Transcriptional regulation of the synthesis and secretion of farnesol in the fungus Candida albicans: examination of the Homann transcription regulator knockout collection

Abstract Candida albicans is an efficient colonizer of human gastrointestinal tracts and skin and is an opportunistic pathogen. C. albicans exhibits morphological plasticity, and the ability to switch between yeast and filamentous morphologies is associated with virulence. One regulator of this switch is the quorum sensing molecule farnesol that is produced by C. albicans throughout growth. However, the synthesis, secretion, regulation, and turnover of farnesol are not fully understood. To address this, we used our improved farnesol assay to screen a transcription regulator knockout library for differences in farnesol accumulation in whole cultures, pellets, and supernatants. All screened mutants produced farnesol and they averaged 9.2× more farnesol in the pellet than the supernatant. Nineteen mutants had significant differences with ten mutants producing more farnesol than their SN152+ wild-type control strain while nine produced less. Seven mutants exhibited greater secretion of farnesol while two exhibited less. We examined the time course for farnesol accumulation in six mutants with the greatest accumulation differences and found that those differences persisted throughout growth and they were not time dependent. Significantly, two high-accumulating mutants did not exhibit the decay in farnesol levels during stationary phase characteristic of wild-type C. albicans, suggesting that a farnesol modification/degradation mechanism is absent in these mutants. Identifying these transcriptional regulators provides new insight into farnesol's physiological functions regarding cell cycle progression, white–opaque switching, yeast–mycelial dimorphism, and response to cellular stress.


Introduction
Candida albicans is an opportunistic pathogen that is present in most human gastrointestinal tracts and is an efficient colonizer of mucosal surfaces (Neville et al. 2015).A weakened immune system (human immunodeficiency virus, chemotherapy, and organ transplantation) or loss of competing flora (antibiotic treatment) allows for C. albicans to colonize and invade host tissues leading to candidiasis (Pfaller and Diekema 2007).Virulence of C. albicans is strongly linked to its ability to switch between yeast and filamentous morphologies (Romano 1966;Kadosh 2019) as mutants that have lost the ability to switch are avirulent (Lo et al. 1997;Saville et al. 2003).Dimorphism is also important for adaptation to different host environments (Noble et al. 2017;Alves et al. 2020) and evading host immune responses (Gow et al. 2012).One regulator of dimorphic switching is the quorum sensing molecule (QSM) farnesol (Hornby et al. 2001).Farnesol is produced by C. albicans cells continuously throughout growth (Boone et al. 2022), except in anaerobically grown cells (Dumitru et al. 2004) and opaque cells (Dumitru et al. 2007) where farnesol production is much lower.However, despite numerous publications on farnesol's role as a QSM and virulence factor, the actual mechanisms regulating its synthesis and secretion remain unclear.
As an approach to identifying these genes and mechanisms, we tested the hypothesis that farnesol synthesis and secretion are regulated by the transcriptional networks involved in morphogenesis and virulence.The transcription regulator (TR) knockout library generated by Homann et al. (2009) has 165 different TR deleted, in most cases with 2 independently derived knockouts.Thus, we grew each of the library strains in liquid culture and then assayed them for total, extracellular, and intracellular farnesol by the improved methods of Boone et al. (2022).Key features of this assay include prevention of analyte loss by avoiding filtration and minimizing evaporation, while incorporating simultaneous cell lysis and analyte extraction by ethyl acetate.The assay enables comparison of whole culture values with the sum of their cell pellets and supernatants.Our results suggest farnesol accumulation is coordinated with cell cycle progression, white-opaque switching, yeast-mycelial dimorphism, and responses to cell stress.Our results provide new insights into the physiological role of farnesol and understanding of the regulatory mechanisms underlying its production.

Strains and media
The homozygous TR deletion mutants were obtained from the TR knockout library provided by Dr. Alexander Johnson's lab (Homann et al. 2009)

Mutant Screen Report
were constructed from C. albicans SN152, an auxotroph for arginine, histidine, and leucine.Each of the TR mutants is auxotrophic for arginine.A wild-type control strain was created by reintroduction of a single allele of HIS1 and LEU2 into the parent strain (Homann et al. 2009).Throughout this paper, we will refer to this control strain simply as SN152 + .All assays with these mutants include the paired wild-type strain of the X2 and Y2 independently derived deletion collections.

