| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Previous Article | Next Article 
Eukaryotic Cell, October 2007, p. 1824-1840, Vol. 6, No. 10
1535-9778/07/$08.00+0 doi:10.1128/EC.00188-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
,
Microbiology and Molecular Pathogenesis Program, Dartmouth Medical School, Hanover, New Hampshire 03755
Received 23 May 2007/ Accepted 7 August 2007
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The hypha-specific gene HWP1 (hyphal wall protein 1) is an attractive candidate for studying morphology-specific gene regulation since it is abundantly expressed in hyphal growth mode and is barely detectable in the yeast morphology (73, 74). Pseudohyphae vary in the level of HWP1 gene expression (71). HWP1 encodes a glycosylphosphatidylinositol-linked cell wall protein that serves as a transglutaminase substrate and enables C. albicans to form covalent attachments to epithelial cells that line the buccal mucosa (73, 74). Recent findings have demonstrated the presence of a critical region of the HWP1 promoter that specifically confers developmental regulation and possesses distinct sets of protein/DNA interactions that are dependent on the morphology of the cell (29).
The regulation of HWP1 gene expression appears to be multifaceted and involves the Cph2/Tec1, Ras/cyclic AMP (cAMP), Bcr1/biofilm, and mitogen-activated protein kinase signal transduction pathways to various extents depending on growth medium (38, 39, 56, 69). A number of transcriptional activators, including Cph1, Efg1, Tec1, Bcr1, and Flo8, and transcriptional repressors, such as Nrg1 and Tup1, have been shown to be critical to both hypha-specific gene expression and hyphal development (11, 27, 54, 56, 64, 69, 79). With the exception of BCR1 (56), the null mutant phenotypes for these factors include aberrant growth morphologies in standard hyphal or yeast growth conditions, leaving open the possibility that the observed phenotype of each null mutant may be indirectly mediated through unrecognized factors that accompany morphogenesis. These findings prompt the need for experiments to investigate the possibility of a direct relationship between the structural changes in cell shape that occur during morphogenesis and HWP1 gene expression.
Polarization of cell growth and alterations to cytoskeletal dynamics are both highly regulated during morphogenesis. Recent work has shown that the Rho-like GTPase Cdc42 plays a central role in polarization and that C. albicans strains deficient for wild-type CDC42 are defective in the bud-hypha transition (50, 77). Strains carrying specific Cdc42 alterations in residue 26, 100, or 158 have revealed roles for CDC42 in both bud-hypha transitioning and hypha-specific gene expression (78). In terms of the actin cytoskeleton, yeast growth is characterized by a uniform cortical distribution of F-actin patches and cables, as shown by staining with rhodamine-conjugated phalloidin, whereas hyphal cells concentrate F-actin at hyphal tips (2). Extensive work with the budding yeast Saccharomyces cerevisiae has shown that actin structures nucleated by Arp2/3 and the formins, Bni1 and Bnr1, form extensive branched networks and linear filaments, respectively, and similar findings have been reported for yeast forms of C. albicans as well (15, 41, 46). These nucleators facilitate the localized rapid polymerization of F-actin, which allows for a high degree of regulation in the generation of new actin filaments at specific locations in the cell in response to the changing environment (52). The importance of actin cytoskeletal structure and dynamics to morphogenesis is shown by experiments in which treatment of C. albicans cells with the microfilament destabilizer cytochalasin A prior to growth in hypha-inducing conditions results in the complete abrogation of hyphal development while preserving cellular growth processes (1).
To better understand the regulatory connections that exist between the cytoskeletal and membrane rearrangements of the C. albicans bud-hypha transition and hypha-specific gene expression, we have utilized chemical-genetic and genetic approaches to analyze HWP1 expression in strains carrying a green fluorescent protein (GFP) reporter gene. Our findings reveal a coupling between the stability of the actin cytoskeleton and the up-regulation of HWP1 gene expression and suggest a new model for HWP1 regulation in which actin dynamics play a critical role in hypha-specific gene regulation at the C. albicans bud-hypha transition.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Nourseothricin (Werner Bioagents, Jena, Germany) was utilized for selection in YPD medium at a concentration of 450 µg/ml.
Drug treatments. The drug concentrations used for the Medium 199 experiments were as follows: cytochalasin A (Sigma-Aldrich, St. Louis, MO), 1 to 10 µg/ml; latrunculin A (Invitrogen, Carlsbad, CA), 10 to 50 µg/ml; jasplakinolide (Invitrogen, Carlsbad, CA), 1 to 20 µg/ml; benomyl (Sigma-Aldrich, St. Louis, MO), 2.5 to 50 µg/ml; propranolol (Sigma-Aldrich, St. Louis, MO), 125 to 175 µg/ml in accordance with previous work (8); brefeldin A (Sigma-Aldrich, St. Louis, MO), 10 to 50 µg/ml; ML-7 (Sigma-Aldrich, St. Louis, MO), 10 to 50 µg/ml. Drugs were dissolved in dimethyl sulfoxide as a concentrated stock solution and were added to Medium 199 or YPD plus 10% BCS at the appropriate concentration immediately before the addition of HB-12 cells.
Three methods were used to ensure that drug treatments did not compromise viability. First, methylene blue (0.01%, vol/vol, in PBS) staining revealed that a minimum of 80% of the cells survived each treatment. Second, C. albicans strain E-15, in which the constitutively active enolase promoter drives GFP expression (72), was exposed to the maximum utilized concentration of each drug. GFP fluorescence did not decrease after overnight exposure of E-15 to each drug under hypha-inducing conditions. Since the estimated half-life of GFP protein in yeast is approximately 7 h (47), the retention of GFP fluorescence was indicative of the continued activity of essential cellular processes following drug treatment. Third, 5 x 106 HB-12 cells were grown in Medium 199 with 10 µg/ml of cytochalasin A, latrunculin A, or jasplakinolide for 3 h at 37°C and 250-rpm shaking. Following this incubation, the cells were washed twice and resuspended in 10 µl PBS. The cells were diluted five times at a 1:10 ratio, and all six dilutions for each treatment were spotted onto a YPD plate. We found no difference in growth rates among any of the drug-treated cells in comparison to untreated control cells (data not shown).
