While virulence properties of Candida albicans, the most commonly isolated human fungal pathogen, are controlled by transcriptional and post-translational mechanisms, considerably little is known about the role of post-transcriptional, and particularly translational, mechanisms. We demonstrate that UME6, a key filament-specific transcriptional regulator whose expression level is sufficient to determine C. albicans morphology and promote virulence, has one of the longest 5′ untranslated regions (UTRs) identified in fungi to date, which is predicted to form a complex and extremely stable secondary structure. The 5′ UTR inhibits the ability of UME6, when expressed at constitutive high levels, to drive complete hyphal growth, but does not cause a reduction in UME6 transcript. Deletion of the 5′ UTR increases C. albicans filamentation under a variety of conditions but does not affect UME6 transcript level or induction kinetics. We show that the 5′ UTR functions to inhibit Ume6 protein expression under several filament-inducing conditions and specifically reduces association of the UME6 transcript with polysomes. Overall, our findings suggest that translational efficiency mechanisms, known to regulate diverse biological processes in bacterial and viral pathogens as well as higher eukaryotes, have evolved to inhibit and fine-tune morphogenesis, a key virulence trait of many human fungal pathogens.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Candida albicans is the major cause of human fungal disease worldwide. While normally a commensal in the mammalian host, this organism is responsible for a wide range of mucosal and systemic infections (Odds, 1988; Calderone and Clancy, 2012). Immunocompromised individuals, including cancer patients on chemotherapy, organ transplant recipients, and recipients of artificial joints and prosthetic devices, are particularly susceptible (Dupont, 1995; Weig et al., 1998). Candida species are now the fourth-leading cause of hospital-acquired bloodstream infections in the USA with a ∼ 40% mortality rate (Edmond et al., 1999; Wisplinghoff et al., 2004) and approximately $1 billion per year is spent in this country on antifungal therapies to treat systemic candidiasis (Miller et al., 2001).
Candida albicans is known to possess a number of properties which contribute to virulence, including the ability to undergo a reversible morphological transition from single oval-shaped yeast cells to pseudohyphal and hyphal filaments (elongated cells attached end-to-end) in response to specific environmental cues in the host environment (e.g. serum and body temperature, 37°C) (Odds, 1988; Calderone and Clancy, 2012). Hyphal filaments are known to play an important role in tissue invasion, lysis of macrophages as well as evasion of the host immune system (Kumamoto and Vinces, 2005). Several key experiments have also indicated that the C. albicans yeast–filament transition is required for virulence in a mouse model of systemic candidiasis (Braun and Johnson, 1997; Lo et al., 1997; Saville et al., 2003). Additional virulence-related processes include adhesion to host epithelial and endothelial cells, secretion of degradative enzymes, phenotypic switching and the ability to form biofilms on host surfaces as well as implanted medical devices. Both phenotypic switching and biofilm formation can lead to the development of antifungal drug resistance (Sundstrom, 1999; Douglas, 2002; Schaller et al., 2005; Hoyer et al., 2008; Calderone and Clancy, 2012).
Given the importance of the virulence properties described above for C. albicans pathogenicity, intense research efforts have focused on elucidating molecular mechanisms by which they are controlled, particularly in response to host environmental cues. Thus far, considerable progress has been made towards the identification and characterization of post-translational mechanisms. Both mitogen-activated protein (MAP) kinase and Ras cAMP-protein kinase A (PKA) signalling pathways have been shown to mediate C. albicans filamentation in response to a variety of environmental cues including starvation, serum and glucose (Biswas et al., 2007). Phosphorylation of septins and other targets by the Hgc1-Cdc28 cyclin/Cdk complex under filament-inducing conditions is important for the physical process of C. albicans hyphal development (Wang, 2009). Histone acetylation and/or deacetylation are also important for C. albicans stress adaptation, survival in macrophages, morphogenesis and phenotypic switching (Hnisz et al., 2010; Lopes da Rosa et al., 2010; Lopes da Rosa and Kaufman, 2012). In addition, ubiquitination and sumoylation have both been shown to control C. albicans morphogenesis, cell cycle progression and stress response (Leach et al., 2011a,b; Leach and Brown, 2012).
Significant progress has also been made towards the identification of transcriptional mechanisms that control C. albicans virulence properties. A variety of transcriptional regulators of filamentous growth (e.g. Efg1, Cph1, Nrg1) have been shown to function as downstream targets of the MAP kinase and/or Ras-cAMP-PKA signalling pathways described above (Biswas et al., 2007; Lu et al., 2011; Calderone and Clancy, 2012). Many of these regulators control the expression of adhesins and secreted degradative enzymes (Felk et al., 2002; Sohn et al., 2003; Kadosh and Johnson, 2005). In addition, transcriptional regulatory mechanisms are known to control phenotypic switching, biofilm formation, stress response, iron acquisition and the development of antifungal drug resistance (Zordan et al., 2007; Sanglard et al., 2009; Chen et al., 2011; Nobile et al., 2012; Sellam et al., 2012).
In contrast to post-translational and transcriptional mechanisms, considerably little is known about post-transcriptional mechanisms that control C. albicans virulence properties. A She3-dependent RNA transport system is important for both hyphal development and the transport of certain filament-specific transcripts to the hyphal tip (Elson et al., 2009). A nuclear localization mechanism has recently been shown to control Sef1, a transcriptional regulator important for virulence and iron uptake in the host (Chen and Noble, 2012). An mRNA stability mechanism is also known to regulate filamentation and virulence (Cleary et al., 2012) and the Ccr4–Pop2 mRNA deadenylase complex controls cell wall integrity, filamentation and antifungal drug resistance (Chen et al., 2001; Tucker et al., 2001; Dagley et al., 2011). Information regarding translational control of C. albicans virulence properties is particularly lacking. Interestingly, however, a recent whole-genome RNA-Seq analysis has revealed that several key transcripts encoding proteins involved in filamentation, biofilm formation, white-opaque phenotypic switching (important for mating), adhesion, degradation of host membrane proteins and other processes important for C. albicans pathogenesis have unusually long (> 500 bp) 5′ untranslated regions (UTRs) (Bruno et al., 2010); similar results were found in an independent experiment using high-resolution C. albicans tiling arrays (Sellam et al., 2010). Previous studies in other organisms have revealed that 5′ UTRs can play an important role in reducing overall translational efficiency by a variety of mechanisms, including the formation of highly stable secondary structures which inhibit ribosome scanning and/or accessibility (Mignone et al., 2002).
