Post-transcriptional gene regulation in the biology and virulence of Candida albicans.

In the human fungal pathogen Candida albicans, remodelling of gene expression drives host adaptation and virulence. Recent studies revealed that in addition to transcription, post-transcriptional mRNA control plays important roles in virulence-related pathways. Hyphal morphogenesis, biofilm formation, stress responses, antifungal drug susceptibility and virulence in animal models require post-transcriptional regulators. This includes RNA binding proteins that control mRNA localization, decay and translation, as well as the cytoplasmic mRNA decay pathway. Comprehensive understanding of how modulation of gene expression networks drives C. albicans virulence will necessitate integration of our knowledge on transcriptional and post-transcriptional mRNA control.

Introduction

The human commensal yeast Candida albicans causes oral and vaginal infections, and disseminated disease in severely ill hosts (Brown et al., 2012). Screens of transcription factor mutant libraries have revealed functions for gene expression networks in pathogenesis‐related biology of C. albicans (Nobile and Mitchell, 2005; Homann et al., 2009; Pukkila‐Worley et al., 2009; Finkel et al., 2012; Nobile et al., 2012; Perez et al., 2013). Further to transcription, proper spatio‐temporal expression of genes necessitates regulation of the subcellular localization, translation and turnover of mRNAs by RNA binding proteins and post‐transcriptional mechanisms. Studies in baker's yeast Saccharomyces cerevisiae and other eukaryotic models have illuminated the functions of multiple RNA binding proteins, and the importance of post‐transcriptional mRNA networks in dictating cellular physiology, for example (Hogan et al., 2008; Freeberg et al., 2013), reviewed in (Keene, 2007; Quenault et al., 2011; Blackinton and Keene, 2014). How post‐transcriptional regulation controls C. albicans biology and pathogenicity is a major knowledge gap. Excitingly, recent publications have focussed on this question, laying the foundation for understanding post‐transcriptional mRNA networks in this pathogen.

Post‐transcriptional regulation of hyphal morphogenesis

The yeast to hyphae transition is central to C. albicans virulence through functions including tissue invasion, cell adhesion, evasion of macrophages and development of clinically relevant biofilm communities. In response to stimuli such as temperature, nutrients or serum, signal transduction pathways and transcription factors induce the hyphal gene expression programme and filamentous growth (Sudbery, 2011). Significant evidence for post‐transcriptional mRNA regulation being important in hyphal growth of C. albicans is (Fig. 1).

Figure 1

Post‐transcriptional regulation in hyphal morphogenesis.

(a) mRNA decay factors required for hyphal morphogenesis. The pathway is shown as understood in S. cerevisiae. For simplicity, not all known protein–protein interactions are depicted, and the mRNA is shown linear instead of circular with 5′ and 3′ ends in proximity. The roles of the Dcp1/Dcp2 decapping enzyme in filamentation are yet to be studied.

(b) Possible mechanisms of 5′ UTR‐dependent repression of UME6 through secondary structure formation, or binding of RNA binding protein(s) (orange circle) to inhibit ribosome association or target to P‐bodies, as proposed by (Childers et al., 2014).

(c) mRNA trafficking to the hyphal tip by She3. The precise architecture of the C. albicans She3 complex is not known. The homolog of She2 is lacking, and the myosin is likely Myo2 (Elson et al., 2009). Whether C. albicans She3 interacts directly with mRNAs is unknown, but RNA binding activity can be predicted based on S. cerevisiae (Muller et al., 2011). Based on studies in S. cerevisiae, particularly ASH1 trafficking, mRNA localization elements can be in the open reading frame and 3′ UTR, and the She3 complex is multimeric (Shi et al., 2014). Therefore, our cartoon is a greatly simplified version of the likely protein‐RNA complex. En route, mRNAs is repressed and activated at the destined location. The mechanism of translational repression–derepression has not been studied in C. albicans.