The drug-induced effects were not growth medium dependent. Similar effects were observed when a different growth medium (YPD plus 10% BCS) was used. However, higher drug concentrations were required in media containing YPD plus 10% BCS, most likely because of the enhanced growth capacity in nutritionally richer media (24, 54). Since similar results were found with both growth media, only the results using Medium 199 are reported.
Assessment of HWP1 gene expression by fluorescence microscopy and fluorometry. Samples were processed for epifluorescence microscopy as previously described (75). Fluorescent images were visualized with a BX-60 microscope (Olympus, Center Valley, PA) using 600x magnification and a GFP filter with 470- to 490-nm excitation and 515- to 550-nm emission. Photomicrographs of cells expressing GFP were taken using a 7.6-second exposure with an Olympus MagnaFire camera and processed using MagnaFire version 2.1 (Optronics, Inc., Goleta, CA) and Image-Pro Plus version 4.1 (Media Cybernetics, Inc., Silver Spring, MD) software. The selected images represented the appearance of the majority of cells in a given population.
Quantitation of fluorescence was performed using a Wallac 1420 Victor2 plate reader (PerkinElmer, Inc., Wellesley, MA). After incubation and washing as described above, 1.25 x 107 cells in 100 µl PBS were resuspended and transferred to a Falcon Microtest 96-well plate (Becton Dickinson Labware, Franklin Lakes, NJ). Assays were performed in duplicate. Four separate readings for optical density (OD) at 600 nm and fluorescence (485-nm excitation and 535-nm emission) were collected for each sample. Normalization of fluorescence data with respect to cell density was accomplished by dividing the fluorescence value by the optical density for each individual reading.
Indirect immunofluorescence assay to detect Hwp1 and rhodamine-phalloidin staining to detect actin. To examine Hwp1 localization, C. albicans cells grown in hypha-inducing conditions were washed twice in ice-cold PBS and resuspended to a concentration of 3 x 107 cells/ml. A 200-µl sample of this resuspension was placed in an Eppendorf tube, and anti-Hwp1 antibody was added at a 1:200 dilution. The cells were incubated at 37°C for 30 min with gentle shaking, washed three times, and resuspended in 200 µl of ice-cold PBS. Alexa Fluor 488 goat anti-rabbit antibody (Molecular Probes, Eugene, OR) was then added at a 1:50 dilution, and the cells were incubated for 30 min at 37°C with gentle shaking. After three additional washes and resuspensions in ice-cold PBS, slides were prepared and visualized using the procedures described above for fluorescence microscopy (75). Images were taken using a 1.2-second exposure.
Actin structures were visualized using rhodamine-phalloidin staining, which was performed as previously described (57) using a rhodamine filter set (510- to 550-nm excitation, 570- to 590-nm emission) and a 2.4-second exposure.
Quantitation of F-actin aggregates. The total level of F-actin at the early stages of yeast and hypha development was determined using ImageJ 1.38x software (NIH, Bethesda, MD) in the manner previously described (28). Ten yeast and hyphal cells were analyzed for each individual time point and statistically analyzed as described below.
Disruption of the C. albicans BNI1 gene. The BNI1 knockout strain MWY1 was constructed using a NAT1 flipper cassette that conferred resistance to nourseothricin. Upstream and downstream regions (300 base pairs in length) of genomic DNA relative to the BNI1 coding region were amplified using primer sets BNI128N and BNI128C, respectively (Table 2), and genomic DNA as the template. The upstream region PCR product and pJK863, containing the NAT1 flipper cassette (70), were digested with KpnI and ApaI and gel purified (QIAGEN, Valencia, CA). After treatment of the vector with calf intestinal alkaline phosphatase (Fisher Scientific, Pittsburgh, PA), the two fragments were joined using T4 DNA ligase (Invitrogen, Carlsbad, CA) and transformed into TOP10 Escherichia coli cells (Invitrogen, Carlsbad, CA). Plasmid DNA was purified from transformants and verified by DNA sequencing. The PCR product corresponding to the region downstream of the BNI1 open reading frame and the plasmid containing the upstream region were digested with SacI and SacII prior to ligation and were transformed into TOP10 E. coli cells as described above.
|
A nourseothricin-sensitive BNI1/bni1
strain was created by excision of the NAT1 cassette using standard techniques (70). Excision of the NAT1 cassette was confirmed by Southern blotting using EcoRV digestion and a PCR-amplified complete genomic copy of BNI1 as the probe (not shown).
For the second round of BNI1 deletion, oligonucleotides that were slightly displaced from the genomic location of the original deletion to prevent recombination between the new flipper cassette and the regions flanking the excised copy of BNI1 were designed. A PCR using primer set BNI518N generated the KpnI-ApaI upstream fragment of BNI1, whereas primer set BNI518C generated the SacII-SacI downstream fragment of BNI1. Ligation of these two fragments to the ends of the NAT1 flipper cassette generated plasmid pBNIDEL518. Deletion of the second copy of BNI1 was performed as described above.
