Transcriptional Circuits Regulating Developmental Processes in Candida albicans

Introduction

C. albicans is a common human commensal that asymptomatically colonizes the skin, oral cavity, and gastrointestinal and genitourinary tracts of healthy individuals (Kennedy and Volz, 1985; Kumamoto, 2002; Achkar and Fries, 2010; Spiliopoulou et al., 2010; Kumamoto, 2011; Nobile and Johnson, 2015; Kan et al., 2020). It is also an opportunistic pathogen that is capable of causing superficial mucosal and life-threatening disseminated infections, especially in immunocompromised individuals (Wenzel, 1995; Calderone and Fonzi, 2001; Hube, 2004; Pappas et al., 2004; Mayer et al., 2013), such as in AIDS, chemotherapy and organ transplant patients, as well as in individuals with implanted medical devices (Wenzel, 1995; Nobile and Johnson, 2015). Multiple regulatory pathways controlling important C. albicans developmental processes allow this opportunistic fungal pathogen to adapt to and proliferate in distinct environmental niches in the host. In this review, we discuss the “core” transcriptional circuits controlling three major developmental processes in C. albicans: biofilm formation, the white-opaque phenotypic switch, and the commensal-pathogen transition. The core circuitry is defined as the direct physical interactions between transcriptional regulators that control these developmental processes and their respective upstream intergenic regions, where at least one direct binding interaction with other members of the circuit has been experimentally observed. These three circuits were chosen because they regulate persistent phenotypic changes in C. albicans that have been characterized using genome-wide transcriptional profiling (RNA-sequencing and/or microarray) and binding (chromatin immunoprecipitation) approaches. In our discussion of these circuits we focus largely on transcription factors (TFs) that bind to DNA in a sequence-specific manner; however, we also include some discussion of important cofactors for which genome-wide transcriptional profiling and binding data are available. In addition, we include information on “auxiliary” transcriptional regulators of these three developmental processes that we define as those that are known to regulate these processes, but that lack direct binding interactions with the core transcriptional regulators or binding data is not available for these transcriptional regulators under the growth condition of interest.

Regulation of Biofilm Formation

Biofilms are communities of adherent microbial cells encased in protective extracellular matrices (Kolter and Greenberg, 2006; Nobile and Johnson, 2015; Gulati and Nobile, 2016). Biofilms are ubiquitous in nature and are typically associated with interfaces, such as solid-liquid, liquid-gas, and liquid-liquid interfaces (Davey and O’toole, 2000; Kolter and Greenberg, 2006; Wilking et al., 2011; Desai and Ardekani, 2020). They are problematic when they form in industrial settings, such as in water distribution systems and on food preparation settings, and even more so when they form inside a host on tissues and on implanted medical devices. C. albicans biofilms are composed of several cell types, including round budding yeast-form cells, oval pseudohyphal cells, and elongated hyphal cells, encased in a protective extracellular matrix (Chandra et al., 2001; Desai and Mitchell, 2015). C. albicans biofilm formation occurs in four basic temporal stages: i) adherence of yeast-form cells to a surface; ii) growth and proliferation of yeast-form cells forming a basal layer of anchoring cells; iii) differentiation of a proportion of yeast-form cells into hyphal cells and production of the extracellular matrix; and iv) dispersion of yeast-form cells out of the biofilm to cause bloodstream infections or to colonize new sites for biofilm formation (Figure 1) (Desai and Mitchell, 2015; Nobile and Johnson, 2015; Gulati and Nobile, 2016). Indeed, C. albicans is a common cause of bloodstream infections worldwide, which often originate from biofilms (Edmond et al., 1999; Richards et al., 1999; Pfaller and Diekema, 2007). Given that cells within C. albicans biofilms are inherently resistant and tolerant to most antifungal drug treatments compared to planktonic (free-floating) cells, biofilm infections are particularly challenging to treat in the clinic. Understanding the genetic regulatory mechanisms that control C. albicans biofilm formation could lead to the development of novel therapeutic strategies effective in treating biofilm infections.

Figure 1 Stages of C. albicans biofilm formation. C. albicans biofilm formation occurs in four basic temporal stages: 1) adherence of yeast-form cells to a surface; 2) growth and proliferation of yeast-form cells forming a basal layer of anchoring cells; 3) differentiation of a proportion of yeast-form cells into hyphal cells and production of the extracellular matrix; and 4) dispersion of yeast-form cells out of the biofilm to cause bloodstream infections or to colonize new sites for biofilm formation.

The C. albicans transcriptional network controlling biofilm formation was first described eight years ago (Nobile et al., 2012). Six “master” biofilm transcriptional regulators (Bcr1, Tec1, Efg1, Ndt80, Rob1, and Brg1) were identified by screening a library of 165 transcription factor (TF) mutant strains (Homann et al., 2009) for defects in biofilm formation under standard in vitro biofilm growth conditions (Nobile et al., 2012). Here, we define a master biofilm transcriptional regulator as one whose deletion impairs biofilm formation throughout a 48-h period of biofilm growth under these standard conditions. All six TF mutant strains identified additionally had clear defects in biofilm formation in at least one of two in vivo animal models for biofilm formation (Nobile et al., 2012). Using genome-wide transcriptional profiling and chromatin immunoprecipitation techniques to study mature 48-h biofilms, a complex interconnected transcriptional network was discovered consisting of those six master transcriptional regulators, along with 1,061 downstream “target” genes (Nobile et al., 2012). These six master transcriptional regulators directly bound to the upstream intergenic regions and positively regulated the expression of each other, forming a tightly knit core biofilm circuit (Fox and Nobile, 2012; Nobile et al., 2012). Additionally, with the exception of Tec1, all of the six master biofilm transcriptional regulators acted as both repressors and activators of their directly bound biofilm target genes; Tec1, on the other hand, primarily acted as an activator (Nobile et al., 2012). Each of the six master biofilm transcriptional regulators controlled target genes that were in common with the other core transcriptional regulators in the circuit, as well as target genes that were unique to each transcriptional regulator. These findings suggest that each master biofilm transcriptional regulator in the circuit controls certain elements of biofilm formation independently, but that they also work together to coordinate concerted efforts important for biofilm formation. For example, Ndt80 regulates the expression of drug transporters independent of the other master biofilm transcriptional regulators in the circuit (such as, CDR4), and some in common with several of the other master biofilm transcriptional regulators in the circuit (such as, CDR3) (Nobile et al., 2012). Additionally, each master biofilm transcriptional regulator likely responds to unique environmental inputs, such as oxygen and nutrient availability, pH, temperature, and waste products. How different environmental inputs influence the biofilm transcriptional circuit is an intriguing area of future research. For example, we know that the six master biofilm transcriptional regulators discovered using in vitro biofilm assays are still required for in vivo biofilm formation in at least one of two in vivo biofilm models (Nobile et al., 2012). The majority (four) of the master biofilm transcriptional regulators discovered in this study were essential for biofilm formation in both in vivo biofilm models used; however, two of the master biofilm transcriptional regulators played different roles depending on the in vivo biofilm model (Nobile et al., 2012). Specifically, Bcr1 was essential for biofilm formation in a rat catheter biofilm model but was dispensable in a rat denture biofilm model (Nobile et al., 2012). Similarly, Brg1 was essential for biofilm formation in a rat denture biofilm model but was dispensable in a rat catheter biofilm model (Nobile et al., 2012). Future work on these master transcriptional regulators will determine their unique influences on biofilm formation dependent on the environmental inputs present.