Mutations in the cytoplasmic mRNA decay pathway impair hyphal morphogenesis

The major eukaryotic mRNA decay pathway consists of poly(A) tail degradation by the mRNA deadenylase Ccr4‐NOT, hydrolysis of the 5′ cap (decapping) and mRNA digestion by Xrn1/Kem1, reviewed in Goldstrohm and Wickens (2008) (Fig. 1a). RNA binding proteins modulate these processes by recruiting deadenylation and decapping factors, reviewed in Goldstrohm et al. (2008) and Quenault et al. (2011). C. albicans mutants in the deadenylase subunits CCR4 and POP2, the decapping activators DHH1 and EDC3 and the exonuclease XRN1/KEM1 are defective in hyphal morphogenesis (Richard et al., 2005; Dagley et al., 2011; Jung and Kim, 2014; Shively et al., 2015). Hyphal defects are not seen in all conditions. For example, ccr4 and pop2 can filament in liquid serum at 37°C, and xrn1 in the Galleria mellonella infection model (Fuchs et al., 2010; Dagley et al., 2011). This suggests the mRNA decay pathway is not necessary for hyphal growth per se, but rather acts in the response to hyphal signals. Consistent with this, in S. cerevisiae Dhh1, Xrn1 and Edc3, together with other post‐transcriptional regulators, are substrates of the kinases that orchestrate filamentous growth (Shively et al., 2015). It will be interesting to determine if in C. albicans mRNA decay factors are regulated post‐translationally by filamentous growth kinases.

It is unclear why the mRNA decay mutants display filamentation defects. Functions important for hyphal growth, such as cell wall biogenesis, nutrient responses and mitochondrial activity, are compromised in mRNA decay mutants in C. albicans and/or S. cerevisiae (Moriya and Isono, 1999; Kaeberlein and Guarente, 2002; Dagley et al., 2011; Braun et al., 2014), reviewed in Panepinto et al. (2013). Consistent with a metabolic reason, in the C. albicans xrn1 mutant protein levels for several metabolic enzymes are distinct from wild type cells, and overexpression of the pyruvate dehydrogenase subunit Lpd1 rescued hyphal defects (Lee et al., 2010). S. cerevisiae Dhh1 has been implicated in translational regulation of Ste12, a transcriptional activator of filamentation (Park et al., 2006). Whether this is conserved in C. albicans is unknown. It should be noted that a picture has emerged whereby regulators of mRNA decay play complex, cytoplasmic and nuclear roles in gene expression, and control and coordinate multiple steps such as transcription, post‐transcriptional mRNA regulation and translational control (Panasenko et al., 2006; Kruk et al., 2011; Haimovich et al., 2013; Sun et al., 2013). Therefore, future studies will need to decipher the precise molecular mechanisms by which mRNA decay factors regulate hyphal morphogenesis.

The expression of key transcriptional regulators of the yeast‐hyphae transition is regulated post‐transcriptionally

Evidence for this comes from studies of the hyphal transcriptional activator UME6 and the transcriptional repressor of the hyphal programme NRG1 (Cleary et al., 2012; Childers et al., 2014; Lee ).

The 5′ UTR of UME6 dictates Ume6 protein levels, without impacting on mRNA level or induction by hyphal signals (Childers et al., 2014). Polysome association data are consistent with the effect being on mRNA translation (Childers et al., 2014). The mechanism(s) are not defined, but the authors proposed that secondary structure formation by the 5′ UTR could inhibit translation, or alternatively binding of RNA binding proteins could impair ribosome association or target the UME6 mRNA to a translationally silent location, such as P‐bodies (Childers et al., 2014). Given abundance of Ume6 alone determines whether C. albicans exists as yeast or hyphal form (Carlisle et al., 2009), translational regulation might enable fast, reversible yeast–hyphae–yeast morphogenesis that could be important for disease.