In this study, we identify an exceptionally long 5′ UTR and examine the role that this element plays in controlling the expression of UME6, which encodes a critical filament-specific transcriptional regulator of the C. albicans morphological transition. UME6 is important for controlling the level and duration of filament-specific gene expression in response to filament-inducing conditions. Strains deleted for UME6 are defective for hyphal extension and attenuated for virulence in a mouse model of systemic candidiasis (Banerjee et al., 2008; Zeidler et al., 2009) and a recent study has specifically demonstrated that Ume6 protein levels play a critical role in these processes (Lu et al., 2013). We have also previously shown that expression levels of UME6 are sufficient to sequentially specify C. albicans yeast, pseudohyphal and hyphal morphologies in a dosage-dependent manner (Carlisle et al., 2009). Here, we demonstrate that the UME6 5′ UTR plays an important role in specifically inhibiting Ume6 protein expression, thus affecting the ability of this regulator to determine C. albicans morphology. We also specifically examine the role of this 5′ UTR in controlling UME6 at the level of translational efficiency and provide new information about translational regulation of morphological transitions in pathogenic fungi.
Analysis of the C. albicansUME6 upstream region identifies an unusually long 5′ UTR
Given the importance of UME6 for C. albicans morphology determination and virulence, we initially sought to identify and characterize upstream regulatory elements that control expression of this transcription factor. We immediately noticed the presence of an unusually long intergenic region (20.6 kb) between UME6 and the nearest upstream gene, orf19.177, on C. albicans chromosome 1 (Fig. 1A). In order to identify upstream elements in the intergenic region important for UME6 function, we first generated a ume6Δ/+ strain in which one allele of the UME6 coding sequence, along with 19.3 kb of upstream sequence, was deleted and replaced with a HIS1 marker (this strain was phenotypically equivalent to a strain deleted for one copy of only the UME6 open reading frame). Next, a variety of deletions of varying lengths were generated in the upstream intergenic region of the second UME6 allele. Morphology phenotypes of all deletion strains, along with ume6Δ/+ and ume6Δ/Δ control strains, were compared on solid medium in the presence of serum at 37°C, a strong filament-inducing condition (Fig. 1B). Consistent with previous observations (Banerjee et al., 2008), the ume6Δ/Δ mutant was significantly defective for filamentation relative to the ume6Δ/+ strain. We observed that deletion strains containing only 4 kb and 5 kb of the upstream UME6 intergenic region showed a filamentation defect equivalent to that of the ume6Δ/Δ mutant. In contrast, strains containing 6 kb or greater of the UME6 upstream region showed a level of filamentation similar to that of the ume6Δ/+ parent strain. These results indicate that at least 6 kb of the upstream region is required for UME6 function with respect to filamentation and suggest that critical promoter elements are located in this region. In order to test this hypothesis, a C. albicans strain was generated in which the 6 kb UME6 upstream region was placed upstream of a heterologous Streptococcus thermophilus lacZ reporter gene. As shown in Fig. 1C, the 6 kb UME6 upstream region was sufficient to drive ∼ 25-fold transcriptional induction of the lacZ reporter in the presence of serum at 37°C versus 30°C only. In contrast, the lacZ reporter alone did not show significant induction. These results indicate that the 6 kb UME6 upstream region contains important promoter elements and suggest that deletion strains containing only 5 kb and 4 kb of upstream sequence are defective for filamentation because they lack these elements. In order to map the transcriptional initiation site within this region, we performed a 5′ RACE (rapid amplification of cDNA ends) analysis using cDNA prepared from wild-type cells grown in the presence of serum at 37°C. This analysis revealed the presence of three transcript start sites located at positions −3041 bp, −2126 bp and −1923 bp relative to the translational initiation codon (+1) (Fig. 1D and Fig. S1). These findings are consistent with a previous report which, using RNA-Seq analysis, indicated that the UME6 5′ UTR is greater than 1500 bp in length (Bruno et al., 2010). Based on the total size of the UME6 transcript, as determined by Northern analysis, as well as known sizes of the UME6 open reading frame and 3′ UTR (Braun et al., 2005; Bruno et al., 2010), the major transcription start site is located at 3041 bp upstream of the UME6 start codon. UME6 transcripts whose sizes are consistent with 2.1 kb and 1.9 kb transcription start sites were not detected by Northern analysis, suggesting that these are minor start sites. The 3041 bp 5′ UTR of UME6 is much longer than the average 5′ UTR (∼ 150 bp, based on our calculations) in C. albicans (Bruno et al., 2010) and, based on a search of the UTRdb database (Grillo et al., 2010), is one of the longest 5′ UTRs identified in fungi to date.