In a subsequent study, three additional transcriptional regulators, Gal4, Rfx2, and Flo8, were added to the core biofilm transcriptional circuit (Fox et al., 2015). Gal4, Rfx2, and Flo8 were found to directly bind to the upstream intergenic regions of one or more of the previously identified six master biofilm transcriptional regulators and vice versa during biofilm development (Nobile et al., 2012; Fox et al., 2015). Gal4, Rfx2, and Flo8 were identified (in addition to the six previously identified transcriptional regulators) by screening a TF mutant library containing 192 TF mutant strains (Fox et al., 2015). This TF library contained the same 165 TF mutants (Homann et al., 2009) from the Nobile et al., 2012 study (Nobile et al., 2012) plus 27 additional newly constructed TF mutant strains. The TF mutants in this larger library were screened for their abilities to form biofilms over time at 90 min, 8, 24, and 48 h of biofilm growth (Fox et al., 2015). Flo8, like the other six previously identified master biofilm transcriptional regulators, was required for biofilm formation throughout a 48-h course of biofilm growth, and thus was deemed to be a master biofilm transcriptional regulator; Gal4 and Rfx2 were only required for normal biofilm formation at specific intermediate time points (Fox et al., 2015). Given that the initial biofilm circuit consisting of six master transcriptional regulators was discovered by assessing biofilm formation at a single mature time point (48 h) (Nobile et al., 2012), performing the genetic screen as a biofilm develops over time, with the additional TF mutant strains, contributed to the expansion of the core biofilm circuit (Fox et al., 2015). Genome-wide binding data was not performed for Gal4, Rfx2, and Flo8 as part of this study; however, directed chromatin immunoprecipitation followed by quantitative PCR was performed to determine that these three new transcriptional regulators are integrated into the core biofilm circuit, which now consists of nine core transcriptional regulators, seven of which are considered to be master biofilm transcriptional regulators (Figure 2) (Nobile et al., 2012; Fox et al., 2015). We note that although genome-wide binding experiments have been performed for Gal4 and Flo8 (Askew et al., 2009; Polvi et al., 2019), these experiments were not performed under biofilm conditions and thus the resulting data cannot be integrated into the biofilm transcriptional circuit. Overall, although the logic of the biofilm transcriptional circuit (defined as how each transcriptional regulator contributes to the regulatory dynamics of the circuit) has yet to be fully elucidated, the high degree of interconnectivity between the core biofilm transcriptional regulators likely contributes to the robustness, yet reversibility, of the biofilm state.

Figure 2 Transcriptional circuit controlling C. albicans biofilm formation. Ovals indicate each of the core biofilm transcriptional regulators with their respective names. Arrows indicate direct binding events. See Data Sheet S1, Tab4 for binding interactions. Data were derived from (Nobile et al., 2012; Fox et al., 2015). Figure was generated using Cytoscape (Shannon et al., 2003).

Although the nine core biofilm transcriptional regulators are known to be important for biofilm formation, how each one specifically contributes to biofilm processes (e.g. adhesion, filamentation, antifungal drug resistance, etc.), through detailed analyses of their mutant strains, has not been systematically determined. Table 1 summarizes the current knowledge of the roles of all known transcriptional regulators in known biofilm-related processes. Eight of the nine core biofilm transcriptional regulators (Bcr1, Brg1, Efg1, Flo8, Ndt80, Rfx2, Rob1, and Tec1) have been implicated in regulating filamentation (Schweizer et al., 2000; Bockmüh and Ernst, 2001; Cao et al., 2006; Elson et al., 2009; Hao et al., 2009; Sellam et al., 2010; Vandeputte et al., 2011; Du et al., 2012b; Nobile et al., 2012), which is a critical process necessary for maintaining the architectural stability of the biofilm structure. Four of the nine core biofilm transcriptional regulators (Bcr1, Efg1, Rfx2, and Tec1) have been implicated in regulating adhesion (Dieterich et al., 2002; Hao et al., 2009; Sahni et al., 2010; Finkel et al., 2012), including both cell-cell and cell-substrate adhesion, which is an essential process for both the initiation of biofilm formation as well as for the maintenance of a mature biofilm. Three of the nine core biofilm transcriptional regulators (Bcr1, Efg1, and Ndt80) are known to be involved in the regulation of antifungal drug resistance and/or tolerance (Chen et al., 2004; Sellam et al., 2009; Prasad et al., 2010; Desai et al., 2013), an important feature that contributes to the overall recalcitrance of established biofilms to antimicrobial compounds. Of the nine core biofilm transcriptional regulators, we know the least about the biofilm specific roles of Gal4, and only that it contributes to the structure of a biofilm at intermediate stages of biofilm development (Fox et al., 2015). In the future, additional roles of the nine core biofilm transcriptional regulators during biofilm formation will certainly be elucidated. For example, it seems likely that some of the core biofilm transcriptional regulators would be involved in the formation of the extracellular matrix; however, this role has not been examined to date in the mutant strains of the core biofilm transcriptional regulators. In addition, the ability of cells within biofilms to communicate with one another, called quorum sensing, is an important process for coordinating biofilm formation of many microorganisms; however, this role has yet to be examined in the mutant strains of the core biofilm transcriptional regulators. In fact, little is known in general on the regulation of quorum sensing during C. albicans biofilm development.

Table 1 Known transcriptional regulators with roles in C. albicans biofilm formation.