Screening the TR deletion mutants for differences in farnesol accumulation and secretion
The TR mutant collection was transferred by a 48-pin replicator to YPD plates and grown at 30°C for 48 h.Mutants were then transferred to 3 mL YPD liquid medium and grown for 16 h at 30°C with rotary agitation at 250 rpm.These cultures were used as the inoculum (1:100) for 75 mL YPD in 250-mL flasks that were incubated 24 h at 30°C, 250 rpm.Fifty milliliters of this culture was used for dry weight determination and 20 mL (2 × 10 mL) was used for determination of farnesol production as described by Boone et al. (2022).In this method, 10 mL was used to measure whole culture values while the other remaining 10 mL was centrifuged to obtain the pellet and supernatant values.Due to the large number of mutants in the collection, the X collection mutants were assayed in batches of 10-20 mutants with the SN152 + control in triplicate for each batch.Mutants with farnesol accumulation of greater than 1 log 2 fold above or below the mean value for SN152 + were chosen for follow-up analysis.In follow-up analyses, E,E-farnesol accumulation was measured in at least 3 independent experiments using both the X and Y independently derived TR deletion mutants (n = 6).

Temporal dynamics of farnesol accumulation
To examine the time course of farnesol production, we assayed the farnesol accumulation of the 4 highest and 2 lowest accumulating mutants at several cell densities throughout growth.Mutants were inoculated into 6-mL YPD liquid medium and grown for 16 h at 30°C, 250 rpm.This culture was used as the inoculum (1:100) for 6 replicate flasks for each time point with 75 mL YPD per mutant strain.Farnesol measurements were taken 12, 18, 24, 36, 48, and 80 h post inoculation as described by Boone et al. (2022).Two independent time courses were performed for the mutant strains, and 3 independent time courses were performed for the SN152 + control.Farnesol accumulation values were normalized to the 50 mL dry weight at the indicated time point.To quantify differences in farnesol accumulation across time, area under the curve (AUC) analyses were performed in GraphPad Prism.

Statistical analysis
Statistical analyses were performed using Microsoft Excel (Version 16.61, Microsoft Office, Las Vegas, NV, USA) and GraphPad Prism Software (Version 9.5.0,San Diego, CA, USA).All biological data are represented as mean ± SD of at least 3 biological replicates for both the X and Y independent mutants unless otherwise stated.Mutants that did not have reproducible differences between the X and Y mutants were excluded and considered false positives.Normality and homogeneity of variance were assessed by the D' Agostino-Pearson omnibus (K2) test and Brown-Forsythe test, respectively.Differences between groups that were normally distributed and homoscedastic (equal variances) were assessed by one-way ANOVA with Dunnett's multiple comparisons test.Differences between groups that were not normally distributed or heteroscedastic were assessed by the Kruskal-Wallis test with Dunn's multiple comparisons tests.

Results
General design and reproducibility of the screen for farnesol regulators Homann et al. (2009) screened their collection of 165 independent TR mutants for possible differences in 55 growth conditions.It was a monumental piece of work.However, for practical reasons, they limited themselves to phenotypes discernible via colonies on agar plates.We have now extended the list of phenotypes to include accumulation of both intracellular and extracellular farnesol as determined by our improved gas chromatography (GC) with detection by flame ionization (FID) assay (Boone et al 2022).For the initial screen, 164 mutants were screened in 12 batches.One mutant ΔΔhfl1 (orf19.3063)was not screened as it exhibited a severe growth defect on YPD.Each batch included duplicate or triplicate SN152 + controls, for a total of 198 cultures and 594 GC-FID runs.Farnesol accumulation values were normalized on a per cell basis using the 50 mL dry weight value of the culture assayed.In this initial screen, the paired WT controls were run a total of 34 times and their whole culture mean E,E-farnesol accumulation value was 5.29 ± 1.12 ng/µL/50 mL dry weight (8.27 ± 2.8 µM).
For the rescreen, 27 strains were selected because their farnesol values differed from the average values for SN152 + by a log 2 fold change of >1.0.To these, we added 3 strains (ΔΔupc2, ΔΔtac1, and ΔΔcsr1) because prior reports from the literature indicated they were likely to have altered farnesol accumulations.These 30 strains were then rescreened 5 more times, 2 from the X plate and 3 from the Y plate (total n = 6) in 10 batches including the SN152 + parent in biological duplicate or triplicate, requiring 180 flasks and 540 GC measurements.Thus, the rescreen data set examines 30 transcriptional regulator mutants (n = 6), reporting the means ± SD for normalized farnesol accumulation as detected in whole cultures, pellets, and supernatants as well as the supernatant/pellet (S/P) ratio.Composite data sets for the initial screen (Supplementary File S1), rescreen (Supplementary File S2), and farnesol growth curves (Supplementary File S3) are available at Figshare.We now describe several highlights from these data sets.

