4-fold induction of erg2, the gene encoding one of the AMOR targets. Similar to that observed in the case of cdr4, no significant transcriptional responses of these tested erg genes to TERB were detected (Supplementary Figure 2). These results together indicate the responsive genes are different upon inhibition at different steps of the ergosterol biosynthesis pathway.

We then examined the sterol content changes in N. crassa treated with different ergosterol biosynthesis inhibitors. In the presence of all the inhibitors tested, ergosterol levels were reduced by at least two folds, excluding TERB which resulted in ergosterol depletion to only 80% of wild-type levels (Supplementary Figure 3 and Supplementary Table 4). Along with the depletion of ergosterol, KTC treatment resulted in accumulation of lanosterol, eburicol and 14α-methyl-3,6-diol, AMOR treatment resulted in accumulation of ignosterol and ergosta-5,8,22-trienol, while TERB treatment resulted in accumulation of squalene and an unknown compound (Supplementary Table 4). Thus, inhibition at different steps in ergosterol biosynthesis pathway resulted in accumulation of different sterol intermediates in addition to depletion of ergosterol.

Based on above results, it is possible that the responses by different genes could be associated with accumulation of different sterol intermediates. If this is the case, when the accumulated intermediates are reduced or other intermediates are additionally accumulated, the transcriptional responses caused may be reduced or changed.

To examine the possibility, we treated the Δerg2 strain with KTC (Figure 1). The sterol profiles were firstly compared between the two strains with or without KTC. When treated with KTC, the accumulation of lanosterol, eburicol and 14α-methyl-3,6-diol along with the depletion of ergosterol was observed in wild type (Supplementary Figure 4 and Supplementary Table 4). Deletion of erg2 resulted in the accumulation of ergosta-5,8,22-trienol as its major sterol, while KTC treatment of this strain also lead to accumulation of lanosterol, eburicol and 14α-methyl-3,6-diol along with the depletion of ergosta-5,8, 22-trienol (Supplementary Figure 4 and Supplementary Table 4). Overall, KTC treatment in the Δerg2 strain could result in the depletion of sterol intermediates accumulated normally in the Δerg2 strain and the accumulation of sterol intermediates resulted from azole inhibition.

We then measured the expression of cdr4 and erg24 which respond to erg2 deletion, erg11 inactivation and corresponding antifungals. The expression of three additional genes, erg11, erg5 and erg6 which respond to only erg11 inactivation or KTC inhibition, was also analyzed. As shown in Figure 6, transcription of cdr4 was induced in a Δerg2 background. Yet when KTC was applied, the quantitative difference between the transcriptional responses of wild type and the mutant was minimal. Thus, it appears that the response of wild type to the drug is not dependent on a functional erg2. The expression of erg24 in the Δerg2 mutant treated with KTC was reduced when compared to that in the Δerg2 mutant (nine folds) and similar to that observed in wild type treated with the drug. The expression of erg5, erg6 and erg11 in the Δerg2 mutant treated with KTC also showed a similar response to KTC induction on its own, which is very different from that in the non-treated Δerg2 strain. All the results indicate that the transcriptional responses by cdr4 and genes in ergosterol biosynthesis pathway are associated with accumulation of sterol intermediates.

During the course of this study we have found that even though cdr4 expression can be induced by KTC and AMOR, when erg11 expression is repressed or erg2 is deleted, cdr4 expression is significantly increased, even in the absence of the sterol biosynthesis inhibitor. Since the CDR4 homologs in S. cerevisiae and C. albicans have been shown capable of transporting steroids (Kolaczkowski et al., 1996; Krishnamurthy et al., 1998), it is possible that N. crassa CDR4 may also function as a transporter of sterol intermediates. As the first step toward analyzing the possibility, we examined whether the expression of cdr4 is related to the accumulation of sterols. As inhibition or inactivation of erg11 causes accumulation of lanosterol, eburicol and the toxic sterol 14α-methyl-3,6-diol (Watson et al., 1989; Chen et al., 2016) (Supplementary Figure 4 and Supplementary Table 4), we monitored erg11 and cdr4 expression along a 24-h period after amending the growth medium with KTC. As can be seen from the results of the time course study, the expression of cdr4 was induced 3 h after KTC treatment while erg11 expression was induced within 1 h in the presence of the drug (Figure 7). This suggests that the accumulation of the excess sterols occurs before the onset of pump activity.

Do sterol intermediates accumulate in the Δcdr4 strain? To determine whether this is the case, we used HPLC-MS to analyze the sterol content in the cdr4 deletion mutant with or without KTC treatment. Deletion of cdr4 in N. crassa resulted in hypersensitive to azoles (Zhang et al., 2012). We firstly determined whether deletion of cdr4 would result in the accumulation of KTC. The accumulation of KTC in the Δcdr4 strain was 2.4-fold of that in wild type (Figure 8A), confirming CDR4 is the transporter of azoles, functionally the same to its ortholog Pdr5p in S. cerevisiae. As shown in Figure 8B, under normal condition, deletion of cdr4 had no obvious effect on the amount of intracellular sterols. In wild type, KTC treatment resulted in a reduction of total ergosterol content and the abnormal accumulation of sterols along the biosynthetic pathway (Figure 8B). In the Δcdr4 strain, a much more pronounced reduction (about threefolds) in total ergosterol content was evident, following KTC treatment (Figure 8B). This was accompanied by the increased accumulation of lanosterol, eburicol and the toxic sterol 14α-methyl-3,6-diol, which were 1.6-, 1.7- and 2.8-folds of that in wild type, respectively (Figure 8B).

To examine whether sterol intermediates can be exported via CDR4, the sterols in the cultured medium were also analyzed and compared between wild type and the Δcdr4 strain. Neither free ergosterol, nor lanosterol, eburicol or 14α-methyl-3,6-diol was detected in the media of any of the treatments. Moreover, other sterols, such as sterol acetate, were also not detected. Thus, the intracellular accumulation of sterols in Δcdr4 mutant is not due to reduced efflux of these sterols.