In addition to these nine transcriptional regulators that make up the core biofilm circuit, there are 50 “auxiliary” transcriptional regulators that have been implicated in biofilm formation (Table 1). The majority of these auxiliary biofilm transcriptional regulators are also bound in their upstream intergenic regions by at least one of the initial six master biofilm transcriptional regulators (Bcr1, Tec1, Efg1, Ndt80, Rob1, or Brg1; note that of the nine core biofilm transcriptional regulators, there is not genome-wide chromatin immunoprecipitation data available for Gal4, Rfx2, and Flo8, and thus we do not know whether they bind to the auxiliary biofilm transcriptional regulators) (Table 1) (Nobile et al., 2012). As such, several of the 50 auxiliary transcriptional regulators are integrated into the larger biofilm network that includes the core nine transcriptional regulators and all of their directly bound target genes (Nobile et al., 2012). Based on existing phenotypic analyses of the mutant strains of the auxiliary biofilm transcriptional regulators, the majority (48) are implicated in the regulation of adhesion and/or filamentation (Brown et al., 1999; Kadosh and Johnson, 2001; Cheng et al., 2003; Uhl et al., 2003; García-Sánchez et al., 2004; Kelly et al., 2004; Mulhern et al., 2006; Kim et al., 2008b; Shen et al., 2008; Wheeler et al., 2008; Homann et al., 2009; Nobile et al., 2009; Pukkila-Worley et al., 2009; Uppuluri et al., 2010a; Uppuluri et al., 2010b; Askew et al., 2011; Bonhomme et al., 2011; Ganguly et al., 2011; Finkel et al., 2012; Kamthan et al., 2012; Langford et al., 2013; Delgado-Silva et al., 2014; Tsai et al., 2014; Chen and Lan, 2015; Fox et al., 2015; Ghosh et al., 2015; Kakade et al., 2016; Kakade et al., 2019; Böttcher et al., 2020; Lagree et al., 2020; Omran et al., 2020; Wang et al., 2020); 16 are implicated in drug resistance and/or tolerance (Bruno et al., 2006; Cornet et al., 2006; Mulhern et al., 2006; Xu et al., 2007; Dunkel et al., 2008; Wheeler et al., 2008; Homann et al., 2009; Prasad et al., 2010; Nett et al., 2011; Vandeputte et al., 2012; Langford et al., 2013; Vasicek et al., 2014); two are implicated in the production of the extracellular matrix (Finkel et al., 2012; Delgado-Silva et al., 2014); and two are implicated in dispersion (Uppuluri et al., 2010b; Uppuluri et al., 2010a). Similar to the core biofilm transcriptional regulators, detailed analyses of the mutant strains of the auxiliary biofilm transcriptional regulators have not been systemically studied for known biofilm processes. Rather, most of their roles in biofilm processes have been determined through large-scale genetic screens. Of the auxiliary biofilm transcriptional regulators, we understand the least about the biofilm specific roles of Bpr1/Orf19.6874, which is only known to contribute to biofilm biomass throughout biofilm development (Fox et al., 2015). Future detailed phenotypic analyses of the auxiliary transcriptional regulator mutant strains in biofilm specific processes will certainly reveal new and additional roles for these transcriptional regulators in biofilm development.

Regulation of the White-Opaque Phenotypic Switch

The white-opaque switch in C. albicans is a form of phenotypic switching that gives rise to two distinct cell types called “white” and “opaque” that display distinct phenotypic characteristics at the single cell and colony levels (Anderson and Soll, 1987; Slutsky et al., 1987; Rikkerink et al., 1988; Bergen et al., 1990; Soll, 1992; Soll et al., 1993). White cells represent the standard budding yeast form of C. albicans, forming shiny, white, dome-shaped colonies on solid media plates, while opaque cells are larger and more elongated than white cells and form dull, off-white, flattened colonies on solid media plates (Slutsky et al., 1987; Soll et al., 1993; Lohse and Johnson, 2009; Noble et al., 2017). White and opaque cells differ in their virulence characteristics, metabolic preferences, mating competencies, interactions with the host innate immune system, and responses to environmental stimuli (Kolotila and Diamond, 1990; Lan et al., 2002; Lockhart et al., 2002; Miller and Johnson, 2002; Bennett et al., 2003; Geiger et al., 2004; Dumitru et al., 2007; Lohse and Johnson, 2008; Ramírez-Zavala et al., 2008; Huang et al., 2009; Huang et al., 2010; Lohse et al., 2013; Xie et al., 2013; Lohse et al., 2016; Du and Huang, 2016; Ene et al., 2016; Dalal et al., 2019). In total, nearly 20% of the transcriptome is differentially expressed, by at least twofold, between the two cell types, highlighting that the white-opaque switch involves major transcriptional rewiring (Tuch et al., 2010; Hernday et al., 2013). Under standard switch permissive growth conditions, switching between the white cell type, considered the “ground” state, and the opaque cell type, considered the “excited” state, occurs stochastically at a frequency of roughly one switch event per 1,000-10,000 cell divisions (Rikkerink et al., 1988; Bergen et al., 1990; Ramírez-Zavala et al., 2008; Alby and Bennett, 2009b). Each cell type is heritably maintained without any change to the primary sequence of the genome, thus fitting the classic definition of an epigenetic switch (Slutsky et al., 1987; Soll et al., 1993; Zordan et al., 2006; Zordan et al., 2007). The switch is responsive to the combined effects of environmental signals, such as carbon source, pH, CO₂ levels, and temperature, which can differentially bias the cell population towards one of the two cell types (Dumitru et al., 2007; Ramírez-Zavala et al., 2008; Alby and Bennett, 2009a; Huang et al., 2009; Huang, 2012; Lohse et al., 2013; Du and Huang, 2016; Ene et al., 2016; Dalal et al., 2019). Mating type can also influence the ability of the cells to undergo white-opaque switching, where MTL heterozygous (a/α) cells are typically “locked” in the white state, while MTL hemizygous (a/Δ, α/Δ), homozygous (a/a, or α/α), and haploid (a or α) cells are capable of undergoing stochastic white-opaque switching (Hull and Johnson, 1999; Lockhart et al., 2002; Miller and Johnson, 2002). This mating type dependency, however, is not exclusive to all strains; in fact, a significant fraction of MTL heterozygous clinical isolates can be induced to form opaque cells under specific growth conditions that promote white to opaque switching in MTL hemizygous, homozygous, or haploid cells (Xie et al., 2013).