An in silico analysis of the UME6 5′ UTR sequence revealed the presence of two upstream open reading frames (uORFs) which are 39 bp and 102 bp in length and located at −1519 bp and −685 bp, respectively, upstream from the UME6 start codon (Fig. 1D). In order to perform a predicted structure analysis of the UME6 transcript we used the RNA folding program mFold (http://mfold.rna.albany.edu/?q=mfold/) (Zuker, 2003). This program is widely used and was selected because it performs calculations based on Minimum Folding Energy (MFE) algorithms, which are typically more accurate in determining major substructures of RNAs with low folding energies. As indicated in Fig. 2, an mFold analysis of the full-length UME6 transcript sequence predicted that the 5′ UTR forms a complex secondary structure. The 5′ UTR alone is predicted to be extremely stable with a very low folding free energy (ΔG) of −468.1 kcal/mol; the full-length UME6 transcript has a predicted ΔG of −1017.7 kcal/mol. An independent analysis using RNAFold (http://rna.tbi.univie.ac.at/cgi-bin/RNAfold.cgi) (Zuker and Stiegler, 1981), a different MFE-based RNA structural prediction program, predicted similar structures and folding energies. The average folding free energy for 5′ UTRs of transcripts encoding transcription factors, growth factors, receptors and other regulatory proteins across species is −50 kcal/mol (Davuluri et al., 2000). Both uORFs and stable secondary structures are characteristic of 5′ UTRs which are known to play important roles in translational regulation (Mignone et al., 2002) and their apparent presence in this 5′ UTR suggested that UME6 may be controlled by such a mechanism.
Deletion of the UME6 5′ UTR increases C. albicans filamentous growth but does not affect UME6 transcript level or induction kinetics
In order to determine whether the long 5′ UTR was important for the ability of UME6 to control C. albicans filamentous growth, we generated a strain in which both copies of the UME6 5′ UTR were deleted. The 5′ UTR deletion spanned from positions −3011 bp to −47 bp upstream of the UME6 start codon and this sequence was replaced with a 34 bp FRT site as well as XhoI and NotI restriction sites (see Supporting information for details as well as a complete sequence of the fusion joint). Importantly, the complete UME6 open reading frame (ORF), as well as 46 bp immediately upstream of the UME6 AUG start codon, were left intact. As shown in Fig. 3A, the UME6 5′ utrΔ/Δ strain showed a mild increase in filamentation, relative to that of the wild-type control strain, when cells were grown on solid YEPD medium in the presence of serum at 37°C. The increased filamentous growth phenotype of the UME6 5′ utrΔ/Δ mutant was more pronounced on solid Spider (nitrogen and carbon starvation) and Lee's pH 6.8 media, but was not observed when cells were grown on solid YEPD medium at 30°C for 3 days (non-filament-inducing conditions). The UME6 5′ utrΔ/Δ mutant also showed greater filamentation than that of a wild-type control strain in liquid Spider medium at 37°C (data not shown). These results indicate that the UME6 5′ UTR functions to inhibit C. albicans filamentation in response to a variety of filament-inducing conditions.
Many UTRs (primarily 3′ UTRs) are known to control the level of their respective transcripts (Mignone et al., 2002). In order to determine whether the UME6 5′ UTR plays a role in controlling UME6 transcript level and/or induction kinetics, a serum and temperature induction time-course experiment was carried out using both UME6 5′ utrΔ/Δ and wild-type control strains. Cells from each strain were harvested at the zero time point, as well as various post-induction time points, and total RNA was prepared for Northern analysis. As shown in Fig. 3B, UME6 induction kinetics, as well as overall transcript level, in the UME6 5′ utrΔ/Δ strain appeared nearly identical to those observed in the wild-type control strain. We have also observed that both of these strains express UME6 at an equivalent level when cells are grown in Spider medium at 37°C (Fig. S2).
The UME6 5′ UTR inhibits the ability of UME6, when expressed at constitutive high levels, to drive complete hyphal growth but does not cause a reduction in UME6 transcript
We have previously demonstrated that constitutive high-level expression of UME6 is sufficient to drive nearly complete hyphal growth of C. albicans in the absence of filament-inducing conditions (Carlisle et al., 2009). This result was obtained using a strain in which the Escherichia coli tet operator (tetO) was placed upstream of the start codon for one allele of UME6 (the 5′ UTR was not present). In the absence of doxycycline (Dox, a tetracycline derivative), a transactivator binds tetO and UME6 is expressed at high constitutive levels; in the presence of Dox, the UME6 allele is shut off. When the tetO-UME6 strain is initially grown in the presence of Dox and then transferred to medium lacking Dox, C. albicans cells sequentially transition from yeast to pseudohyphae to a nearly complete hyphal population over a time-course in the absence of filament-inducing conditions. In order to determine the effect of the UME6 5′ UTR on the ability of UME6 levels to specify C. albicans morphology, a similar strain was generated in which the tet operator was placed immediately upstream of the 3 kb 5′ UTR. Both tetO-5′ UTR-UME6 and tetO-UME6 strains were grown under non-filament-inducing conditions in the presence of Dox, and then cultures were diluted into medium lacking Dox, as described above. As a control, cultures from both strains were also diluted into medium containing Dox. Cell morphology of both strains was examined at various time points following Dox depletion. By the 3 h time point, the tetO-5′ UTR-UME6 strain grew as pseudohyphae (Fig. 4A). We also observed that this strain showed a higher proportion of yeast cells (32%) compared to that of the tetO-UME6 strain (14%). By the 10 h time point, the tetO-UME6 strain had transitioned to a nearly complete hyphal population [note: in our previous study (Carlisle et al., 2009) this transition was completed within 9 h, most likely due to differences in culture volume] whereas the tetO-5′ UTR-UME6 strain grew as a mixture of yeast, pseudohyphae and hyphae. Even following growth overnight in the absence of Dox, the tetO-5′ UTR-UME6 strain showed a mixed population of cell morphologies and was unable to transition completely to hyphae (Fig. S3). Also, consistent with previous observations (Carlisle et al., 2009; Carlisle and Kadosh, 2010), control cells of both strains grown in the presence of Dox remained in the yeast morphology (data not shown). Overall, these results indicate that the 5′ UTR is important for the ability of UME6 expression levels to specify C. albicans morphology.