Cloning and disruption of C. albicans MEA1. The Aspergillus nidulans MesA amino acid sequence (60) was used as a query to search the predicted open reading frames in the C. albicans genome (4) using BLASTP, leading to the identification of ORF19.5919. Global alignment using Enboss Needle at EMBL-EBI (http://ebi.ac.uk) showed 57.4% similarity and 20.7% identity between the two sequences. In four previously identified conserved regions (60), homology was as high as 65% similarity and 47% identity (see Fig. S1 in the supplemental material for complete alignment analysis). The gene containing ORF19.5919 was named MEsA1 or MEA1. A 2,482-bp DNA segment containing MEA1 was synthesized by PCR using primer set MEA366 and genomic DNA as the template. This PCR product, which included 303 bp of upstream flanking sequence, 1,879 bp of coding sequence, and 300 bp of downstream flanking sequence, was sequenced and cloned into the pBluescript SK(–) vector to generate pBSMEA366. The MEA1 deletion strain MWY2 was generated in the manner described above for the deletion of BNI1. Primer sets MEA500N and MEA500C were utilized to generate MEA1 flanking regions for ligation to the NAT1 flipper cassette and deletion of the first copy of MEA1, whereas primer sets MEA700N and MEA700C were used for deletion of the second copy. pBSMEA366 was used as template for the PCRs.
Insertion of the HWP1 reporter plasmid into bni1/bni1, bnr1/bnr1, and pfy1/pfy1 deletion strains and Cdc42 point mutant strain AV03.
The nourseothricin resistance marker was inserted into p1902GFP3 prior to transformation into the bni1/bni1, bnr1/bnr1, and pfy1/pfy1 deletion strains as well as the Cdc42 point mutant strain AV03. Plasmid JK795 (70), containing the complete NAT1 genomic region, was used as the template for primer set NAT1. The resulting PCR-amplified fragment contained the complete NAT1 genomic region flanked by SacI and SapI restriction sites at its 5' and 3' ends, respectively. This fragment and p1902GFP3 were both digested with SacI and SapI and ligated together to generate p1902GFP3NAT1, an HWP1 reporter plasmid that contains URA3 and NAT1 markers on opposite sides of HWP1-driven GFP. This plasmid was targeted to the enolase locus of each C. albicans strain of interest following ClaI linearization. Each strain was transformed via electroporation, selected on YPD plates containing nourseothricin, and verified via Southern blotting using BglII digestion and genomic enolase as a probe to recognize corresponding sequences present in p1902GFP3NAT1 (not shown).
Examination of HWP1 gene expression in the CDC42 inducible deletion strain.
The p1902GFP3NAT1 plasmid was transformed by electroporation into SMC7A and SMC8A (50) using the above methods. Transformants were selected on YPD plates containing nourseothricin and confirmed by Southern blotting using BglII digestion and genomic enolase as a probe (29) (not shown). Overnight growth in yeast carbon base-bovine serum albumin media led to the excision of the ectopic copy of CDC42 in SMC7A localized to the ACT1 locus in the manner previously described for excision of the BNI1 flipper cassette to create a cdc42/cdc42 strain. Analysis of the induced cdc42/cdc42 strain was carried out under hypha-inducing conditions as described above.
Kinetic analysis of HWP1 gene expression. To examine HWP1 gene expression and actin structures during the development of yeast buds and hyphal germ tubes, strain HB-12 was grown under yeast- and hypha-inducing conditions as described above. Cell samples were taken at 15-min intervals over a 120-min period and examined for morphology, HWP1 gene expression, and actin by differential interference contrast (DIC) and epifluorescence microscopy. To quantitate HWP1 up-regulation rates, strain HB-12 with or without 1, 5, or 10 µg/ml cytochalasin A or 10 µg/ml jasplakinolide was analyzed every 10 min over a 150-min period by fluorometry. Strains MWY1 and AV03 were analyzed in a similar manner. Experiments were performed in triplicate.
Addition of exogenous cAMP to actin-deficient cells. Strain HB-12 was induced to form hyphae in the presence or absence of 5 µg/ml or 10 µg/ml cytochalasin A and 10 mM dibutyryl-cAMP (Roche Diagnostics, Basel, Switzerland). Measurements were taken every 30 min over a 150-minute time period. Strain MWY3 was analyzed in a similar manner. Experiments were performed in triplicate and statistically analyzed as described below.
cAMP assay. Intracellular cAMP in Medium 199 was extracted as previously described (16) and measured using the cAMP Biotrak enzyme immunoassay system according to the manufacturer's instructions (GE Healthcare, Piscataway, NJ). Cells were prepared for hypha formation in the manner described above, and samples were taken at 30-minute intervals over a 120-min time period. Experiments were performed in triplicate and statistically analyzed as described below.
Statistical analysis. Unless otherwise indicated, statistical analysis was performed using the one-way analysis of variance and Bonferroni's multiple-comparison test performed with Prism 4.03 software (GraphPad, San Diego, CA).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
By contrast, abrogation of hypha formation by the F-actin stabilizer jasplakinolide (10, 68) and the reversible disruptor of Golgi complex-mediated secretion brefeldin A (43) was not accompanied by the abolition of HWP1 gene expression in 
90% of cells (Fig. 1C). Experiments using S. cerevisiae have shown that jasplakinolide causes the formation of actin clumps that are extremely stable, cannot be dispersed by the monomer sequestering agent latrunculin A, and are therefore not dynamic (5). Nonetheless, the actin structures stabilized by jasplakinolide are structurally and functionally similar to normal F-actin in that they continue to associate with proteins that normally associate with F-actin and bind anti-actin antibodies (5). As expected, stabilized actin clumps were observed in C. albicans treated with jasplakinolide when detected with rhodamine-conjugated phalloidin (Fig. 2A).
