Over the last 15 years, yeast pseudohyphal growth (PHG) has been the focus of intense research interest as a model of fungal pathogenicity. Specifically, PHG is a stress response wherein yeast cells deprived of nitrogen form filaments of elongated cells. Nitrogen limitation also induces autophagy, a ubiquitous eukaryotic stress response in which proteins are trafficked to the vacuole/lysosome for degradation and recycling. Although autophagy and filamentous growth are both responsive to nitrogen stress, a link between these processes has not been investigated to date. Here, we present several studies describing an interrelationship between autophagy and filamentous growth. By microarray-based expression profiling, we detect extensive upregulation of the pathway governing autophagy during early PHG and find both processes active under conditions of nitrogen stress in a filamentous strain of budding yeast. Inhibition of autophagy results in increased PHG, and autophagy-deficient yeast induce PHG at higher concentrations of available nitrogen. Our results suggest a model in which autophagy mitigates nutrient stress, delaying the onset of PHG; conversely, inhibition of autophagy exacerbates nitrogen stress, resulting in precocious and overactive PHG. This physiological connection highlights the central role of autophagy in regulating the cell's nutritional state and the responsiveness of PHG to that state.

FROM the human pathogen Candida albicans to the corn smut fungus Ustilago maydis, many diverse fungal species possess the ability to switch between a cellular yeast form and a filamentous invasive form in response to appropriate environmental cues (Gimeno et al. 1992; Madhani and Fink 1998). Constituting an essential determinant of fungal pathogenicity in both plants and humans (Lo et al. 1997), this morphogenetic switch has garnered increased attention over the last 15 years, particularly in the budding yeast Saccharomyces cerevisiae (Gimeno et al. 1992). Like its pathogenic counterparts, certain strains of S. cerevisiae also undergo a shift to a filamentous growth form (Kron 1997; Madhani and Fink 1998; Gancedo 2001). Presumably as a means of foraging for nutrients, diploid yeast cells grown under conditions of nitrogen starvation differentiate into branching chains of elongated cells (Gimeno et al. 1992; Liu et al. 1993). The morphogenetic changes associated with filamentous differentiation are extensive; during filamentous growth, yeast cells delay in G2/M, exhibit an elongated morphology, bud in a unipolar fashion, remain physically attached, and invade their growth substrate (Gimeno et al. 1992; Kron et al. 1994). The resulting filaments are called pseudohyphae, and hence this form of growth is referred to as pseudohyphal growth (PHG).

In S. cerevisiae, PHG is regulated by at least two signaling pathways: (1) the nutrient-sensing cyclic AMP-protein kinase A (PKA) pathway, and (2) a mitogen-activated protein kinase (MAPK) pathway. During filamentous growth, the GTP-binding protein Ras2p is activated through a sensor system that is not well characterized at present. Activated Ras2p, in turn, stimulates the synthesis of cAMP, which activates PKA (Robertson and Fink 1998). The yeast PHG MAPK cascade also functions downstream of Ras2p (Mosch et al. 1996). Activated Ras2p acts through the G-protein Cdc42p to stimulate the p21-activated kinase Ste20p (Peter et al. 1996). Ste20p, in turn, initiates a MAPK signaling cascade consisting of Ste11p, Ste7p, and the MAPK itself, Kss1p (Cook et al. 1997). These well-characterized signaling modules act upstream of a diverse and incompletely defined set of genes, including many transcription factors such as Ste12p, Tec1p, Phd1p, Flo8p, and Mss11p (Kobayashi et al. 1996; Liu et al. 1996; Madhani and Fink 1997; Webber et al. 1997; Bardwell et al. 1998; Gagiano et al. 2003; Prinz et al. 2004; Van Dyk et al. 2005; Borneman et al. 2006). The PHG PKA and MAPK pathways have been linked with pathways governing cell polarity, bud site selection, and cell cycle progression (Rua et al. 2001), but the extensive changes associated with PHG likely encompass additional pathways as well.

Like PHG in yeast, autophagy is also a stress response initiated under conditions of nutrient deprivation. Autophagy is an intracellular catabolic pathway conserved among all eukaryotes in which cytosol, organelles, and other structures are sequestered within double membrane vesicles (autophagosomes) for delivery to the vacuole/lysosome, where they are consumed by resident hydrolases (Reggiori and Klionsky 2002, 2005; Levine and Klionsky 2004). Autophagy plays a principal role in the degradation and recycling of long-lived proteins and organelles; as such, it is an important cellular stress response, enabling eukaryotic cells to survive starvation conditions by generating an internal pool of nutrients (Reggiori and Klionsky 2002; Shintani and Klionsky 2004a). Although nitrogen deprivation is the most common stimulus for autophagy in laboratory studies, carbon stress (Takeshige et al. 1992), amino acid stress (Yang et al. 2006), and organelle stress, in the form of endoplasmic reticulum stress and mitochondrial dysfunction (Yorimitsu et al. 2006; Abeliovich 2007), also result in activation of the autophagy pathway; these stresses, however, do not induce filamentous growth. Through extensive studies, this pathway is known to encompass >20 autophagy-related (ATG) genes in the budding yeast (Levine and Klionsky 2004). In particular, Atg1p is a serine/threonine kinase essential for autophagy (Matsuura et al. 1997; Stephan and Herman 2006). Atg1p is required for the induction of autophagy and is thought to function as part of a protein complex with several other components of the autophagy pathway (Reggiori et al. 2004; Klionsky 2005). ATG7 encodes an activating enzyme (E1) that is part of two ubiquitin-like systems essential for vesicle expansion and completion (Mizushima et al. 1998; Ichimura et al. 2000). While autophagy-related functions have been identified for the majority of ATG genes, functional relationships between autophagy and other cell signaling pathways remain to be determined.