To lift repression of the hyphal programme, the mRNA and protein levels of the repressor Nrg1 drop rapidly upon hyphal induction, and NRG1 is regulated transcriptionally (Braun et al., 2001; Murad et al., 2001; Sellam et al., 2010; Lassak et al., 2011; Lu et al., 2011; Childers and Kadosh, 2015). Further to this, the stability of the NRG1 mRNA is controlled by the transcription factor Brg1, via production of an anti‐sense transcript located in the NRG1 open reading frame (Cleary et al., 2012). The exact mechanism of this regulation remains to be elucidated. Nrg1 protein levels are further controlled by the RNA binding protein Ssd1 (Lee ). In S. cerevisiae Ssd1 interacts with, and regulates the translation and subcellular localization of, a suite of mRNAs encoding cell wall remodelling factors (Hogan et al., 2008; Jansen et al., 2009; Kurischko et al., 2011). The translational repressor activity of Ssd1 is negatively regulated through phosphorylation by the RAM (regulation of Ace2 and morphogenesis) network kinase Cbk1, and this process is important for cell wall remodelling during cell division (Jansen et al., 2009). In C. albicans the RAM network is required for hyphal morphogenesis (Song et al., 2008). The cbk1 mutant does not reduce Nrg1 protein levels upon hyphal signalling, but this can be rescued by deletion of SSD1 (Lee ). This indicates that in the absence of Cbk1, Ssd1 promotes NRG1 translation even when hyphal signalling is ‘on’. This, together with the C. albicans ssd1 mutant not being compromised for hyphal morphogenesis (Song et al., 2008), suggests that Ssd1 is a repressor of filamentation.

mRNA trafficking via She3 impacts on hyphal morphogenesis

In S. cerevisiae, the She3‐system transports mRNAs to the site of bud growth and repress their translation en route, reviewed in Haag et al. (2015). The RNA binding proteins She2 and She3 interact together and with localization sites in the mRNA, and form a complex with the myosin V motor Myo4 that enables mRNA trafficking along actin cables (Haag et al., 2015). As shown by the prototypical example of the ASH1 mRNA, two further RNA binding proteins, Puf6 and Khd1, bind mRNA and are required for translational repression (Paquin et al., 2007; Deng et al., 2008). Three of the RNA binding proteins (She3, Puf6 and Khd1) have orthologues in C. albicans, but only She3 has been characterized. C. albicans She3 interacts with 31 and 38 mRNAs during yeast and hyphal growth respectively (Elson et al., 2009). The targets represent several functions, including cell wall, hyphal‐induced genes and transcription factors, and She3 is required for asymmetric mRNA localization (Elson et al., 2009). The she3 mutant is defective in invasive filamentation on solid media and is further defective in causing damage to epithelial cells (Elson et al., 2009). However, she3 could initiate hyphal morphogenesis in liquid media (Elson et al., 2009), suggesting that mRNA targeting by She3 is dispensable for hyphal initiation, but important for long‐term filamentous growth and invasion.

Further to the regulators mentioned earlier, the serine‐arginine RNA binding protein of C. albicans Slr1 is also required for hyphal morphogenesis (Ariyachet et al., 2013). Slr1 could play multiple roles in mRNA physiology, including splicing and translation, but its RNA targets for hyphal morphogenesis remain to be identified.

Post‐transcriptional regulation of biofilm formation

Candida albicans grows drug‐resistant biofilms on various substrates and medical devices, seeding life‐threatening infections. Given that hyphae are important structural components of biofilms, it can be predicted that post‐transcriptional regulators described earlier will play roles in biofilm formation. This is exemplified by the xrn1 mutant, which was identified as biofilm‐defective because of impaired hyphal morphogenesis (Richard et al., 2005).