|
Detection of Hwp1 on drug-treated cells using indirect immunofluorescence. To determine if Hwp1 protein was detectable on cell surfaces, drug-treated cells of strain HB-12 were examined for the presence of Hwp1 protein by the indirect immunofluorescence assay using monospecific anti-Hwp1 antibody. Hwp1 was found exclusively on the surfaces of hyphae in almost all untreated cells (Fig. 3, panel 2) and at discrete regions on the surfaces of the vast majority of cells treated with jasplakinolide or brefeldin A despite the lack of a discernible hyphal structure (Fig. 3, panels 6, 8, and 10). In the case of jasplakinolide-treated cells, multiple polarization sites containing Hwp1 could be detected, suggesting a loss in the ability to organize and commit to a single site of hypha emergence.
|
50% of cells that were blocked in hypha formation by cytochalasin A (Fig. 3, panel 4). Although GFP reflecting HWP1 gene expression was not evident in the merged epifluorescence and DIC image (Fig. 3, panel 3), GFP was detectable at a low level in the dark-field fluorescent image of cytochalasin A-blocked cells compared to yeast form cells (compare Fig. 1B, panel 6, and 1A, panel 2). Similar results were also found for latrunculin A-treated cells (not shown). In both cases, the majority of cells that showed the presence of Hwp1 patches had a slightly oval or pear-shaped appearance, a shape that is characteristic of cells immediately prior to germ tube emergence. These results showed that the processes controlling polarization of cell surface proteins in hypha-inducing conditions were retained during drug treatments that inhibited germ tube elongation and revealed the presence of a low level of HWP1 gene expression under these conditions that was barely visible by epifluorescence microscopy to detect the GFP reporter. Reduced HWP1 gene expression in mutants defective in F-actin nucleation. The opposite effects of jasplakinolide treatment versus cytochalasin A or latrunculin A treatment on HWP1 gene expression as assessed by measuring GFP production pointed to the possibility that actin dynamics plays a pivotal role in HWP1 expression. The observation that cells treated with jasplakinolide, which stabilizes actin and promotes the formation of polymeric actin structures (5), retained HWP1 gene expression despite the lack of true hyphal development suggested that F-actin in cells undergoing bud-hypha transitions may be important for the up-regulation of HWP1. To obtain genetic evidence in support of this hypothesis, mutants with defects in actin cable nucleation were analyzed for the presence of reduced HWP1 gene expression.
Previous work with S. cerevisiae and C. albicans has demonstrated the importance of the formin gene BNI1 for efficient actin cable formation and for C. albicans hyphal development (15, 41, 46). Therefore, a bni1/bni1 mutant strain (MWY1) containing p1902GFP3NAT1 was created. Upon growth under hypha-inducing conditions, the majority of MWY1 cells produced hyphae that were wider and less consistent in width than wild-type hyphae (Fig. 4A). MWY1 also showed a reduction in HWP1 gene expression, which was 63% of control levels as detected by fluorometry (Fig. 4A and C). Identical results were obtained using a bni1/bni1 strain produced in Yue Wang's laboratory (Institute of Molecular and Cell Biology, Singapore) (41).
|
/bni1 background and is consistent with the predicted effect of a loss of interaction between Cdc42 and Bni1 (Fig. 4C). The reductions in fluorescence of the bni1/bni1 and AV03 strains transformed with p1902GFP3NAT1 compared to HB-12 were statistically significant.
The effect of the formins on HWP1 gene expression was restricted to BNI1. The bnr1/bnr1 knockout did not deviate from control cells in hypha formation or in the ability to activate HWP1 (Fig. 4C). Likewise, the actin monomer recruiter profilin was not required. C. albicans profilin has been shown to be functionally similar to S. cerevisiae profilin (59), and the two proteins share extensive sequence homology (87% similarity, 72% identity). The pfy1/pfy1 strain showed alterations to hyphal structure similar to those observed in the bni1/bni1 strain (not shown); however, HWP1 gene expression was not reduced in pfy1/pfy1 cells (Fig. 4C).
The genetic studies were consistent with the drug studies in that reduced levels of HWP1 gene expression were found using two mutants that are deficient in F-actin nucleation.
Kinetics of expression of the HWP1 gene relative to actin polarization. To investigate the reasons for the differences in GFP levels at 2.5 h between cells treated with jasplakinolide and those treated with cytochalasin A or latrunculin A, the kinetics of increases in GFP reporter expression over time during germ tube inducing conditions were assessed.
(i) Microscopic analysis. The actin cytoskeleton and GFP fluorescence reflecting HWP1 gene expression were analyzed at 15-min intervals in cells grown under hypha-inducing or yeast growth conditions (Fig. 5A). In hyphal cells at 30 min postinduction, F-actin began to polarize to a site of germ tube emergence but GFP fluorescence was not observed (Fig. 5A, panels 7 to 9), confirming work from the Liu laboratory (24).
|
(ii) Quantitation of F-actin in yeast and hyphal growth. Previous work by Hazan et al. (24) suggested the presence of permanently polarized actin at the point of germ tube emergence that was not dispersed during the G1 phase of the yeast cell cycle, when cell cycle-regulated actin became depolarized. This prompted questions regarding whether or not this apical F-actin constituted an absolute increase in the amount of F-actin during the emergence of germ tubes compared to that during bud emergence. To address this question, we quantified the total cellular F-actin from yeast and hyphal cells during the first 60 min of induction according to previously described procedures (28). Although a slight increase in F-actin levels in hyphal cells relative to yeast cells was observed at 60 min, this difference was not statistically significant (Fig. 5B), suggesting that the observed aggregation of F-actin to both the nascent germ tube and the nascent bud in their respective growth conditions represents a redistribution of existing cellular F-actin populations in response to polarization signals.