In order to examine whether the 5′ UTR affects C. albicans morphology determination by causing a reduction in UME6 transcript levels, total RNA was prepared from cells of both tetO-5′ UTR-UME6 and tetO-UME6 strains at each time point of the time-course experiment described above and used for quantitative RT-PCR analysis. As shown in Fig. 4B, in the tetO-UME6 strain the UME6 transcript was induced 8.7-fold at the 1 h time point following Dox depletion and remained induced at a high level (12- to 19-fold) from the 2 h time point through the remainder of the time-course. UME6 was expressed at roughly equivalent levels in both the tetO-5′ UTR-UME6 and tetO-UME6 strains at the 1 h time point in the absence of Dox. Interestingly, over the remaining time points in the absence of Dox, UME6 levels in the tetO-5′ UTR-UME6 strain were consistently higher (38- to 54-fold induced) than those of the tetO-UME6 strain. As expected, UME6 was not expressed to a significant degree in either strain when cells were grown in the presence of Dox. Overall, these results clearly indicate that the inability of the tetO-5′ UTR-UME6 strain to drive complete hyphal growth in the absence of Dox cannot be attributed to a reduction in UME6 transcript levels.
The UME6 5′ UTR functions to inhibit translational efficiency of UME6
Given the failure of the 5′ UTR to cause a reduction in UME6 transcript level or control UME6 induction kinetics, we hypothesized that this region may control the translational efficiency of UME6. This hypothesis was supported by previous reports documenting the role of 5′ UTR regions in translational regulation, as well as our observation that the UME6 5′ UTR is predicted to possess a highly stable secondary structure and contains two uORFs, all of which have been associated with translational control in prior studies (Mignone et al., 2002; Pickering and Willis, 2005). In order to test this hypothesis, we performed a polysome profiling assay. Wild-type and UME6 5′ utrΔ/Δ strains showed similar polysome profiles when cells were grown under filament-inducing (serum at 37°C) and non-filament-inducing (30°C) conditions (Fig. 5A). In addition, treatment with EDTA, a known inhibitor of polysome formation, disrupted the polysome profile in both strains, as expected. As indicated in Fig. 5B, the UME6 transcript generally showed significantly greater association with the polysome fractions of the UME6 5′ utrΔ/Δ versus wild-type control strain when cells were induced by serum at 37°C. The greatest differences in association (∼ 10-fold) were observed in the final fractions (10 and 11), which are richest in polysomes. The UME6 transcript of the wild-type strain generally showed a similar abundance, regardless of whether qRT-PCR primers for the open reading frame or 5′ UTR were used for detection, indicating that the 5′ UTR was still present. In addition, an ACT1 control transcript generally did not show large differences in abundance between the wild-type and UME6 5′ utrΔ/Δ strains (Fig. S4). As expected, we also observed an overall shift in transcript abundance from polysome-bound to unbound fractions upon treatment with EDTA (data not shown). Also consistent with previous observations that UME6 is transcriptionally induced in a filament-specific manner (Banerjee et al., 2008; Zeidler et al., 2009), very low levels of UME6 transcript were observed in all fractions when cells of both wild-type and UME6 5′ utrΔ/Δ strains were grown under non-filament-inducing conditions (30°C) (Fig. 5B). These results strongly suggest that the 5′ UTR functions to inhibit UME6 translational efficiency by reducing association of the UME6 transcript with polysomes under filament-inducing conditions.
In order to confirm that the UME6 5′ UTR functions to inhibit Ume6 protein expression, the 5′ UTR was deleted in a strain expressing Myc-tagged Ume6 and cells were induced to form filaments by growth at 37°C in the presence of serum. As shown in Fig. 6A, Ume6-Myc showed a significantly greater induction in the 5′ utrΔ-UME6-MYC strain compared to that observed in the UME6-MYC control strain. We also observed increased expression of Ume6-Myc in the 5′ utrΔ-UME6-MYC versus UME6-MYC strain in the presence other filament-inducing conditions, including 37°C only, Spider at 37°C and Lee's pH 6.8 media (Fig. 6A and B). Interestingly, the extent to which Ume6 levels rose upon deletion of the 5′ UTR appeared to vary between filament-inducing conditions (e.g. compare 37°C + Spider and Lee's pH 6.8). These differences were reproducible based on multiple replicates (two replicates for 37°C and 37°C + Serum and three replicates for 37°C + Spider and Lee's pH 6.8), suggesting that translational inhibition by the 5′ UTR may, to some degree, be modulated by environmental signals which control C. albicans filamentation. A Northern analysis indicated that for all filament-inducing conditions there was not a significant difference in the UME6 transcript level in UME6-MYC versus 5′ utrΔ-UME6-MYC strains (Fig. S5). As expected, UME6 levels were generally higher in the strongest filament-inducing condition, serum at 37°C, compared to those observed in Spider at 37°C, Lee's pH 6.8 and 37°C only, which are weaker inducing conditions. Overall, these findings suggest that reduced translational efficiency directed by the UME6 5′ UTR results in a significant decrease in Ume6 protein expression in the presence of a variety of filament-inducing conditions.
To determine whether the UME6 5′ UTR was sufficient to inhibit translation, we placed the 5′ UTR immediately upstream of a heterologous GFPγ reporter gene driven by a constitutive ACT1 promoter. ACT1pr-UME6 5′ UTR-GFP and ACT1pr-GFP strains were grown under non-filament-inducing conditions and GFP protein expression was quantified by fluorometry. As shown in Fig. 7A and Fig. S6, the 5′ UTR caused a significant reduction in GFP protein levels. Indeed, the ACT1pr-UME6 5′ UTR-GFP strain showed fluorescence values equivalent to those of a wild-type control strain which does not express GFP. In addition, removal of both uORFs did not affect the ability of the UME6 5′ UTR to inhibit GFP protein expression. Importantly, the ratio of GFP protein to transcript (as determined by qRT-PCR) was significantly higher in the ACT1pr-GFP strain when compared to that of the ACT1pr-UME6 5′ UTR-GFP and ACT1pr-UME6 5′ UTR uorf1Δ uorf2Δ-GFP strains (Fig. 7B). A Northern analysis also confirmed that the GFP transcript was expressed at equivalent levels in all three of these strains and not expressed in a wild-type control strain (Fig. S7). These results strongly suggest that the UME6 5′ UTR is sufficient to inhibit translation, but not transcription, via a uORF-independent mechanism when placed in the context of a heterologous promoter.