Through a combination of forward and reverse genetic approaches, a total of 112 transcriptional regulators and one protein binding cofactor (Ssn6) have been identified which, when deleted, significantly impact the frequency of white-opaque switching (Table 2) (Huang et al., 2006; Srikantha et al., 2006; Zordan et al., 2006; Zordan et al., 2007; Hernday et al., 2013; Lohse et al., 2013; Du et al., 2015; Hernday et al., 2016; Lohse et al., 2016; Lohse and Johnson, 2016). Of these 113 switch regulating proteins, eight (Wor1, Wor2, Wor3, Wor4, Czf1, Efg1, Ahr1, and Ssn6) are considered to be core switch regulators, and have been extensively characterized by genome-wide transcriptional profiling and chromatin immunoprecipitation approaches in white and opaque cell types; the remaining 105 switch regulating proteins are considered to be auxiliary switch regulators (Table 2) (Sonneborn et al., 1999; Srikantha et al., 2000; Huang et al., 2006; Zordan et al., 2006; Zordan et al., 2007; Vinces and Kumamoto, 2007; Lohse and Johnson, 2010; Wang et al., 2011; Hernday et al., 2013; Lohse et al., 2013; Lohse and Johnson, 2016; Hernday et al., 2016). Together, these eight core switch regulators form complex cell type specific networks, with 203 bound target genes in white cells and 756 bound target genes in opaque cells (Hernday et al., 2013; Lohse et al., 2013; Hernday et al., 2016; Lohse and Johnson, 2016). At the center of the white and opaque specific regulatory networks are two distinct transcriptional circuits (see Figure 3A for the white circuit, Figure 3B for the opaque circuit, and Figure 3C for the combined white and opaque overlayed circuits) that consist of interconnected positive and negative feedback loops that govern the cell fate and heritable maintenance of the white and opaque cell types (Vinces et al., 2006; Vinces and Kumamoto, 2007; Zordan et al., 2007; Hernday et al., 2013; Hernday et al., 2016; Lohse and Johnson, 2016). Although several groups have identified kinases, chromatin modifiers, and other proteins that also affect white-opaque switching (Hnisz et al., 2009; Noble et al., 2017; Rai et al., 2018); here, we focus on the eight core switch regulators (TFs: Wor1, Wor2, Wor3, Wor4, Czf1, Efg1, Ahr1; and cofactor: Ssn6) for which genome-wide transcriptional profiling and chromatin immunoprecipitation data are available.

Table 2 Known transcriptional regulators and a protein cofactor with roles in the C. albicans white-opaque switch^‡.

Figure 3 Transcriptional circuits controlling the C. albicans white-opaque phenotypic switch. (A) Transcriptional circuit of the white state. (B) Transcriptional circuit of the opaque state. (C) Overlayed transcriptional circuits regulating the white and opaque states. Ovals indicate each of the core regulators with their respective names. Arrows indicate direct binding events. See Data Sheet S1, Tab4 for binding interactions. Data were derived from (Zordan et al., 2007; Hernday et al., 2013; Lohse et al., 2013; Hernday et al., 2016; Lohse and Johnson, 2016). Figure was generated using Cytoscape (Shannon et al., 2003).

Wor1 is considered to be the master regulator of the white-opaque switch, as it is the only switch regulator that is known to be required for both the transition to, and heritable maintenance of, the opaque cell type (Huang et al., 2006; Srikantha et al., 2006; Zordan et al., 2006; Zordan et al., 2007; Hernday et al., 2013; Lohse et al., 2013; Hernday et al., 2016; Lohse and Johnson, 2016). Furthermore, ectopic WOR1 expression can rescue opaque cell formation in all known mutant backgrounds that fail to spontaneously switch to the opaque cell type (Zordan et al., 2007; Du et al., 2012a; Lohse and Johnson, 2016). WOR1 expression is repressed in white cells, where Wor1 protein levels have been found to be nearly undetectable (Huang et al., 2006; Srikantha et al., 2006; Zordan et al., 2006; Zordan et al., 2007; Lohse and Johnson, 2010). In opaque cells, WOR1 is highly transcribed, and Wor1 protein levels have been found to accumulate to elevated levels (Huang et al., 2006; Srikantha et al., 2006; Zordan et al., 2006; Zordan et al., 2007; Lohse and Johnson, 2010). Stochastic white to opaque switching is thought to be the result of transcriptional noise within the white cell circuit that occasionally allows Wor1 levels to surpass a critical threshold necessary to induce the transition to the opaque state  (Srikantha et al., 2006; Hernday et al., 2010; Lohse and Johnson, 2010; Nobile et al., 2012; Guan and Liu, 2015; Horwitz et al., 2015; Lohse and Johnson, 2016; Lohse et al., 2016a; Tandonnet and Torres, 2017). Once established, the excited opaque cell circuit is stably maintained by a series of nested feedback loops, including a positive autoregulatory feedback loop generated by Wor1 binding to the upstream intergenic region of WOR1 (Zordan et al., 2007; Hernday et al., 2013). This Wor1-induced positive feedback loop, along with other opaque specific binding interactions between the white and opaque regulators and their respective upstream intergenic regions, is proposed to be a central mechanism that mediates the epigenetic heritability of the opaque cell type (Huang et al., 2006; Srikantha et al., 2006; Zordan et al., 2006; Zordan et al., 2007; Lohse and Johnson, 2010; Wang et al., 2011; Hernday et al., 2013; Lohse et al., 2013; Lohse and Johnson, 2016; Hernday et al., 2016). Stochastic opaque to white switching is believed to occur when transcriptional noise causes Wor1 levels to drop below a critical threshold, thus leading to a collapse of the excited opaque cell transcriptional program and a return to the ground white cell transcriptional program (Srikantha et al., 2006; Zordan et al., 2006; Lohse and Johnson, 2010).

The core transcriptional circuit in white cells consists of a series of feed-forward loops that ultimately repress the expression of WOR1 and WOR2, both of which are key players in the establishment and/or maintenance of the opaque cell type (Zordan et al., 2007). Efg1, Ahr1, and Ssn6 all contribute to the stability of the white cell circuit and are believed to directly or indirectly repress the expression of WOR1 and WOR2 (Zordan et al., 2007; Tuch et al., 2010; Hernday et al., 2013; Hernday et al., 2016). Deletion of EFG1, AHR1, or SSN6 destabilizes the white cell circuit such that most, if not all, of the cells in the population transition to the opaque state (Sonneborn et al., 1999; Srikantha et al., 2000; Vinces et al., 2006; Vinces and Kumamoto, 2007; Zordan et al., 2007; Wang et al., 2011; Hernday et al., 2016). Czf1, Wor3, and Wor4 are capable of destabilizing the white cell circuit, and induced expression of CZF1, WOR3, or WOR4 in white cells can promote white to opaque switching in a Wor1 dependent manner (Zordan et al., 2007; Hernday et al., 2013; Lohse et al., 2013; Lohse and Johnson, 2016). Interestingly, neither Czf1 nor Wor3 is required for the heritable maintenance of the opaque state once switching has occurred (Zordan et al., 2007; Lohse et al., 2013). Based on these results and the structure of the white cell regulatory circuit (Figure 3A), Czf1 and Wor4 are thought to destabilize the white cell type by directly and indirectly antagonizing the white cell stabilizing activities of Ssn6, Ahr1, and Efg1, and by inducing opaque promoting factors such as WOR3, thus introducing the transcriptional noise that leads to the stochastic activation of the WOR1 positive feedback loop and the transition to the opaque state. In addition to repression of WOR1 and WOR2, the white cell transcriptional program results in repression of opaque enriched transcripts (e.g. WOR3 and CZF1) as well as the activation of white enriched transcripts (e.g. EFG1), thus creating a series of feed-forward loops that act to stabilize the white cell circuit and prevent activation of the opaque state (Zordan et al., 2007; Hernday et al., 2013; Lohse et al., 2013).