(iii) Slow and fast kinetics of HWP1 gene expression during hypha induction. Previous work showed that the rate of HWP1 gene expression increased after 60 min of induction (29). To explore the kinetics of HWP1 expression within the first 2.5 h of induction, GFP fluorescence levels were monitored by fluorometry at 10-minute intervals. Two distinct rates of increasing fluorescence were observed; these reflected the presence of a slow phase followed by a fast phase, with average rates of 135.7 and 362.7 fluorescence units/min, respectively (Fig. 6A). The difference between the slow and fast rates of expression was statistically significant (P < 0.001 using the paired Student t test on data from three independent experiments). The transition period at 50 to 70 min postinduction roughly correlated with both the appearance of intense actin polarization at the emerging hypha (Fig. 5A) and the spike in cAMP that naturally occurs following the shift to hypha-inducing conditions (7).
|
The kinetics of HWP1 gene expression were also altered in mutants with deficiencies in actin nucleation. The bni1/bni1 and Cdc42 E100G strains showed similar kinetics to HB-12 during the slow phase of expression (Fig. 6C). Although the bni1/bni1 strain showed an increase in HWP1 gene expression during the fast phase, this increase was reduced relative to HB-12. The bni1/bni1 strain appeared to enter the fast phase of HWP1 gene expression, but the fast phase was not maintained. The Cdc42 E100G strain showed slight increases in HWP1 expression levels that were markedly reduced in comparison to those shown by HB-12.
The results of the kinetic analysis suggested that activation of HWP1 gene expression occurs by a two-step process consisting of a slow actin-independent phase followed by a fast F-actin-dependent phase. While GFP reporter levels reflecting HWP1 activation are measurable by fluorometry during the slow phase of expression, it is only after 50 to 70 min of induction, when actin is polarized to the site of apical growth in the majority of cells (Fig. 5A), that rapid up-regulation of HWP1 occurs and GFP is readily visualized by epifluorescence microscopy of single cells. Only the fast phase is disrupted by the use of chemical or genetic disruptors of F-actin, implicating F-actin as a central player in the fast phase of HWP1 expression. Although GFP levels are slightly reduced at each time point in jasplakinolide-treated cells in the fast phase, the kinetics of the fast phase of activation are neither diminished nor enhanced by jasplakinolide, which stabilizes and extends existing filaments and causes filament nucleation de novo (68).
HWP1 expression and localization in the absence of polarization. (i) CDC42. CDC42 is an essential gene in S. cerevisiae and C. albicans (26, 51). However, the loss of CDC42 can be studied by using strain SMC7A, in which CDC42 can be inducibly deleted following growth in yeast carbon base-bovine serum albumin medium and subsequent excision of an ectopic copy of CDC42 from the genome (50). Loss of CDC42 gave rise in most cases to large, round cells that had lost the ability to exhibit polarized growth and were terminal (Fig. 7A). Measurable GFP fluorescence was found under hypha-inducing conditions in strain SMC7A, which had been transformed with p1902GFP3NAT1, despite the inability of the cells to form true hyphae. This result added additional genetic evidence for the independence of the regulatory pathways governing the formation of true hyphal structures and HWP1 expression. Indirect immunofluorescence assays showed that Hwp1 was present on cell surfaces but was not confined to specific areas; rather, Hwp1 was extended across the entire cell surface (Fig. 7A). Polarization of the cell surface was not required for high levels of HWP1 expression, which could be determined by measuring GFP fluorescence.
|
Expression of HWP1 by stabilized actin involves the cAMP pathway. (i) Studies using cAMP hypo- and hypermorphic strains of C. albicans.
Recent work using S. cerevisiae endocytosis mutants that display reduced actin dynamics and the presence of F-actin aggregates has shown that an up-regulation of the cAMP signaling pathway is associated with F-actin aggregation and that the adenylyl cyclase-associated protein, Srv2p (also called Cap1), is required for increases in cAMP (21). To investigate whether HWP1 gene expression in the presence of stabilized actin structures was also related to activation of the cAMP pathway, GFP fluorescence and actin structures were monitored in a strain that constitutively overexpresses PDE2 (EPDE2-3) (6), causing enhanced degradation of cAMP. The HWP1GFP reporter gene was introduced into strain EPDE2-3 using plasmid p1902GFPNAT1. Overexpression of PDE2 eliminated observable GFP fluorescence among jasplakinolide-treated cells despite the presence of stabilized actin structures (Fig. 8A, panels 1 and 2). HWP1 gene expression in the presence of jasplakinolide was also not observed in the srv2/srv2 mutant strain, which is deficient in the activation of the cAMP pathway (7) (Fig. 8A, panels 3 and 4).
|
/srv2 and EPDE2-3 strains (Fig. 8B, panels 2 and 4), arguing against a role for the cAMP signaling pathway in the initial stage of HWP1 gene expression. However, the data strongly suggest that activation of the cAMP signaling pathway is related to the expression of HWP1 to the high levels required for detection via GFP reporter fluorescence in the presence of stabilized actin. (ii) The effect of chemical or genetic interference with F-actin on cAMP levels. If the loss of HWP1 gene expression in cytochalasin A- and latrunculin A-treated cells is related to reduced cAMP levels, then cells treated with these drugs should have reduced levels of cAMP compared to untreated cells. Furthermore, mutants with defects in actin nucleation should also exhibit reduced levels of cAMP. The addition of 10 µg/ml cytochalasin A to HB-12 cells reduced intracellular levels of cAMP to one-half the level seen in untreated cells following 60 min of induction, a time that coincides with the slow/fast phase transition (Fig. 8C). A gradual decline in cAMP levels followed the 60-minute time point. By contrast, latrunculin A (50 µg/ml) did not alter the peak of cAMP at 60 min; however, a reduction in intracellular cAMP to 66% of untreated-cell levels by 120 min postinduction was observed. The difference in magnitude between the cytochalasin A and latrunculin A results up to 90 min postinduction could be related to the slow monomer-sequestering mechanism by which the latter acts on actin filaments (13). Also, the disruption to the actin cytoskeleton by latrunculin A-based actin monomer sequestration may result in an overall decline of cAMP production in latrunculin A-treated cells despite the initial increase. This decline, in turn, would be reflected in a decline in HWP1 gene expression (Fig. 1B, panels 7 to 12).