In C. albicans, the most commonly isolated human fungal pathogen, a variety of post-translational and/or transcriptional mechanisms are known to be involved in the regulation of morphogenesis, adhesion, secretion of degradative enzymes, biofilm formation, phenotypic switching and other virulence properties. However, significantly less is known about post-transcriptional, and especially translational, mechanisms that control these processes [although several genes are known to be translationally regulated during filamentation in the non-pathogenic model yeast Saccharomyces cerevisiae (Park et al., 2006)]. Here, we describe a 5′ UTR-mediated translational efficiency mechanism that plays an important role in inhibiting C. albicans morphogenesis by controlling the expression of Ume6, a key filament-specific transcription factor. We provide several lines of evidence to support this mechanism: (i) deletion of the UME6 5′ UTR causes increased filamentation and hyphal growth but does not affect the induction kinetics or level of the UME6 transcript, (ii) the 5′ UTR inhibits the ability of UME6, when expressed at constitutive high levels, to drive complete hyphal formation but does not cause a reduction in UME6 transcript, (iii) the 5′ UTR specifically inhibits the ability of the UME6 transcript to associate with polysomes and also inhibits Ume6 protein expression in the presence of a variety of filament-inducing conditions, and (iv) the UME6 5′ UTR is sufficient to inhibit translation, but not transcription, of a heterologous reporter gene. In addition, the UME6 5′ UTR is exceptionally long and contains several features which have previously been associated with translational control (Mignone et al., 2002; Pickering and Willis, 2005).
How exactly does the 5′ UTR function to inhibit translational efficiency of UME6? Based on previous studies (Mignone et al., 2002; Pickering and Willis, 2005), four possible mechanisms may explain our observations. First, as indicated by our in silico analysis, the UME6 5′ UTR contains two putative uORFs. Because, in eukaryotes, the small ribosomal subunit typically initiates translation at the first scanned AUG codon, the large majority of translation in the 5′ UTR may be initiating at the uORFs, rather than at the UME6 start codon, depending on the context of ribosome capture. In this case, translation of UME6 would occur as a result of either leaky ribosome scanning or re-initiation. However, our observation that removal of both uORFs does not significantly affect the ability of the UME6 5′ UTR to inhibit translation appears to exclude this mechanism.
A second mechanism by which the 5′ UTR may control UME6 translational efficiency is by forming an extremely stable secondary structure (Fig. 8). Stable 5′ UTR secondary structures have previously been shown to inhibit the ability of ribosomes to access and/or efficiently scan mRNA transcripts and reach the start codon (Mignone et al., 2002; Pickering and Willis, 2005). This mechanism is supported by our in silico analysis indicating the UME6 5′ UTR is predicted to form a complex and highly stable secondary structure with very low folding free energy. The predicted free energy of the 5′ UTR is about nine times the free energy required for hairpin structures to block ribosome scanning (−50 kcal mol−1) (Pelletier and Sonenberg, 1985; Kozak, 1989). Unfortunately, due to the exceptionally large size and predicted complexity of the UME6 5′ UTR, standard approaches (e.g. compensatory base change and chemical probing experiments) to determine directly whether its secondary structure is important for controlling translational efficiency are not feasible. A deletion series analysis of the 5′ UTR suggested the possible involvement of certain regions in translational control, but was generally difficult to interpret since many of the partial deletion mutants also showed alterations in transcript levels when compared to those observed in UME6-MYC and UME6-5′ utrΔ-MYC strains (data not shown). In addition, partial deletions in the 5′ UTR caused significant changes in predicted overall secondary structure, which further complicated the analysis and made it difficult to determine the role of specific native 5′ UTR structures in translational control.
A third mechanism involves trans-acting RNA-binding proteins, which may bind to specific structural or sequence elements located within the UME6 5′ UTR. These factors can compete with ribosomes for access to the transcript or induce secondary structures which inhibit ribosome scanning (Fig. 8) (Adeli, 2011). Another related possibility is that the UME6 5′ UTR functions to prevent the formation of an internal ribosomal entry site (IRES) which allows ribosomes to directly enter the transcript at the start codon, instead of scanning from the 5′ end. However, IRES elements have typically been identified in viral and mammalian, rather than yeast, mRNAs (Kozak, 2001; Mignone et al., 2002; Pickering and Willis, 2005).
A fourth mechanism for UME6 5′ UTR-mediated translational inhibition may involve alternative mRNA localization (Fig. 8). 5′ UTRs have been shown to affect the subcellular localization of their respective transcripts and C. albicans is known to transport certain mRNAs in a She3-dependent manner to the apical tip of hyphal filaments (Mignone et al., 2002; Elson et al., 2009). P-bodies, known to be important for mRNA degradation and storage, have recently been shown to accumulate during C. albicans hyphal development (Jung and Kim, 2011). Localization of the UME6 transcript to a P-body or other cellular compartment could therefore possibly lead to restricted translation at specific subcellular locations or storage for translation at a later time. The RNA predicted structural analysis, in combination with our demonstration that the 5′ UTR inhibits association of UME6 mRNA with polysomes, strongly suggests that the secondary structure and/or RNA-binding protein mechanisms play an important role. Importantly, the mechanisms described above are not mutually exclusive and could act either alone, in combination, or in conjunction with alternative mechanisms, to control the translational efficiency of UME6.