In contrast to the core transcriptional circuit of the white cell type (Figure 3A), the core transcriptional circuit of the opaque cell type is extensively intertwined (Figure 3B). All of the core switch regulators are active in opaque cells, and they are each found to bind to their own upstream intergenic regions, along with the upstream intergenic regions of most, if not all, of the other core switch regulators (Figure 3B) (Huang et al., 2006; Srikantha et al., 2006; Zordan et al., 2006; Zordan et al., 2007; Wang et al., 2011; Hernday et al., 2013; Lohse et al., 2013; Hernday et al., 2016; Lohse and Johnson, 2016). To highlight this point, 58 of the 64 possible binding interactions between the core switch regulators and their respective upstream intergenic regions are observed in opaque cells (Data Sheet S1, Tab1). Although the logic of the opaque transcriptional circuit has yet to be fully elucidated, the high degree of interconnectivity between the core opaque regulators likely contributes to the robustness, yet reversibility, of the opaque cell state. Similar to the white cell circuit, Wor1 is a critical player in the opaque cell circuit; however, it is the sustained high levels of WOR1 expression, rather than its repression, that is required for the formation and stable maintenance of the opaque cell type (Huang et al., 2006; Srikantha et al., 2006; Zordan et al., 2006). Although not strictly required for the formation of an opaque cell, Wor2 and Wor4 also play important roles in the heritable maintenance of the opaque transcriptional program (Zordan et al., 2007; Lohse and Johnson, 2016). Strains lacking WOR2 or WOR4 are locked in the white cell type and fail to undergo spontaneous white to opaque switching, yet can be induced to form opaque cells by ectopic expression of WOR1 (Frazer et al., 2020). These induced opaque cells, however, are unstable, and quickly revert to the white cell type when ectopic WOR1 expression is repressed, indicating that Wor2 and Wor4 play essential roles in the heritability of opaque cells (Zordan et al., 2007). Interestingly, with the exception of Ahr1, all switch regulators discovered to date have been found to contain prion-like domains that enable liquid-liquid demixing and the formation of phase-separated condensates (Frazer et al., 2020). Several of the switch regulators, including Wor1 and Wor4, have been shown to undergo phase separation in vitro, and to form condensates at genomic loci in vivo, in a manner similar to the formation of mammalian super-enhancers (Frazer et al., 2020). Combined with the observation that many of the target genes bound by the switch regulators are flanked by unusually large upstream intergenic regions (Zordan et al., 2007; Hernday et al., 2013), and the discovery that specific residues within the Wor1 prion-like domain are required for condensate formation and white to opaque switching, it seems likely that these phase-separated condensates formed by the core switch regulators in opaque cells are critical factors that contribute to the formation and heritable maintenance of the opaque cell type.

Regulation of the Commensal-Pathogen Transition

C. albicans typically exists as a commensal member of the healthy human microbiota. It can also transition into a pathogen in response to specific host environmental cues. In its pathogenic state, C. albicans can cause a wide range of infections, from acute to chronic superficial mucosal infections to severe and life-threatening disseminated bloodstream infections (Wenzel, 1995; Hube, 2004; Pappas et al., 2004). Although immunocompetent individuals with healthy and balanced microbiota are typically not adversely affected by C. albicans, immunocompromised individuals can suffer severe infections with significant morbidity and mortality (Wenzel, 1995; Nobile and Johnson, 2015). Understanding the genetic regulatory mechanisms that control the C. albicans commensal-pathogen transition has the potential to lead to the development of targeted therapeutic strategies against C. albicans in its pathogenic state, without affecting its commensal state and the delicate balance of the microbiota.

Two distinct C. albicans transcriptional networks controlling the commensal-pathogen transition were described in 2011 and 2013, one governing iron homeostasis, and the other governing proliferation in the host, respectively (see Figure 4A for the iron homeostasis circuit, Figure 4B for the proliferation in the host circuit, and Figure 4C for the combined commensal-pathogen overlayed circuits) (Chen et al., 2011; Pérez et al., 2013). As a commensal of the gastrointestinal (GI) tract, C. albicans is exposed to varying and often abundant levels of iron from food, and thus a tightly regulated transcriptional response is important for C. albicans to control iron assimilation and to avoid iron toxicity in the GI tract (McCance and Widdowson, 1938; Martin et al., 1987; Miret et al., 2003). On the other hand, when C. albicans causes a disseminated bloodstream infection, iron is extremely limiting, and to survive, C. albicans must conserve and scavenge iron from the bloodstream. Three transcriptional regulators, Sef1, Sfu1, and Hap43, were found to form a tightly knit transcriptional network, encompassing 214 downstream target genes (Chen et al., 2011). These three transcriptional regulators control iron homeostasis and were found to be essential for C. albicans to survive as both a commensal and as a pathogen within the mammalian host (Chen et al., 2011). Iron homeostasis in many other fungi (such as in other ascomycetes and the basidiomycete, Cryptococcus neoformans) is commonly regulated by a bipartite regulatory circuit composed of orthologs of Sfu1 and Hap43, where Sfu1 orthologs repress iron acquisition genes and HAP43 orthologs, while Hap43 orthologs repress nonessential iron utilization genes and SFU1 orthologs. This mutually repressive regulatory interaction between orthologs of Sfu1 and Hap43 in other fungi is significantly altered in C. albicans by the intercalation of Sef1 as a third player within this circuit (Figure 4A) (Chen et al., 2011). In C. albicans, Sfu1 directly represses SEF1 and iron acquisition genes under iron replete conditions (Chen et al., 2011). In response to iron limitation, Sef1 serves to directly activate HAP43 and iron uptake genes, while Hap43 directly represses SFU1 and iron utilization genes (Chen et al., 2011). Although the roles for Hap43 in C. albicans are similar to those of other fungi, the reciprocal interaction between Sfu1 and HAP43 is altered in C. albicans by the inclusion of Sef1, which serves as an intermediary between Sfu1 and HAP43. C. albicans SEF1 and SFU1 are differentially expressed between growth in the GI tract versus growth in the bloodstream (Chen et al., 2011), thus providing dual inputs into the circuit controlling iron acquisition and utilization. While both Sef1 and Sfu1 serve to promote commensalism in a mouse GI commensal model, only Sef1 is required for virulence in a mouse disseminated infection model (Chen et al., 2011). Interestingly, deletion of SFU1 conferred a significant competitive advantage over wildtype cells in the disseminated infection model (Chen et al., 2011), indicating that Sfu1 serves not only to promote commensalism in the GI tract, but also to attenuate virulence in the bloodstream. (See Table 3 for information on these three core transcriptional regulators in the commensal-pathogen transition.) Ultimately the C. albicans iron homeostasis circuit produces a well conserved transcriptional output consisting of increased iron uptake and reduced iron utilization in iron limited environments, and decreased iron uptake and increased iron utilization in iron replete conditions. Despite being well conserved in its transcriptional output, the iron homeostasis circuit appears to be uniquely evolved in C. albicans to control the delicate balance between its commensal and pathogenic growth states.