By contrast, jasplakinolide treatment (10 µg/ml) led to cAMP levels that were 75% of those in untreated cells at 60 min and surpassed levels found in untreated cells at 120 min postinduction, demonstrating a sustained gradual increase over time in cAMP levels in response to jasplakinolide-mediated stabilized actin. The sustained increase in cAMP levels seen in jasplakinolide-treated cells with stabilized actin was similar to the sustained increase previously observed in a pde2/pde2 strain in which cAMP degradation is inhibited (6).
Examination of intracellular cAMP levels in the bni1/bni1 and Cdc42 E100G strains revealed a loss in the peak of cAMP production that occurred in strain HB-12 at 60 min postinduction (Fig. 8D). At the 60-min time point, the bni1/bni1 and Cdc42 E100G strains showed cAMP levels at 46% and 26%, respectively, of the peak cAMP levels observed from HB-12. By 120 min postinduction, however, the cAMP levels in the actin-compromised mutant strains were close to HB-12 levels.
(iii) Partial recovery of HWP1 gene expression by exogenous cAMP.
The addition of 10 mM dibutyryl-cAMP to HB-12 cells treated with 5 µg/ml and 10 µg/ml cytochalasin A resulted in 1.5- and 1.75-fold increases in HWP1-derived GFP fluorescence following 150 min of induction, to levels that were 38% and 18% of HB-12 levels, respectively (Fig. 8E). The results were statistically significant and suggested that cAMP was able to partially rescue the effects of cytochalasin A concentrations that normally curtail hypha formation and HWP1 gene expression. Higher concentrations of dibutyryl-cAMP did not further increase GFP reporter fluorescence. Treatment of the srv2/srv2 strain with 10 mM dibutyryl-cAMP yielded a twofold increase in HWP1 gene expression compared to that in untreated cells, although after partial rescue levels were only 13% of those seen in untreated HB-12 cells (Fig. 8F). These partial recoveries of HWP1 gene expression by dibutyryl-cAMP support the relationship between actin cytoskeletal aggregates and activation of the cAMP pathway.
Together, the data show that conditions that interfere with F-actin formation also affect the kinetics of cAMP levels in wild-type cells, which peak at 60 min when polarized F-actin becomes present in cells prior to germ tube emergence, beginning at 30 min (Fig. 5A). The results suggest a correlation between the presence of stabilized actin and increasing cAMP levels during the slow phase of HWP1 gene expression, which precedes the rapid phase of induction. The results also underscore the importance of rapid cAMP production at the bud-hypha transition prior to the fast phase of HWP1 gene expression and the strong contribution of stabilized F-actin to the development and maintenance of this cAMP activation. F-actin aggregation alone was not sufficient to override mutations in the cAMP pathway, and the disruption of actin aggregation resulted in marked declines in intracellular cAMP levels among cytochalasin A-treated cells as well as actin-compromised mutant strains.
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
It is important to acknowledge the potential difficulties with interpreting results from drug studies. Cytochalasin A is chemically similar to the thiol-alkylating agent N-ethylmaleimide (NEM) and has secondary effects on cell metabolism and thiol alkylation in mammalian cells that are unrelated to effects on actin (18, 32, 35). In this study, unlike studies with mammalian cells in which NEM did not interfere with actin filaments (18), NEM did in fact interfere with actin polymerization in C. albicans (not shown). The differential effect of NEM on actin from C. albicans compared to mammalian cells points to an intrinsic difference in sensitivity to oxidation between the two actins in their respective environments that may reflect the observation by us and others (32) that yeast and fungi are sensitive to cytochalasin A but not to cytochalasin B. Support for a role for oxidation in actin depolymerization is provided by recent studies showing that oxidation of mammalian actin by peroxide can cause a complete loss of polymerization potential, an effect involving cysteine 374 (40), which is conserved in C. albicans actin. Thus, the influence of thiol alkylation on hyphal development and HWP1 expression does not preclude the correlation between F-actin and HWP1 expression.
Consistent with the importance of reduced sulfhydryl groups in F-actin function, the simultaneous addition of dithiothreitol (DTT) (1 mM) and cytochalasin A (0.02 mM) prevented the complete depolymerization of actin and partially rescued HWP1 gene expression (not shown). GFP levels attained 70% of the level seen in untreated cells in the presence of DTT and cytochalasin A, which was equivalent to GFP levels in the presence of jasplakinolide. Importantly, DTT alone also blocked germ tube formation, but actin was not depolarized or depolymerized and HWP1 gene expression was again retained at the level of 70% of that seen in untreated cells. The effect of DTT alone on blocking germ tube formation could be explained by the selective retention of secretory proteins in the endoplasmic reticulum (25). The effect of DTT was similar to that of brefeldin A in preventing germ tube formation but not actin polymerization or HWP1 gene expression (see below). Thus, the results of germ tube blockage using NEM, DTT, and DTT plus cytochalasin A were consistent with the correlation between the absence or presence of F-actin and the loss or presence, respectively, of HWP1 gene expression as assessed by measuring GFP fluorescence using epifluorescence microscopy.