Transcription factors have been extensively studied for biofilm phenotypes (Nobile et al., 2005, Nobile et al., 2012). How post‐transcriptional regulators control the biofilm transcriptome is poorly defined. The first indication that post‐transcriptional mRNA regulation might play a role came from a bioinformatics approach performed by our colleagues and us (Verma‐Gaur et al., 2015). Thirty six genes related to mitochondria and down‐regulated in biofilms (Nobile et al., 2012) are putative targets of the RNA binding protein Puf3 (Verma‐Gaur et al., 2015). Puf3 is a PUF family member that in S. cerevisiae regulates a network of mRNAs necessary for mitochondrial biogenesis (Gerber et al., 2004). It does so by binding to sequence elements in 3′ UTRs to control mRNA decay, as well as transcript localization to mitochondria (Olivas and Parker, 2000; Gerber et al., 2004; Saint‐Georges et al., 2008). Bioinformatics and functional data in C. albicans are consistent with regulation of mitochondrial biogenesis being a conserved role, as is the function of Puf3 in promoting mRNA decay (Verma‐Gaur et al., 2015). We proposed therefore that Puf3 is involved in controlling mitochondria‐related genes, as part of metabolic changes characteristic of C. albicans biofilms (Verma‐Gaur et al., 2015). However, the puf3 mutant displayed a normal biofilm phenotype. This could result from redundancy, as studies in S. cerevisiae have shown transcripts tend to interact with several RNA binding proteins (Hogan et al., 2008). For example, although an earlier study reported minimal correspondence between Puf3 targets and those of other PUFs in yeast (Gerber et al., 2004), more recent works found that close to one third (30.5%) of mRNAs bound by Puf3 are also bound by Puf5 (Wilinski et al., 2015). An alternative explanation is that although the mRNA targets of Puf3 are differentially expressed in C. albicans biofilms, their proper regulation is not required for biofilm maturation. It could also be that alternative mechanisms compensate when Puf3 targets are dis‐regulated. In contrast to the puf3 mutant, deadenylase mutants ccr4 and pop2 displayed altered biofilms, with the most intriguing phenotype being overproduction of extracellular matrix (Verma‐Gaur et al., 2015). Ccr4 likely regulates matrix production by two mechanisms, through the expression of genes involved in matrix carbohydrate regulation (specifically mannan), as well as via metabolic effects (Verma‐Gaur et al., 2015). Extracellular matrix production is related to biofilm drug resistance (Nett et al., 2010a; Nett et al., 2010b), but its control remains to be fully understood. In addition to Ccr4 the transcription factor Zap1 is a further negative regulator of biofilm matrix production (Nobile et al., 2009). Understanding the interface between transcriptional and post‐transcriptional regulation of this important biofilm phenotype is a topic for future studies.

Post‐transcriptional regulators in stress responses and virulence

Mutants in several mRNA regulators display reduced virulence in the mouse systemic candidiasis model: ccr4 (Dagley et al., 2011), slr1 (Ariyachet et al., 2013) and ssd1 (Gank et al., 2008). In addition, the xrn1 mutant is less virulent in Galleria mellonella (Fuchs et al., 2010). The reasons for reduced virulence are likely multifactorial, including crippled fitness, hyphal defects and altered cell wall integrity. The cell wall is particularly relevant, because its biogenesis is targeted by the echinocandin drugs used to treat Candida infections. Given that the C. albicans ccr4 and pop2 mutants display changes to cell wall composition and increased susceptibility to the echinocandin drug caspofungin (Dagley et al., 2011), we proposed that inactivation of the deadenylase could be considered for combinatorial therapy (Panepinto et al., 2013). Several cell wall‐related genes are dis‐regulated in the ccr4 mutant, but their levels are mostly up‐regulated compared with controls suggestive of compensatory activation (Dagley et al., 2011; Verma‐Gaur et al., 2015). Ccr4 has further been implicated in regulating cell wall genes in hypoxia (Sellam et al., 2014).

Ssd1 is also required for cell wall integrity in C. albicans (Gank et al., 2008; Song et al., 2008), consistent with roles in S. cerevisiae (Kaeberlein et al., 2002, Hogan et al., 2008). In S. cerevisiae, deletion of SSD1 and CCR4 results in a negative genetic interaction, suggesting that they act in parallel pathways for wall integrity (Kaeberlein et al., 2002). Ssd1 has an additional role in regulating the susceptibility of C. albicans to antimicrobial peptides (Gank et al., 2008; Jung et al., 2013). This might stem from functions in cell surface biogenesis, and a role in the expression of BCR1, a transcription factor necessary for antimicrobial peptide resistance (Jung et al., 2013). The C. albicans decapping factor Edc3 also has a stress responsive role in oxidative stress and translational regulation of superoxide dismutase Sod1 and the catalase Cat1 (Jung et al., 2014). Collectively, these observations suggest that post‐transcriptional regulators could be considered as drug targets, because of requirements for virulence and potential for combinatorial treatment with current antifungal drugs or antimicrobial peptides.

Evolution of post‐transcriptional gene regulation in fungi

Changes to gene expression control are a known mechanism of evolutionary divergence between species. Evolutionary rewiring of transcriptional regulation has been extensively studied in fungi and other eukaryotes. For this, C. albicans has served as a useful comparison to S. cerevisiae because it is a related, but biologically divergent yeast (e.g. Tsong et al., 2003; Ihmels et al., 2005; Hogues et al., 2008; Brown et al., 2009). As with transcription factors, RNA binding proteins interact with functionally related genes, thereby establishing post‐transcriptional RNA networks, reviewed in Keene (2007) and Blackinton et al. (2014). Understanding the evolution of post‐transcriptional regulatory networks in fungi is still in its infancy.