Figure 4 Transcriptional circuits controlling the C. albicans commensal-pathogen transition. (A) Transcriptional circuit controlling iron homeostasis. (B) Transcriptional circuit controlling proliferation in the host. (C) Overlayed transcriptional circuits controlling the commensal-pathogen transition. Ovals indicate each of the core regulators with their respective names. Arrows indicate direct binding events. Note that since Rtg1 and Rtg3 function as a heterodimer, and do not appear to bind DNA independently, they are represented as a single node in these regulatory circuit diagrams. See Data Sheet S1, Tab4 for binding interactions. Data were derived from (Chen et al., 2011; Pérez et al., 2013). Figure was generated using Cytoscape (Shannon et al., 2003).

Table 3 Known transcriptional regulators with roles in the C. albicans commensal-pathogen transition.

A subsequent study identified eight transcriptional regulators (Tye7, Orf19.3625, Lys144, Zcf21, Lys14, Hsm1, Rtg1, and Rtg3) that influence C. albicans proliferation in the commensal and/or pathogenic growth states (Pérez et al., 2013). These regulators were identified by screening a subset of the commonly used C. albicans TF mutant library (Homann et al., 2009) for defects in a commensal (GI colonization) mouse model and a pathogenic (disseminated infection) mouse model. This subset of the TF mutant library consisted of those mutant strains that revealed no phenotypes in a diverse panel of in vitro growth conditions, and was screened to identify transcriptional regulators that were specifically required for normal (wildtype) levels of growth in either of the two mouse models (Homann et al., 2009; Pérez et al., 2013). Of the eight regulators that were identified, six (Rtg1, Rtg3, Tye7, Hms1, Orf19.3625, and Lys144) were required for GI colonization, while five (Rtg1, Rtg3, Hms1, Lys14, and Zcf21) were required for robust growth in the disseminated infection model (Pérez et al., 2013). Overall, Tye7, Orf19.3625, and Lys144 were found to be specific to commensal colonization of the GI tract; Zcf21 and Lys14 were found to be specific to disseminated infections; and Rtg1, Rtg3, and Hms1 were found to be associated generally with growth in the host (Pérez et al., 2013). Based on genome-wide transcriptional profiling and chromatin immunoprecipitation data, seven of these regulators (Tye7, Lys144, Zcf21, Lys14, Hsm1, Rtg1, and Rtg3) were found to form a transcriptional network consisting of 808 directly bound target genes. Significant overlap was observed between the bound target genes of this network and those genes that were upregulated in the mouse GI model compared to growth in vitro. Orf19.3625 was excluded from this analysis as it is a predicted subunit of a histone remodeling complex, and thus was not considered to be a specific regulator within the commensal-pathogen network. In contrast to the transcriptional network defined by Sef1, Sfu1, and Hap43, which is primarily responsible for regulating genes involved in iron homeostasis (Chen et al., 2011), the transcriptional network defined by Tye7, Lys144, Zcf21, Lys14, Hsm1, Rtg1, and Rtg3 appears to primarily regulate genes involved in the acquisition and metabolism of carbon and nitrogen, as well as genes that encode transporters and cell surface proteins (Pérez et al., 2013). The binding profiles for Rtg1 and Rtg3 were observed to be identical, and thus they likely function as a heterodimer to bind DNA (Pérez et al., 2013), which is consistent with their orthologs in Saccharomyces cerevisiae (Liu and Butow, 2006). Of the 153 direct target genes in this network that are upregulated during GI colonization and disseminated infection, 108 of them are bound by the Rtg1/3 heterodimer (Pérez et al., 2013), highlighting the central role that Rtg1/3 plays in this network. We note that a subsequent study by the same group identified five transcriptional regulators that influence fitness in an oropharyngeal candidiasis model (Cup9, Zcf8, Zcf21, Zcf27, and Orf19.217), and identified a set of genes that are differentially regulated in response to deletion of CUP9 (Meir et al., 2018). We did not include these data in our analyses since binding experiments that would be necessary to integrate these additional regulators into the commensal-pathogen transcriptional circuit have not been performed.

At the core of this commensal-pathogen transcriptional network lies a tightly interwoven regulatory circuit defined by the binding interactions between five of these transcriptional regulators (Hms1, Zcf21, Tye7, Rtg1, and Rtg3) and their respective upstream intergenic regions (Figure 4B). While Lys14 and Lys144 are clearly important for pathogenic and commensal growth, respectively, they are not integrated into the core transcriptional circuit and instead appear to function as auxiliary regulators. Interestingly, RTG1 and RTG3 are not regulated at the transcriptional level in response to growth in the GI tract and are not direct targets of any of the members of this commensal-pathogen transcriptional circuit (Pérez et al., 2013). Instead, Rtg1/3 seems to function as a major regulatory input into, rather than target of, this commensal-pathogen circuit. In S. cerevisiae, the Rtg1/3 heterodimer is known to be post-translationally modified and translocated into the nucleus in response to growth on poor nitrogen sources or mitochondrial dysfunction, suggesting that nitrogen assimilation and metabolic adaptation could be critical factors for the proliferation of C. albicans in the host (Liao and Butow, 1993; Jia et al., 1997; Liu and Butow, 2006). Hms1, which is also required for both commensal and pathogenic growth in the host, is known to be activated in response to elevated temperatures (Shapiro et al., 2012), indicating that temperature, along with nitrogen source(s), represent two critical environmental signals that influence the commensal and pathogenic growth programs of C. albicans. Zcf21 represses a variety of genes that encode key virulence factors, and plays a major role in pathogenesis by balancing the positive effects of these virulence factors during disseminated infection against the increased susceptibility to host immune system recognition and clearance that is correlated with their expression (Böhm et al., 2016). Finally, Tye7 has been implicated in the metabolism of carbohydrates, such as oligosaccharides and polysaccharides, as well as in the regulation of hyphal growth and biofilm formation (Askew et al., 2009; Bonhomme et al., 2011). (See Table 3 for information on these five core transcriptional regulators in the commensal-pathogen transition.) Although both the iron homeostasis and the host proliferation transcriptional networks are critical to the ability of C. albicans to grow as a commensal and as a pathogen, there is limited interconnectivity between these networks at the level of the core regulators of each circuit (Figure 4C). SFU1 serves as the sole point of integration between the two circuits, being bound by Rtg1/3 and Tye7. There are no binding interactions observed between the iron homeostasis regulators (Sef1, Sfu1, and Hap43) and the genes encoding the host proliferation regulators, suggesting that under certain growth conditions which alter the binding of Rtg1/3 and/or Tye7, the iron homeostasis circuit may function as a sub-circuit of the host proliferation circuit. Together, the transcriptional regulators involved in iron homeostasis and acquisition, and host proliferation, confer C. albicans with the ability to proliferate in different niches of the host as well as to transition between commensal and pathogenic states in response to changes in the host environment.