The importance of F-actin in HWP1 gene expression was not solely based on treatments with cytochalasin A. Latrunculin A, an actin filament destabilizer like cytochalasin A, produced identical losses in hypha formation and HWP1 gene expression as cytochalasin A but required higher drug concentrations (0.12 mM versus 0.02 mM) to attain them, a fact reflective of the monomer-sequestering mechanism by which latrunculin A functions (13). Jasplakinolide, an actin filament stabilizer, produced the opposite effect of the actin-destabilizing drugs. The correlation between F-actin and HWP1 gene expression was also supported by genetic studies and kinetic analyses discussed in the previous section as well as below.
The effect of brefeldin A on hypha formation and HWP1 gene expression is also consistent with a central role for stabilized F-actin in the regulation of both processes. Brefeldin A acts as a specific disruptor of Golgi complex-mediated secretory processes (9, 43), while leaving other cellular processes, such as those involving the actin cytoskeleton, unaltered (30). Whereas recent work has revealed that Bni1-mediated actin cables are necessary for positioning the Golgi complex to a putative site of germ tube emergence and for coordinating the transport and deposition of membrane and cell wall material to a growing hypha (65), actin dynamics would not be expected to change in response to the bud-hypha transition in cells with a chemically dispersed Golgi complex. However, we found that the loss of Golgi complex-mediated secretion prevented hypha development in cells treated with brefeldin A, implicating the Golgi complex as a central player in bud-hypha transitions. These results vary from the recently published findings of Rida et al. (65), since brefeldin A was added to cells simultaneously with the shift to hypha-inducing conditions in our experiments rather than allowing hyphal structures to develop prior to drug addition as in the study by Rida et al. (65). Addition of brefeldin A during the bud-hypha transition would not be expected to interfere with actin dynamics and subsequent expression of HWP1. However, the inability of these cells to form a coherent Golgi complex at the site of hypha formation (65) prevented hypha development, resulting in cells with a yeast-like morphology despite the expression of hypha-specific genes. The results using brefeldin A-treated cells further support the importance of hypha-inducing environmental conditions in HWP1 expression as well as the idea that the yeast morphology is not incompatible with the type of actin dynamics that can promote cAMP pathway up-regulation, leading to the expression of genes that are usually hypha specific.
The effects of benomyl, propranolol, and ML-7 on hyphal development and HWP1 gene expression may also be examined within the context of F-actin. It has been previously shown that the actin cytoskeleton is disrupted along with microtubules upon benomyl treatment (1), which accounts for the cytochalasin A- and latrunculin A-like phenotype observed in our benomyl experiment. Actin bundle disruption and filament depolymerization have also been observed in connection with the use of propranolol (23, 55) and ML-7 (36, 53), suggesting that the stabilization of F-actin may be a common explanatory thread linking the results of all seven of the drugs analyzed in this study.
The contribution of F-actin structures to HWP1 gene expression is supported by genetic data. First, in terms of the ability to activate HWP1, strains lacking the formin gene BNI1 differed markedly from strains lacking the profilin gene PFY1 despite strong similarities in their morphological defects to hyphal cells. Specifically, HWP1 gene expression was decreased in bni1/bni1 cells and unaffected in pfy1/pfy1 cells. Recent work has shown that Bni1 is localized primarily to the developing hypha tip and that its presence is necessary for the development of a wild-type hyphal structure though not for overall apical growth (41). The reduction in HWP1 gene expression in the bni1/bni1 mutant could be related to a compromise to the formin-mediated actin cytoskeletal network that is essential for wild-type hyphal development. While not preventing hyphal morphogenesis per se, the loss of BNI1 may prevent the synthesis of appropriate F-actin structures necessary at the bud-hypha transition for the formation of a cell exhibiting wild-type levels of hypha-specific gene expression (41). Therefore, the finding that deletion of BNI1 compromises HWP1 expression both in the strain generated in this work and the strain generated by Li et al. (41) provides direct genetic evidence supporting the drug studies that specific F-actin structures contribute to HWP1 gene expression. By contrast, loss of Pfy1 would reduce the efficiency of actin rearrangements during hyphal morphogenesis but not preclude them, allowing HWP1 gene expression to persist in a pfy1/pfy1 null mutant.
The importance of stabilized F-actin to the HWP1-regulatory process was also supported by the reduction in HWP1 gene expression in a strain carrying a specific glutamic acid-to-glycine alteration (Cdc42 E100G) encoded by CDC42 (78) that leads to deficient interactions with Bni1 (51). Work with S. cerevisiae has implicated Cdc42 as a direct mediator of actin cytoskeletal dynamics during polarized morphogenesis through its physical interactions with Bni1 (14). More-recent work suggests that this Cdc42-Bni1 interaction exists in C. albicans as well and that compromises to this interaction have deleterious effects on hypha formation (78). The level of HWP1 gene expression in the Cdc42 E100G strain compared to that in HB-12 (47%) mirrored the level observed in the bni1/bni1 strain (63%), which is consistent with the idea that formin-based actin filament rearrangements play a critical role in HWP1 expression. Assessing further compromises to HWP1 expression by the simultaneous deletion of both C. albicans formins is not possible because C. albicans bni1/bni1 bnr1/bnr1 cells are inviable (41).
Although the exact nature of the native actin structures that may support the expression of HWP1 during the bud-hypha transition is not known, it is possible that F-actin aggregates that are similar to those observed in a fraction of the cells in stationary-phase cultures of S. cerevisiae may be present (20) except that the aggregated F-actin in germ tubes is polarized at the apex of the emerging germ tube. The possibility that overall increases in F-actin levels are involved is considered to be unlikely based on the low level of G-actin reserves in S. cerevisiae, which have been measured to be 1% of total actin (5). Indeed, our quantitative analysis did not reveal a difference between the level of total F-actin in cells forming nascent germ tubes or nascent buds. An alternative possibility is that F-actin from cortical patches coalesces and is incorporated into hypha-specific F-actin structures in the emerging germ tube in a manner analogous to the convergent-evolution model of filopodial initiation (76). F-actin aggregates could also arise following the dissolution of preexisting F-actin and nucleation of new structures that would support HWP1 gene expression. The presence of polarized F-actin that is specific to hyphal growth and is not influenced by cell cycle cues has been previously described (24). Concentrated F-actin (2) at hyphal tips may be functionally similar to jasplakinolide-induced structures, a fact that may account for the effect of jasplakinolide on HWP1 gene expression despite the lack of germ tube development that occurs upon jasplakinolide treatment.