Mutants altered in farnesol accumulation
Our initial screen of 164 strains, all from the X plate, was done with an n = 1 comparing whole culture values vs the sum of the cell pellet and supernatant values.The 164 TR mutants were assayed in 12 batches that always included a biological duplicate or triplicate of the SN152 + parent.The whole culture accumulation and S/P ratios are shown in Fig. 1.The distribution of the 34 values for SN152 + (red) is shown in Fig. 1 where they can be seen in comparison with the distribution of the 164 TR mutants (black).As expected (Boone et al. 2022), the whole culture (W ) and pellet (P) + supernatant (S) values were always very close to one another, with the average WPS relative error values (W − (P + S)/W ) being 0.07 ± 0.06 for E,E-farnesol.
The vast majority of the 164 TR mutants were unchanged from the SN152 + replicates (Fig. 1).We then identified all mutants that were >1 log 2 above or below the mean values, as indicated by the 4 dotted lines in Fig. 1, and designated them as mutants of interest to be rescreened with greater accuracy.There were 27 mutants with higher or lower farnesol accumulation.Six mutants (ΔΔcwt1, ΔΔrim101, ΔΔcrz2, ΔΔzcf1, ΔΔtea1, and ΔΔrme1) were designated as false positives because their rescreened values from the X and Y plates differed significantly or their rescreen values differed significantly from the initial screen.Thus, we identified 10 mutants with elevated farnesol (Fig. 2a), 9 with reduced farnesol (Fig. 2b), 7 with elevated S/P ratios (Fig. 2c), and 2 with lowered farnesol S/P ratios (Fig. 2c).Each of these mutants is described in greater detail in Table 1.
The 3 mutants chosen because of literature precedent had elevated farnesol accumulation but did not reach statistical significance (ΔΔcsr1, ΔΔupc2, and ΔΔtac1) (Table 1).Ganguly et al. (2011) reported decreased accumulation of farnesol by the ΔΔcsr1 mutant.Notably, the assay conditions used in our screen (YPD, 30°C) differ significantly from those used in prior studies of ΔΔcsr1 that employed biofilm-forming conditions at 37°C in SPIDER medium (Ganguly et al. 2011).Thus, the dynamics of farnesol accumulation are influenced not only by temperature but also by growth medium (Boone et al. 2022).

Mutants altered in the farnesol S/P ratio
The S/P ratios for farnesol were determined for each mutant (n = 1) in comparison to those for their SN152 + control (Fig. 1).The value for SN152 + (n = 22 batches) was 0.12 ± 0.02 for farnesol.As we had observed previously (Boone et al. 2022), farnesol was predominantly retained in the cell pellet.The S/P ratios were all   Log 2 fold change production values ± SD represent the mean farnesol accumulation value of 3 X and 3 Y (n = 6) independently derived mutants compared to the mean farnesol accumulation value of the paired SN152 parent (n = 22).For reference, a log 2 fold change of 1 represents a fold change of 2. Both over-and underaccumulating mutants are ordered by how much they exceed the SN152 parent.Differences between mutants (n = 6) and the SN152 + parent (n = 22 batches) were accessed by one-way ANOVA with Dunnett's multiple comparisons test using GraphPad Prism.Differences were considered significant at P < 0.05 ( * P < 0.05, ** P < 0.01, and *** P < 0.001).

b
The 4 mutants added because of literature precedent.
Two mutants, ΔΔhap5 and ΔΔcas5, are of special interest in that they exhibit greater secretion without being accompanied by greater total farnesol production (Fig. 1).Thus, they target secretion specifically.Mutants such as ΔΔswi4, ΔΔahr1, and ΔΔrap1 with much greater farnesol production (Fig. 2a and Table 1) could exhibit a higher S/P ratio because internal and external farnesol are in equilibrium or there is a need to keep internal farnesol levels below a critical threshold value.In both cases, the greater secretion would be an indirect consequence of greater production.