Integration of Networks

In total, the three larger regulatory networks, consisting of the core regulators and all of their directly bound target genes involved in biofilm formation, the white-opaque phenotypic switch, and the commensal-pathogen transition in C. albicans encompass at least 1657 directly bound individual target genes, making up a little over 25% of genes in the entire genome (note that Flo8, Gal4, and Rfx2 were excluded from this analysis since there is not genome-wide chromatin immunoprecipitation data available for them) (Data Sheet S1, Tab2) (Zordan et al., 2007; Chen et al., 2011; Nobile et al., 2012; Hernday et al., 2013; Lohse et al., 2013; Pérez et al., 2013; Fox et al., 2015; Hernday et al., 2016; Lohse and Johnson, 2016). These three networks are highly intertwined, with 40% (667/1657) of the target genes shared between at least two of the networks, and 11% (188/1657) of the target genes shared between all three networks (Data Sheet S1, Tab2). This high degree of interconnectivity is even more pronounced at the level of the core transcriptional circuits that control these three networks, as is evident by the extensive binding interactions present between the core regulators themselves (Figure 5 and Data Sheet S1, Tab1). Together, the twenty transcriptional regulators for which we have genome-wide chromatin immunoprecipitation data available form a total of 225 binding interactions within and between their core circuits, distributed roughly evenly between intra-circuit (49%) and inter-circuit (51%) interactions (note that the Rtg1/3 heterodimer is counted as a single regulator since neither subunit is known to bind independently) (Data Sheet S1, Tab3). The commensal-pathogen circuit and the biofilm circuit are highly intertwined with the regulators in the other circuits, with 66% and 59% inter-circuit interactions, respectively, while the opaque cell circuit appears to be much more isolated, with the majority (64%) of its interactions being intra-circuit (Data Sheet S1, Tab3). Perhaps the most striking example of integration between the circuits is exemplified by Ndt80 in the biofilm circuit, which binds to the upstream intergenic regions of 22 out of 24 of the core regulators (all but the upstream intergenic regions of RTG1 and RTG3) (Data Sheet S1, Tab1). The percentage of inter-circuit binding events is highest for Tye7 (79%), Zcf21 (75%), Bcr1 (71%), Brg1 (67%), and Rtg1/3 (67%), accounting for at least two out of three binding events for each of these regulators within the three core circuits (Data Sheet S1, Tab3). At the opposite end of the spectrum, Hap43, Hms1, and Sfu1 are exclusive to the commensal-pathogen circuit. In addition, at least two thirds of the binding events observed for Wor3 (88%), Czf1 (75%), Rob1 (71%), Ahr1 (70%), and Wor4 (70%) within the three core circuits occur within their respective core circuits (Data Sheet S1, Tab3). Interestingly, the degree of Efg1 inter-circuit interaction is unique to the circuit within which it lies, where 61% inter-circuit interactions are observed for Efg1 in the biofilm circuit, while only 42% inter-circuit interactions are observed for Efg1 in the white-opaque circuit (Data Sheet S1, Tab3). BRG1 is the most highly integrated target within the three circuits, where it is bound by seventeen of the twenty core regulators evaluated (leaving out Gal4, Rfx2, and Flo8, and considering Rtg1 and Rtg3 as a single regulator) (Data Sheet S1, Tab1). Overall, more than half (thirteen out of twenty-four) of the regulators that make up the three core circuits are bound by at least half (eleven or more) of the twenty core regulators evaluated (Data Sheet S1, Tab1). These rather striking numbers highlight the degree to which these circuits are intertwined, and these numbers are only likely to increase as additional core regulators are identified and incorporated into the three transcriptional circuits.

Figure 5 Integrated transcriptional circuits of C. albicans biofilm formation, the white-opaque switch and the commensal-pathogen transition. Ovals indicate each of the core regulators with their respective names. Arrows indicate direct binding events. See Data Sheet S1, Tab4 for binding interactions. Data were derived from (Zordan et al., 2007; Chen et al., 2011; Nobile et al., 2012; Hernday et al., 2013; Lohse et al., 2013; Pérez et al., 2013; Fox et al., 2015; Hernday et al., 2016; Lohse and Johnson, 2016). Figure was generated using Cytoscape (Shannon et al., 2003).

The extensive integration between these core transcriptional circuits appears to have significant functional relevance. For example, 14 of the 24 regulator genes discussed (AHR1, BCR1, BRG1, CZF1, GAL4, HAP43, HMS1, RFX2, SEF1, SFU1, TEC1, WOR1, WOR3, ZCF21) are differentially expressed by at least twofold between planktonic and biofilm growth conditions; of these fourteen genes, all but GAL4 are upregulated in biofilms (Data Sheet S1, Tab1) (Nobile et al., 2012). A similar trend is observed during white-opaque switching, where eleven of the twenty-four regulator genes (BRG1, CZF1, EFG1, GAL4, HMS1, RFX2, ROB1, TYE7, WOR1, WOR2, WOR3) are differentially expressed by at least twofold between white and opaque cell types (Data Sheet S1, Tab1) (Tuch et al., 2010). The interactions between the biofilm circuit and the white-opaque circuit are particularly striking. All eight of the core white-opaque regulator genes are bound by at least four of the six core biofilm regulators, and six of the eight white-opaque regulator genes (all but EFG1 and WOR4) are differentially expressed by twofold or more between planktonic and biofilm conditions (WOR1, AHR1, CZF1, and WOR3 are upregulated by 3-, 5-, 8-, and 32-fold, respectively, while WOR2 and SSN6 are both downregulated by 2-fold) (Data Sheet S1, Tab1). Conversely, five of the nine core biofilm regulator genes are bound by at least four of the eight white-opaque regulators in opaque cells (EFG1, BRG1, BCR1, TEC1, and RFX2 are bound by eight, eight, five, five, and four white-opaque regulators, respectively), and five of the nine biofilm regulator genes are differentially expressed by at least 2-fold between white and opaque cells (BRG1 and RFX2 are upregulated in opaque cells, while EFG1, GAL4, and ROB1 are upregulated in white cells) (Data Sheet S1, Tab1). The commensal-pathogen circuit regulators are closely intertwined with the biofilm circuit; however, there is relatively little overlap between the overlayed white-opaque circuit and the overlayed commensal-pathogen circuit. Six of the eight commensal-pathogen regulator genes (all but RTG1 and RTG3) are bound by at least one biofilm core regulator, half of which (SFU1, TYE7, and ZCF21) are bound by at least four of the biofilm regulators (Data Sheet S1, Tab1). All six of the commensal-pathogen regulator genes that are bound by biofilm regulators are differentially expressed by twofold or more between planktonic and biofilm conditions, with all but TYE7 being upregulated in biofilms (Data Sheet S1, Tab1). In contrast to the high degree of functional interaction between the biofilm circuit and the overlayed commensal-pathogen circuit, only three of the eight commensal-pathogen regulator genes (SFU1, TYE7, and ZCF21) are bound by any of the white-opaque regulators, and of the three target genes, only TYE7 is differentially expressed between white and opaque cells (upregulated twentyfold in opaque cells). The effect of growth under conditions relevant to the overlayed commensal-pathogen circuit (i.e. low iron or growth in the GI tract) is relatively limited when compared to the effects of biofilm formation and white-opaque switching. Upon growth in low iron, only the three regulator genes involved in iron homeostasis (HAP43, SEF1, SFU1) are differentially expressed (Data Sheet S1, Tab1) (Chen et al., 2011). While growth in the GI tract does affect the expression of core regulator genes in the other circuits, the impact of this expression is relatively limited, where AHR1 and TEC1 are upregulated and ROB1 is downregulated in the GI tract versus growth in vitro (Rosenbach et al., 2010).