In many eukaryotic systems, post-transcriptional control mechanisms play an important role in rapidly fine-tuning the expression of genes involved in important developmental processes and the mechanism we have described here appears to be no exception. UME6 encodes a key filament-specific transcriptional regulator which controls C. albicans morphology and virulence, as well as the level and duration of filament-specific gene expression. UME6 also serves as an important downstream target for multiple filamentous growth signalling pathways and we have previously shown that the morphology of C. albicans cells is exquisitely sensitive to UME6 transcript levels (Banerjee et al., 2008; Carlisle et al., 2009; Zeidler et al., 2009). Our observations that the UME6 5′ utrΔ/Δ strain shows a generally mild increase in filamentation only in the presence of filament-inducing conditions and that constitutive high-level expression of UME6 in the presence of the 5′ UTR generates a mixed population of cell morphologies (rather than all yeast) suggest that the 5′ UTR-mediated translational efficiency mechanism serves to rapidly control and fine-tune Ume6 expression levels. Consistent with this observation, we have found that the level of translational inhibition directed by the UME6 5′ UTR can vary in the presence of different filament-inducing conditions, suggesting that translation of Ume6 can be modulated by environmental signals. Because UME6 transcript levels are very low in non-filament-inducing conditions, it was not possible to determine whether the 5′ UTR affects UME6 translation under these conditions. However, our GFP reporter experiment does suggest that the 5′ UTR can inhibit translation under non-filament-inducing conditions as well. A recent report has indicated that stabilization of Ume6 protein by multiple filamentous growth signalling pathways is critical for C. albicans hyphal development and maintenance (Lu et al., 2013). In this respect, the 5′ UTR-mediated translational inhibition mechanism may serve to rapidly reduce Ume6 protein levels, thus preventing unnecessary hyphal growth until appropriate host environmental cues are present.
Our results also suggest that the 5′ UTR may possess elements that can increase UME6 mRNA levels, at least when UME6 is expressed from a heterologous tet operator under non-filament-inducing conditions. Consistent with this observation, a recent report indicates that Ume6 can bind to its own upstream region in the vicinity of the 5′ UTR to increase transcription in a positive feedback loop (Lu et al., 2013). In addition, a transcription factor important for temperature-induced C. albicans morphogenesis, Hms1, has recently been shown to bind the UME6 5′ UTR and induce UME6 expression (Shapiro et al., 2012). Our finding that natural UME6 transcript levels are not altered upon deletion of the 5′ UTR is most likely explained by the observation that a significant number of additional transcriptional regulators appear to play an important role in the activation of UME6 (likely via the promoter) under filament-inducing conditions (Zeidler et al., 2009). Alternatively, changes in chromatin structure resulting from introduction of the 5′ UTR in the context of the tetO cassette could possibly account for increased UME6 transcript levels in the tetO-5′ UTR-UME6 versus tetO-UME6 strain. Given the importance of Ume6 for determining morphology and controlling filament- and virulence-specific gene expression, it is not surprising that C. albicans has evolved multiple mechanisms (transcriptional, translational and post-translational) to carefully adjust the levels of this regulator under a variety of different environmental conditions.
Interestingly, a recent whole-genome RNA-Seq analysis has indicated that, similar to UME6, a significant number of C. albicans genes involved in a wide variety of processes important for pathogenicity possess unusually long (> 500 bp) 5′ UTRs (Bruno et al., 2010). Many of these genes encode transcriptional regulators that control filamentous growth, biofilm formation, white-opaque switching and/or antifungal drug resistance (e.g. EFG1, CPH1, RFG1, CZF1, FKH2, SFL1, CRZ1, UPC2, GCN4, SIR2) (Liu et al., 1994; Stoldt et al., 1997; Brown et al., 1999; Perez-Martin et al., 1999; Sonneborn et al., 1999; Kadosh and Johnson, 2001; Bensen et al., 2002; Ramage et al., 2002; Tripathi et al., 2002; Onyewu et al., 2004; Silver et al., 2004; Santos and de Larrinoa, 2005; Li et al., 2007; Vinces and Kumamoto, 2007; Zordan et al., 2007). Several genes encoding adhesins of the α-agglutinin-like (ALS) family (ALS4, ALS5, ALS9), which are important for interaction of C. albicans with host cells (Hoyer, 2001), secreted aspartyl proteases (SAP1, SAP2) and lipases (LIP4, LIP8), important for the ability of C. albicans to degrade host cell membranes (Hube et al., 2000; Schaller et al., 2003), a superoxide dismutase (SOD4) involved in responding to oxidative stress in the host (Martchenko et al., 2004; Frohner et al., 2009), and a key regulator of iron-uptake genes and gastrointestinal commensalism (SFU1) (Chen et al., 2011) also have long 5′ UTRs. Finally, several genes that play critical roles in the mechanics of C. albicans hyphal development (HGC1, CDC24, RGA2) and cell cycle control (CLN3, CLB4) fall in this category as well (Bassilana et al., 2003; Zheng et al., 2004; 2007; Bachewich and Whiteway, 2005; Bensen et al., 2005; Chapa y Lazo et al., 2005; Wang, 2009). While not all of these genes may necessarily be controlled by a translational efficiency mechanism, the presence of a long 5′ UTR upstream of so many genes involved in virulence-related processes suggests, based on our findings, that this mechanism may play a significant role in controlling and fine-tuning C. albicans pathogenicity in the host.
There is also evidence to suggest that 5′ UTR-mediated translational efficiency mechanisms may play an evolutionarily conserved role in the regulation of morphogenesis and pathogenicity in non-albicans Candida species. Many of these species possess UME6 orthologues and the synteny of the long upstream intergenic region is conserved in Candida dubliniensis, Candida tropicalis, Candida parapsilosis, Candida guilliermondii and Candida lusitaniae (Candida Gene Order Browser, http://cgob.ucd.ie/), but has diverged in the non-pathogenic model yeast S. cerevisiae. In addition, a recent whole-genome RNA-Seq experiment has identified over 250 C. parapsilosis genes with 5′ UTRs greater than 500 bp in length, many of which appear to be involved in filamentous growth and pathogenicity (Guida et al., 2011) and several of which encode orthologues of the C. albicans genes with long 5′ UTRs discussed above (including UME6).