A possible mechanism by which F-actin contributes to HWP1 gene expression is through involvement of the cAMP signaling pathway. The importance of activation of the cAMP signaling pathway to HWP1 up-regulation was demonstrated in our previous work using hypo- and hypermorphic cAMP mutants (6). Here, multiple approaches pointed to a connection between F-actin and cAMP. First, the cAMP increase that normally occurs and peaks within 60 min of hypha induction was largely abrogated in the presence of cytochalasin A treatment as well as in mutants that are deficient in the nucleation of F-actin. Although latrunculin A did not inhibit the cAMP peak, cAMP levels during the fast phase of HWP1 gene expression were significantly reduced compared to those in untreated cells after 120 min. Second, measurements of intracellular cAMP in cells treated with 10 µg/ml jasplakinolide revealed a gradual but sustained increase in cAMP levels during hypha induction that did not decline over time but rather continued to increase during the fast phase of HWP1 gene expression. Third, the absence of Srv2 as well as the overexpression of PDE2, which encodes the high-affinity phosphodiesterase that degrades cAMP, led to the loss of HWP1 gene expression in the presence of jasplakinolide-stabilized actin structures. The absence of HWP1 gene expression despite the presence of jasplakinolide-stabilized actin in hypha-inducing conditions was consistent with the idea that elevated levels of cAMP are responsible for HWP1 gene expression in association with F-actin. A similar experiment using overexpression of the PDE2 gene was used in S. cerevisiae to demonstrate the importance of cAMP signaling in the presence of stabilized actin (20). Fourth, the addition of dibutyryl-cAMP partially reversed the genetic and chemical-genetic interference with F-actin polymerization.
The data from this study suggest that the mechanisms leading to HWP1 gene expression share similarities to the F-actin-related increases in release of ROS from mitochondria that precede apoptosis in S. cerevisiae. Mutations in the sla1 or end3 genes (21) interfere with the actin dynamics that accompany endocytosis, stabilize F-actin, and hyperactivate the Ras/cAMP pathway. Adenylyl cyclase and the cyclase-associated protein Srv2 were observed to colocalize with stabilized actin. The colocalization of stabilized actin and cAMP pathway components was accompanied by concomitant increases in cytoplasmic cAMP levels, oxygen radicals, and apoptosis, leading to the conclusion that stabilized actin structures and hyperactivation of the cAMP signaling pathway were interrelated. Apoptosis and the presence of ROS have also been described for C. albicans (61, 62), and connections between these processes and germ tube formation are indicated by an increase in ROS at the time of germ tube formation (67). The incomplete rescue of C. albicans HWP1 gene expression by dibutyryl-cAMP in the presence of genetic and chemical-genetic compromises to F-actin parallels the lower levels of ROS generated upon addition of exogenous cAMP to a pde2 mutant compared to ROS levels generated by the endocytosis mutants of S. cerevisiae and is suggestive of a similar relationship between the cAMP signaling pathway and F-actin in the two organisms. The possibility that hypha-inducing conditions impact mitochondrial function is suggested by the interference of these conditions with electron transport, a shift to fermentative metabolism, and impairment in oxygen uptake (37). Finally, a recent result from the literature that is also consistent with the correlation between the presence of stabilized actin and activation of the cAMP pathway in C. albicans was the finding that a C. albicans sla2/sla2 mutant did not show alterations in cAMP-regulated gene expression despite the inability to form true hyphae (58). C. albicans sla2/sla2 null mutants would be predicted to have a block in endocytosis-associated actin dynamics and stabilization similar to that of S. cerevisiae sla1/sla1 mutants given the related roles of Sla1p and Sla2p in S. cerevisiae endocytosis (52).
Despite the intriguing similarities mentioned in the previous paragraph, there are unexplained differences between pathways leading to HWP1 gene expression and pathways leading to apoptosis in both S. cerevisiae and C. albicans. Little is known about the exact signaling mechanisms that limit HWP1 gene expression to hypha-inducing growth conditions, the potential colocalization of the C. albicans cAMP pathway components Srv2 and Cyr1 (adenylyl cyclase) with aggregated F-actin, the functional roles of the C. albicans protein kinase A (PKA) isoforms Tpk1 and Tpk2 at the bud-hypha transition in comparison with the established roles of S. cerevisiae PKA isoforms in apoptosis, and the differences in F-actin organization and interactions with cAMP between S. cerevisiae and C. albicans that permit balanced hyphal growth and reversion to yeast growth conditions in the latter case. Future research to investigate these questions is likely to have a high impact on the understanding of the mechanisms of candidiasis that relate to interconversion between C. albicans growth morphologies.
Srv2 of C. albicans could be important for the coupling of F-actin dynamics and cAMP pathway activation that leads to HWP1 gene expression (7). In this work, we found that HWP1 was not expressed in the srv2/srv2 mutant even though stabilized actin structures could be induced by jasplakinolide, suggesting that cAMP activation is necessary for HWP1 gene expression and that the Srv2 protein is essential to this process. In S. cerevisiae, it has been shown that Srv2 localizes to actin patches at the site of endocytosis as well as to stabilized actin structures that are generated in the mutants with defects in endocytosis (21). It is possible that Srv2 serves as a sensor of changes in actin dynamics that