Perspectives

The C. albicans transcriptional regulatory circuits controlling the developmental processes of biofilm formation, the white-opaque phenotypic switch, and the commensal-pathogen transition are individually tightly knit and we show that they are integrated together by extensive regulatory crosstalk between the core regulators that comprise each circuit. If we take into consideration all of the target genes in each of the larger transcriptional networks, each regulator controls individual subsets of target genes regulating distinct functions as well as subsets of target genes with functions in common with the other core regulators in each network. Strikingly, these three major transcriptional networks, together, encompass a little over 25% of genes in the entire genome, indicating that there is a high degree of functional redundancy across the networks. The complexity and functional redundancy of these network structures often make dissecting the logic of each network extremely challenging. The networks we discuss here in this review are overall structurally very similar to networks controlling complex transcriptional developmental processes in higher eukaryotes, such as the mammalian embryonic stem cell state (pluripotency) network (Boyer et al., 2005; Kim et al., 2008a). Given that mammals and C. albicans diverged from a common ancestor around 1.5 million years ago (Wang et al., 1999), it is notable that the structures of these independently evolved transcriptional networks are so similar. There are a couple hypotheses as to how these transcriptional networks could appear so structurally similar (Sorrells and Johnson, 2015). The first hypothesis is that these complex transcriptional networks represent the optimal solutions for organizing the biological processes they control (François and Hakim, 2004; Prill et al., 2005). The second hypothesis is that these transcriptional networks are not optimal solutions but are rather non-adaptive structures that have been retained over evolutionary time scales by purifying selection and are thus the result of high-probability evolutionary trajectories (Sorrells and Johnson, 2015). As we begin to discover and deconvolute complex transcriptional networks, we will begin to test these hypotheses and shed new light on the logic of these complex network structures.

Author Contributions

Conceptualization: DR, MQ, AH, and CN. Formal Analysis: DR, MQ, AH, and CN. Investigation: DR, MQ, AH, and CN. Resources: AH, and CN. Data Curation: DR, MQ, AH, and CN. Writing—Original Draft Preparation: DR, MQ, AH, and CN. Writing—Review and Editing: DR, MQ, AH, and CN. Visualization: DR, AH, and CN. Supervision: AH and CN. Project Administration: AH and CN. Funding Acquisition: AH and CN. All authors contributed to the article and approved the submitted version.

Funding

This work was supported by the National Institutes of Health (NIH) National Institute of Allergy and Infectious Diseases (NIAID) and National Institute of General Medical Sciences (NIGMS) awards R21AI125801 and R35GM124594, respectively, by a Pew Biomedical Scholar Award from the Pew Charitable Trusts, and by the Kamangar family in the form of an endowed chair to CN. This work was also supported by NIH NIAID award R15AI37975 to AH. DR was supported by the National Science Foundation (NSF) Graduate Research Fellowship Program (GRFP) award 1744620. The funders had no role in the study design, data collection and interpretation, or the decision to submit the work for publication.

Conflict of Interest

CN is a cofounder of BioSynesis, Inc., a company developing inhibitors and diagnostics of biofilms.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer AS declared a past co-authorship with one of the authors AH to the handling editor.

Acknowledgments

We thank all members of the Nobile and Hernday labs for insightful discussions on the topic of this manuscript.

Supplementary Material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcimb.2020.605711/full#supplementary-material

Supplementary Data Sheet 1 | Compilation and analysis of regulator binding interactions and target gene expression. Tab1 labeled “Tab1_Combined Core Circuits” contains compiled genome-wide chromatin immunoprecipitation (ChIP-seq or ChIP-chip) and expression profiling (RNA-seq or microarray) data for the core circuit regulators and their respective target genes. ChIP data and RNA-seq data values are in log₂ format. Biofilm regulators ChIP data and differential gene expression data were derived from (Nobile et al., 2012). White-opaque regulators ChIP data were derived from (Zordan et al., 2007; Hernday et al., 2013; Lohse et al., 2013; Hernday et al., 2016; Lohse and Johnson, 2016). White-opaque differential gene expression data were derived from (Tuch et al., 2010). Iron homeostasis regulators ChIP data and differential gene expression data were derived from (Chen et al., 2011). Host proliferation ChIP data were derived from (Pérez et al., 2013). Host proliferation differential gene expression data were derived from (Rosenbach et al., 2010). Tab2 labeled “Tab2_Combined Networks” contains compiled genome-wide chromatin immunoprecipitation (ChIP-seq or ChIP-chip) and expression profiling (RNAseq or microarray) data for the core circuit regulators and all possible target genes in the C. albicans genome. ChIP data and RNA-seq data values are in log₂ format. Biofilm regulators ChIP data and differential gene expression data were derived from (Nobile et al., 2012). White-opaque regulators ChIP data were derived from (Zordan et al., 2007; Hernday et al., 2013; Lohse et al., 2013; Hernday et al., 2016; Lohse and Johnson, 2016). White-opaque differential gene expression data were derived from (Tuch et al., 2010). Iron homeostasis regulators ChIP data and differential gene expression data were derived from (Chen et al., 2011). Host proliferation ChIP data were derived from (Pérez et al., 2013). Host proliferation differential gene expression data were derived from (Rosenbach et al., 2010). Tab3 labeled “Tab3_Inter- vs Intra-circuit” contains an analysis of the genome-wide ChIP data from Tab1, tabulating the total number of bound targets for each of the regulators for which genome-wide binding data is available, and calculating the percentage of binding events that represent inter- versus intra-circuit binding interactions. Tab4 labeled “Tab4_Cytoscape Interactions” contains a representation of the genome-wide ChIP data from Tab1 in an interaction table format for visualization in Cytoscape (Shannon et al., 2003). This dataset was used to generate all of the circuit diagrams shown in Figures 2–5.

References