Temporal dynamics of farnesol accumulation
The initial screen and rescreen for farnesol accumulation were done at a single time point 24 h post inoculation.To verify that farnesol differences we observed were not time dependent, we extended the assay to 12, 18, 24, 36, 48, and 80 h post inoculation for 4 overaccumulating (ΔΔswi4, ΔΔahr1, ΔΔrap1, and ΔΔnrg1) and 2 underaccumulating (ΔΔmsn4 and ΔΔ19.6874) mutants (Figs. 3  and 4, respectively), with n = 2 for each mutant and n = 3 for the SN152 + control.Growth of the 6 mutant strains paralleled the growth of SN152 + (Figs.3a and 4a).The 4 overaccumulating mutants had elevated farnesol at all time points in both the whole culture (Fig. 3b) and supernatant fractions (Fig. 3d) while the 2 underaccumulating mutants had decreased farnesol in both the whole culture (Fig. 4b) and supernatant fractions (Fig. 4d).To compare total farnesol accumulations, AUC analyses were performed.The 4 overaccumulating mutants ΔΔswi4, ΔΔrap1, ΔΔahr1, and ΔΔnrg1 had significantly higher AUC in the whole culture (Fig. 3c) and supernatant (Fig. 3e) than the SN152 + parent while ΔΔorf19.6874 and ΔΔmsn4 had significantly lower supernatant AUC (Fig. 4e).Of great interest, 2 of the overaccumulating mutants (ΔΔswi4 and ΔΔrap1) still had significant farnesol remaining at 80 h (Fig. 3b and d).They did not exhibit the decay in farnesol levels during stationary phase characteristic of wild-type C. albicans (Boone et al. 2022) as well as SN152 + and the other 2 overaccumulating mutants tested (Fig. 3b), suggesting that a farnesol modification or degradation mechanism is absent from these 2 mutants.

Opaque phenotype of ΔΔtup1
One surprising result of the screen (Fig. 1) was that ΔΔtup1 did not appear to over accumulate farnesol in these assay conditions (Table 1).This result was surprising because we had previously reported that the ΔΔtup1 mutant produced 17 times more farnesol than its parent (Kebaara et al. 2008).A possible explanation for this discrepancy is that C. albicans can undergo phenotypic switching between 2 heritable states: white and opaque, each of which is normally stable for thousands of cell divisions.Switching between the 2 cell types is reversible and occurs without any chromosomal rearrangements or sequence changes.Critically, Tup1 is a key repressor of the opaque state and Alkafeef et al. (2018) have shown that loss of TUP1 is sufficient to induce the opaque phase, even in a MTL a/α background (Alkafeef et al. 2018).The pressure of ΔΔtup1 to spontaneously convert to the opaque phase is strong (Alkafeef et al. 2018), and we already know that wild-type opaque cells usually produce far less farnesol than do white cells (Dumitru et al. 2007).Given this, we investigated whether the ΔΔtup1 mutant in the Homann collection is in the opaque phase under the assay conditions used for this screen (YPD, 30°C).It is important to note that ΔΔtup1 cells are locked in the filamentous morphology (Kebaara et al. 2008), making visual determination between white and opaque cells more difficult.When plated on YPD medium supplemented with phloxine B and incubated at 30°C, the ΔΔtup1 colonies were pink while the SN152 + colonies plated under the same conditions were white (Fig. 5a).The pink colonies for ΔΔtup1 are consistent with the cells being in the opaque phase.This view was supported by microscopic examination of the ΔΔtup1 cells that showed numerous short but elongated cells consistent with the opaque state (Fig. 5b).The typical budding yeast cells of SN152 + are included for comparison (Fig. 5b).We are currently investigating the conditions that influence the farnesol production by white and opaque ΔΔtup1 cells.