In bacterial and viral pathogens, 5′ UTR-mediated translational efficiency mechanisms have been shown to control a variety of biological processes, several of which are important for virulence. In Listeria monocytogenes, 5′ UTRs are important for controlling the translation of key virulence factors involved in the production of listeriolysin O (Johansson et al., 2002; Wong et al., 2004; Shen and Higgins, 2005). The Haemophilus influenzae sxy gene, which encodes an important regulator of DNA uptake, is also known to be translationally regulated by a 5′ UTR which forms an inhibitory secondary structure (Cameron et al., 2008). In several viral pathogens, including poliovirus and Hepatitis C virus, 5′ UTRs contain IRES sequences important for controlling translational efficiency and viral replication (Balvay et al., 2009). In higher eukaryotes, UTRs are known to mediate translational control of a wide variety of genes that function in diverse cellular processes, including cell cycle, stress response, oncogenesis, fertilization and development (Pickering and Willis, 2005; Chatterjee and Pal, 2009). 5′ UTRs have also been shown to affect the expression of genes associated with a number of human diseases including breast cancer, Alzheimer's disease, and bipolar affective disorder (BPAD) (Chatterjee and Pal, 2009). Our findings are significant because they suggest that in the major human fungal pathogen C. albicans, a 5′ UTR-mediated translational efficiency mechanism has evolved to inhibit and fine-tune morphogenesis, a key developmental process important for pathogenicity in the host environment.
Given that 5′ UTRs are likely to control the expression of a variety of important regulators of C. albicans morphology and/or virulence, what mechanisms are responsible for mediating translational control? How exactly do translational mechanisms modulate and/or fine-tune the expression of key virulence factors in response to host environmental conditions? Do certain components of the translation machinery respond to specific environmental cues? Is there cross-talk between translational mechanisms and known transcriptional and post-translational mechanisms which have previously been shown to control virulence properties? How and why did translational mechanisms apparently evolve to control so many genes associated with fungal pathogenicity? It is hoped that future research in this area will help to address these questions and shed more light on the important, but poorly understood, role that translational mechanisms may play in controlling a variety of processes important for fungal pathogenesis.
Strains and DNA constructions
A complete listing of strains used in this study is shown in Table S1. A detailed description of the plasmids and methods used to generate additional strains is provided in the Supplemental Materials and Methods section. All primers used for plasmid and strain constructions are described in Table S2.
Media and growth conditions
Standard non-filament-inducing growth conditions were YEPD (yeast extract-peptone-dextrose) medium at 30°C. Induction of filamentous growth by 10% serum at 37°C was performed as described previously (Banerjee et al., 2008). Spider and Lee's media were prepared as previously described (Lee et al., 1975; Liu et al., 1994). Induction of filamentous growth in Spider medium at 37°C was performed by first growing strains overnight in YEPD medium at 30°C. Cells were then washed and resuspended in 10 ml YEPD medium or Spider medium at a concentration of 1 × 106 cells ml−1. Ten microlitres of cells from each suspension were inoculated into 50 ml pre-warmed YEPD medium at 30°C or Spider medium at 37°C respectively. Cultures were grown for 36 h and cells were harvested for RNA extraction and microscopic analysis. DK318 was used as the wild-type control strain for the filament induction and polysome profiling experiments. The C. albicans morphological transition time-course experiment was performed by initially growing tetO-UME6 and tetO-5′ UTR-UME6 strains overnight at 30°C in YEPD medium + 1.0 μg ml−1 Dox to OD600 ∼ 0.5. Fifty millilitres aliquots of cells from each strain were next washed once in pre-warmed YEPD medium at 30°C and used to inoculate 1.5 l of YEPD medium in the presence or absence of 1.0 μg ml−1 Dox. Cultures were grown at 30°C and cells were harvested for RNA extraction at each hour for 10 h. Cells for the zero hour time point were collected from the tetO-UME6 overnight culture just prior to washing. For the 5′ RACE analysis, cells were induced with serum at 37°C as described previously (Banerjee et al., 2008) and harvested at 30 min (for identification of the −2126 transcript start site) or 1 h (for identification of the −1923 and −3041 transcript start sites) for total RNA and cDNA preparation. For the GFP expression experiment (Fig. 7, Figs S6 and S7) strains were grown overnight in YNB minimal medium at 30°C, diluted into fresh YNB minimal medium the next day, and grown for 3 h at 30°C, as described by Wolyniak and Sundstrom (2007); CAF2-1 was used as the wild-type control strain. For the Ume6-Myc expression experiment (Fig. 6) strains were grown overnight in YEPD at 30°C and diluted 1:10 into the indicated pre-warmed media. Cells for the 0 h time point sample of this experiment were harvested immediately prior to dilution.
RNA preparation and analysis
Total RNA preparation and Northern analysis were performed as described previously (Banerjee et al., 2008). Primers used for Northern probes are described in Table S2. RNA from the polysome profiling experiments was extracted with 1:1 phenol : chloroform, precipitated overnight at −20°C with two volumes of 70% ethanol in 2.5 M LiCl, washed with 70% ethanol and resuspended in nuclease-free water prior to qRT-PCR analysis (del Prete et al., 2007). RNA for qRT-PCR analysis and for 5′ RACE analysis was prepared using the SV Total RNA Isolation Kit (Promega) according to the manufacturer's directions with the following modification: cells were resuspended in 225 μl buffer RLT and placed in a bead beater for 2.5 min (yeast) or 5 min (hyphae); cells were rested on ice for 1 min per every 2.5 min in the bead beater.