Examples of differences in post‐transcriptional mRNA control have been uncovered between S. cerevisiae and C. albicans, with orthologous RNA binding proteins displaying distinct regulation of the same mRNA, or having distinct sets of targets in the two species. Within fungi, the PUF family of RNA binding proteins is best studied for evolutionary aspects (Jiang et al., 2012; Hogan et al., 2015; Verma‐Gaur et al., 2015). There are five PUF family members in C. albicans, all with orthologues in S. cerevisiae. Bioinformatics suggests that several C. albicans PUFs predominantly share their mRNA targets with S. cerevisiae (Jiang et al., 2010; Jiang et al., 2012; Hogan et al., 2015; Verma‐Gaur et al., 2015). However, in one case uncovered so far, regulation by Puf3 appears to be distinct in the two yeasts (Verma‐Gaur et al., 2015). In C. albicans Puf3 facilitates mRNA decay of the mitochondrial ribosomal subunit MRPL25 in two carbon sources important in human body environments, fermentable glucose and non‐fermentable lactate (Verma‐Gaur et al., 2015). In S. cerevisiae this is the case in glucose, but not lactate (Verma‐Gaur et al., 2015). This recent work from our laboratory fits with previous studies showing that S. cerevisiae Puf3 is an mRNA decay factor only in glucose (Miller et al., 2014), and moreover, it can mediate translation of mitochondrial biogenesis factors in non‐fermentable carbon, with the change from repressor to activator regulated by phosphorylation of Puf3 via glucose‐regulated signalling (Lee and Tu, 2015a). Importantly, transcription of mitochondrial ribosomal genes is a well‐known example of distinct control in C. albicans and S. cerevisiae, likely reflecting their differences in fermentative growth. Of note, although numerous mitochondrial ribosomal subunits are Puf3 targets in both S. cerevisiae and C. albicans, for two other transcripts tested a much smaller degree of difference was seen in their regulation between the two species (Verma‐Gaur et al., 2015). It will be important to comprehensively address similarities and differences in Puf3‐dependent regulation of mitochondrial biogenesis factors in C. albicans and S. cerevisiae by transcriptome‐wide approaches, as well as determine how C. albicans Puf3 regulates mRNA translation and localization to mitochondria in response to environmental and nutritional signals. This will lead to more complete understanding of how evolutionary changes shaped distinct regulation of mitochondrial biogenesis on multiple levels of control.

In contrast to the conservation of PUF targets between S. cerevisiae and C. albicans, the mRNAs localized by the She3‐system differ substantially, with only two shared targets (Elson et al., 2009). The authors proposed that this reflects the two yeasts being subject to divergent environmental conditions in the human body (C. albicans) versus habitats such as fruit (S. cerevisiae) (Elson et al., 2009). Similarly, the targets of the C. albicans tristetraprolin family RNA binding protein Zfs1 have diverged from its S. cerevisiae orthologues Cth1 and Cth2 (Wells et al., 2015). Moreover, bioinformatic searches for Zfs1 binding sites in very closely related Candida species from the CTG clade suggested rapid evolution of Zfs1 functions (Wells et al., 2015). In S. cerevisiae Zfs1 orthologues regulate the response to iron deficiency through decay of mRNAs with functions in iron‐dependent processes (Puig et al., 2005; Puig et al., 2008). The ability to respond to iron is important for commensalism and pathogenicity of C. albicans (Chen et al., 2011). It will be interesting to determine whether an RNA binding protein different to Zfs1 is involved in iron homeostasis.

Outlook

With studies showing that post‐transcriptional mRNA regulators are important for C. albicans biology and virulence, the challenge now is to understand the functions and identify the mRNA targets of many more RNA binding proteins in this pathogen and decipher how regulation of post‐transcriptional mRNA networks mediates the response of C. albicans to its environment. Constructing libraries of RNA binding protein mutants, coupled with systems‐biology approaches that have been developed in S. cerevisiae, will achieve these goals. Ultimately, integrating this knowledge with transcriptional circuits will provide a more accurate picture of how C. albicans adapts to host and antifungal drug stresses and might offer new avenues for therapy.