Discussion
We examined 164 independent transcriptional regulator knockout mutants for their accumulation and secretion of farnesol.This project was undertaken to define the genes needed for farnesol's synthesis, secretion, and regulation.For logistical reasons, the initial screen and rescreen cultures were grown at 30°C in liquid YPD medium and sampled after 24 h when the cells were in early stationary phase.The 24-h time point was chosen based on our prior measurements with wild-type C. albicans SC5314 over 72 h of culture that showed that farnesol accumulation generally followed increased cell mass, peaked in early stationary phase, and then declined sharply after ∼30 h (Boone et al. 2022).Thus, the 24-h time point was intended to cover the peak of farnesol accumulation.Similarly, our choice of YPD as a growth medium was to insure that all of the TR mutants grew well by 24 h and indeed 164 out of 165 had done so.
With the advantage of now having reliable measures for farnesol presence in whole cultures, cell pellets, and supernatants, we can now approach the genetic basis for the synthesis, secretion, regulation, and turnover of farnesol.As expected, most of the TR mutants were unchanged for these parameters (Fig. 1) while a few potentially useful mutants gave significantly higher and lower levels of farnesol.Significantly, no mutants were entirely devoid of farnesol; its levels often varied but they never dropped to 0. A tentative conclusion from this observation is that farnesol has an essential physiological function over and above acting as a QSM and virulence factor or that it is the product of a redundant biosynthetic pathway.In examining mutants with significant differences in farnesol production, several patterns in the transcriptional networks emerge that may highlight farnesol's physiological functions.These networks include (1) white-opaque switching, (2) yeast-mycelia dimorphism, (3) response to cell stress, and (4) cell cycle progression.These networks integrate environmental signals to influence the morphology of C. albicans.Firstly, 3 of the TR studied, Ahr1, Ssn6, and Tup1, have crucial roles in white-opaque switching (Hernday et al. 2013(Hernday et al. , 2016;;Alkafeef et al. 2018).Given that farnesol is toxic to opaque cells (Dumitru et al. 2007), it is intriguing that the AHR1 deletion mutant overaccumulates farnesol while also promoting a higher frequency of white-opaque switching (Hernday et al. 2013).Additionally, it is unclear whether Ssn6 and Ahr1 participate in the response to farnesol, but there is ample precedent given that another transcriptional regulator of the white-opaque switch, Czf1, is involved in farnesol response.Secondly, in addition to Nrg1 and Tup1 (Kebaara et al. 2008), 4 more transcriptional regulators of yeast-mycelia dimorphism were found by this study to influence farnesol accumulation including Cph2, Ash1, Ssn6, and Tec1.Thirdly, C. albicans is especially resistant to farnesol compared to other fungi (Semighini et al. 2006;Brasch et al. 2013;Wang et al. 2014), but the mechanism of this tolerance is elusive.One possibility concerns the protective effect of using longer chain ubiquinones (UQ).C. albicans and Candida dubliniensis use UQ9 rather than UQ7 as in other Candida species or UQ6 as in Saccharomyces cerevisiae (Pathirana et al. 2020).Another possibility could be due to differences in their response to cellular stress.Given that Msn4 and Cas5 have roles in stress response (Nicholls et al. 2004;Xie et al. 2017) and influence farnesol accumulation and secretion (Fig. 2), it is likely they may also contribute to farnesol tolerance.Fourthly, Swi4 and Ahr1 have known roles in regulating cell cycle progression (Hussein et al. 2011;Sellam et al. 2019).Cell cycle arrest can induce filamentous growth of C. albicans, either through depletion of cyclins such as Cln3, Clb2, or Clb3 (Bachewich et al. 2005;Bensen et al. 2005), depletion of cell cycle polo-like kinase Cdc5 (Bachewich et al. 2003), or through treatment with cell cycle inhibitors like hydroxyurea (Bachewich et al. 2005;Chen et al. 2018).Interestingly, expression of the G1/S cyclins PCL2, CLN3, and HGC1 are farnesol regulated (Enjalbert and Whiteway 2005).Given that PLC2 and CLN3 are Swi4 regulated (Hussein et al. 2011), how these cyclins influence farnesol accumulation is an important question for subsequent studies.
There are many questions that can now be approached via the genetic information gleaned from this screen.Among them are as follows: (1) How is carbon flow through the farnesyl pyrophosphate (FPP) branch point regulated?(2) Is farnesol made directly from FPP by an appropriate pyrophosphatase (e.g.Dpp1p, Dpp2p, or Dpp3p) or are there other more easily regulated and/ or safer biosynthetic pathways?A pathway suggested from Arabidopsis invokes the necessary turnover of farnesylated proteins, where proteolytic degradation of those proteins releases farnesylated cysteine that can be cleaved by a ligase to release farnesaldehyde that is then converted to farnesol by a farnesol dehydrogenase (Bhandari et al. 2010).( 3) Is the disappearance of farnesol later in stationary phase due strictly to evaporation of a volatile molecule or to enzymatic conversion of farnesol to, for instance, 2,3-dihydrofarnesol (Brasch et al. 2013;Costa et al. 2020)?Significantly, regarding its ability to block germ tube formation and hyphal development, 2,3-dihydrofarnesol is inactive, with only 0.34% of the activity of farnesol (Shchepin et al. 2003).This question has now been partially resolved by extending the ΔΔswi4 and ΔΔrap1 mutant analyses to 80 h post inoculation (Fig. 3b).While wild-type SC5314 (Boone et al. 2022), SN152 + , and most of the mutants examined had little farnesol remaining by 80 h, ΔΔswi4 and ΔΔrap1 differed in that they maintained high levels of farnesol.Clearly, something other than evaporation is occurring.There is ample precedent for farnesol-modifying enzymes in fungi.For instance, trans-trans-farnesol can be converted to a cis-trans-farnesol by Helminthosporium sativum (Imai and Marumo 1974).Presumably, the corresponding enzymes for C. albicans are not made by Δswi4 and ΔΔrap1, and we are currently using these genetic clues to follow the metabolic disappearance of farnesol during stationary phase.