5′ RACE analysis
5′ RACE analysis was performed using an Ambion FirstChoice® RLM-RACE kit (Applied Biosystems). Briefly, total RNA was first treated with calf intestinal phosphatase (CIP) to remove 5′ phosphates from rRNA, tRNA, degraded mRNA and genomic DNA. Next, the RNA was treated with tobacco acid pyrophosphatase (TAP) to remove cap structures from the full-length mRNAs. A 45 bp 5′ RACE adaptor was then ligated to the decapped mRNA. Following a random-primed reverse transcription reaction, the 5′ end of the UME6 transcript was then amplified by nested PCR. All 5′ RACE PCR products were run on 0.8% agarose gels and directly sequenced to determine size and identity.
Real-time quantitative RT-PCR analysis
cDNA for qRT-PCR analysis was prepared from 2 μg of total RNA treated with DNase I (Invitrogen) using an Applied Biosystems – High Capacity cDNA Reverse Transcription Kit, according to the manufacturer's instructions. Real-time PCR was performed in duplicate in 96-well plates using the Chromo4 Four-Color Real-Time PCR Detection System (Bio-Rad). PCR reactions were carried out in 25 μl volumes containing 5 μl 1:25 diluted cDNA (original cDNA volume was 20 μl), 12.5 μl GoTaq qPCR Master Mix (Promega) and 4.3 μl dH2O. Primers used for qRT-PCR analysis are described in Table S2. Real-time PCR was performed using the following cycling conditions: Step 1: 95°C for 2 min, Step 2: 95°C for 30 s, Step 3: annealing temperature (determined for each primer pair) for 1 min, Step 4: read plate, Step 5: repeat steps 2–4 for 39 times, Step 6: 72°C for 5 min, Step 7: Melting Curve 50°C–95°C every 0.4°C, hold 1 s and read plate. Standard curves were generated using seven serial dilutions of a pool of cDNA from each experiment to determine primer efficiency. Expression levels of each gene were normalized to levels of an internal ACT1 control using the Pfaffl method (Pfaffl, 2001). For the polysome profiling experiment, spike-in mRNA (Solaris) was added to each sample at a final concentration of 1× prior to RNA isolation and expression levels of UME6 were normalized to spike-in mRNA using the Pfaffl method.
Polysome profiling analysis
Cells were treated with 0.1 mg ml−1 cycloheximide and incubated on ice for 5 min. Next, cells were washed twice in lysis buffer (20 mM Tris-Cl, pH 8.0, 140 mM KCl, 1.5 mM MgCl2, 0.5 mM DTT, 1% Triton X-100, 0.1 mg ml−1 cycloheximide, 1 mg ml−1 heparin), lysed by vortexing with 2/3 volume beads for 4 × 20 s and centrifuged 5 min at 4500 r.p.m. at 4°C. Fifty OD260 units of supernatant were loaded on the top of a 10–50% continuous sucrose gradient and centrifuged at 35 000 r.p.m. for 160 min at 4°C in a SW41 rotor (Beckman Coulter). Fractions were collected manually in 1 ml aliquots for RNA isolation or 200 μl aliquots to monitor OD254 absorbance. RNA was isolated from each fraction for qRT-PCR analysis. For the EDTA release assay, following treatment with 0.1 mg ml−1 cycloheximide, 25 mM EDTA was added to the lysis buffer and sucrose gradient.
Protein isolation was performed as described previously (Cao et al., 2006). Twenty micrograms of total protein extract was separated by 8% (for Ume6-Myc) or 12% (for Actin) SDS-PAGE and transferred to a PVDF membrane (Invitrogen). Membranes were blocked with 5% milk in 1× PBS with 0.01% Tween-20, incubated with primary antibody for Myc (Cell Signaling Technology #2272) or actin (Sigma #A5060) overnight at 4°C, washed three times in 1× PBS with 0.01% Tween-20, then incubated with HRP-conjugated goat anti-rabbit secondary antibody (Zymed). The ECL system (GE Healthcare) was used for detection. Densitometry quantification of Western blots was performed using GelQuant.NET software provided by biochemlabsolutions.com (http://biochemlabsolutions.com/GelQuantNET.html).
Cellular fluorescence levels were quantified as described previously (Wolyniak and Sundstrom, 2007) using COSTAR 96-well plates and a Biotek Synergy 2 microplate reader. Plate wells contained 1.25 × 107 cells of each strain. Fluorometry assays were performed in biological triplicate and technical duplicate. Fluorescence values for each sample were normalized to cell density as determine by OD600.
We thank Brian Wickes and other members of the San Antonio Center for Medical Mycology for fruitful discussions and advice during the course of the experiments and for useful comments and suggestions on this manuscript. We especially thank Terri Kinzy (Rutgers Robert Wood Johnson Medical School) as well as Chris Browne and Andrew Link (Vanderbilt University Medical Center) for useful comments and suggestions regarding the polysome profiling analysis. We are grateful to Alistair Brown (University of Aberdeen, UK), Hironobu Nakayama and Mikio Arisawa (Nippon Roche Research Center, Kamakura, Japan), James Konopka (Stony Brook University) and Haoping Liu (University of California, Irvine) for plasmids and/or strains. We also thank Ian Morris for assistance with statistical analysis and microscopy and Haoping Liu and Yang Lu for assistance and advice with Western analysis. D.S.C. was supported by a COSTAR training grant (National Institute of Dental and Craniofacial Research Grant T32DE14318) as well as a Ruth L. Kirschstein National Research Service Award for Individual Predoctoral Fellows (National Institute of Dental and Craniofacial Research Grant F31DE021930). D.K. was supported by National Institute of Allergy and Infectious Diseases Grant 5RO1AI083344 in addition to a Voelcker Young Investigator Award from the Max and Minnie Tomerlin Voelcker Fund. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Allergy and Infectious Diseases, the National Institute of Dental and Craniofacial Research or the National Institutes of Health.