Fig. 1 .
Fig. 1.Initial screen E,E-farnesol S/P ratio vs E,E-farnesol whole culture accumulation at 24 h post inoculation.Data represent the accumulation values for each of 164 TR mutants (circle) (n = 1) and each SN152 + replicate (square) (n = 34).Dotted lines represent a log 2 fold change increase or decrease of 1 for farnesol from the mean for SN152 + .

Fig. 2 .
Fig. 2. TR mutants differing in farnesol accumulation and localization.Accumulation values were measured 24 h post inoculation at 30°C in YPD for each TR mutant in both the X (3) and Y (3) independent mutants (total n = 6) and the SN152 + parent (n = 22 batches).Data represent the mean accumulation value ± SD a, b) or mean farnesol S/P ratio ± SD c).The mutants added because of literature precedent are indicated by a number sign.Differences between groups that were normally distributed and that had equal variances c) were accessed by one-way ANOVA with Dunnett's multiple comparisons test.Differences between groups that were not normally distributed or heteroscedastic a, b) were accessed by Kruskal-Wallis test with Dunn's multiple comparisons tests.Differences were considered significant at P < 0.05 (*P < 0.05, **P < 0.01, and ***P < 0.001).

Fig. 3 .Fig. 4 .
Fig. 3. Temporal dynamics of farnesol accumulation inΔΔswi4, ΔΔahr1, ΔΔrap1, and ΔΔnrg1.Farnesol accumulation was accessed at 12, 18, 24, 36, 48, and  80  h post inoculation at 30°C in YPD for each TR mutant in 2 independent growth curves (n = 2) and 3 independent growth curves for SN152 + (n = 3).Dry weights were accessed as each time point and are presented as the mean dry weight ± SEM a).Farnesol accumulation data represent the mean dry weight normalized farnesol accumulation value ± SEM in either the whole culture b) or the supernatant fraction d).To access overall farnesol production, AUC analyses were performed and data are the mean AUC ± SE for the whole culture c) or supernatant fraction e).Differences in the AUC between groups were accessed by one-way ANOVA with Dunnett's multiple comparisons test.Differences were considered significant at P < 0.05 (*P < 0.05 and ***P < 0.001).

Table 1 .
Summary of TR mutants with statistically significant differences in farnesol accumulation in YPD media at 30°C.
a Mutant S.