To date, autophagy has not been investigated in a filamentous strain of S. cerevisiae, and thus, no connection between autophagy and PHG has been considered. Here, we present several studies indicating a physiological interrelationship between these processes. Through microarray-based expression profiling, assays for autophagic induction, filamentous growth analyses, and cell survival assays, we derive a model of yeast PHG and autophagy in which PHG is responsive to the degree of nitrogen stress, and autophagy plays a critical role in determining the degree of this stress.



All nonfilamentous lab strains are of the S288c genetic background and are derived from those used by the Yeast Deletion Consortium (e.g., BY4743 described in Winzeler et al. 1999). All filamentous lab strains are derived from the Σ1278b genetic background (Gimeno et al. 1992). The filamentous strains Y825 and Y826 were used to generate homozygous diploid deletion strains. The genotype of Y825 is as follows: MATa ura3-52 leu2Δ0. Y826 is a haploid strain of opposite mating type otherwise isogenic to Y825. Modified forms of Y825 and Y826 were constructed containing URA3 (Y825 Ura+) and LEU2 (Y826 Leu+) for the subsequent generation of Y825/6 diploid mutants.

Media and growth conditions:

PHG was induced according to standard protocols using low-nitrogen growth media (Gimeno et al. 1992), except as noted. Briefly, a 50-ml yeast culture was grown at 30° to an OD600 of 0.6 (cell density of ∼4.3 × 106) in YPD medium (1% yeast extract, 2% peptone, 2% glucose). Cells were harvested by centrifugation and washed twice before being transferred to SLAD medium (2% glucose, 50 μm ammonium sulfate, 0.17% yeast nitrogen base without amino acids, and ammonium sulfate, supplemented with essential amino acids for nutritional auxotrophies) for varying times as indicated (Gimeno et al. 1992).

PHG was assessed in autophagy mutants by growth in SLAD medium and by growth in SLAD medium supplemented with 1% ethanol. Strains were incubated on plates at 30° for ∼5–6 days, followed by continued growth at room temperature for an additional 3–4 days as needed.

Gene deletions and ATG1 overexpression:

Gene deletions were performed using the one-step gene replacement strategy of Baudin et al. (Baudin et al. 1993) with the KanMX6 disruption cassette from plasmid pFA6a-KanMX6 (Longtine et al. 1998). To generate homozygous diploid deletion mutants, gene replacement was performed individually in Y825 Ura+ (MATa) and in Y826 Leu+ (MATα); transformants were selected on YPD plates containing 200 μg/ml G418. These strains were subsequently mated and selected on SC–Ura–Leu medium to generate homozygous diploid deletion mutants. In all cases, correct integration was verified by PCR.

ATG1 overexpression was achieved using the pRS416-derived plasmid pCUP1-ATG1 carrying a gene fusion between the copper-inducible CUP1 promoter and ATG1 (Sikorski and Hieter 1989). Expression was induced using media supplemented with 10, 50, or 100 μm copper sulfate.

Microarray experiments and data analysis:

Yeast strains were cultured as described above. RNA was prepared according to standard protocols using the Poly(A) Purist kit (Ambion, Austin, TX). RNA concentration and purity were determined spectrophotometrically and by gel electrophoresis. Microarray hybridization was performed with the Yeast Genome S98 Array using standard protocols (Affymetrix, Santa Clara, CA). All microarray experiments were performed in quadruplicate (four biological replicates) for each strain and indicated time point. Prinz et al. (2004) previously profiled gene expression upon nitrogen deprivation in a filamentous strain of budding yeast; in their study, a derivative of the Σ1278b strain was grown in liquid culture under conditions of nitrogen sufficiency (SLAD medium supplemented with 32 μm ammonium) and was subsequently transferred to solid low ammonium plates (50 μm ammonium sulfate) for growth between 1 and 10 hr. RNA was extracted hourly from cultures. Thus, the study by Prinz and colleagues differs from this study with respect to the time points sampled and growth conditions used.

Here, differentially expressed genes were identified by significance analysis of microarrays (SAM) (Tusher et al. 2001; Rieger and Chu 2004). SAM is a statistical technique in which genes exhibiting significant changes in expression can be identified by assimilating a set of gene-specific t-tests. Briefly, SAM computes a nonparametric score for each gene by dividing the between-group difference of (normalized log) gene expression levels and the within-group difference of gene expression levels. The score is then compared with random permutation scores. The random permutation scores for a gene are computed in the same manner as the original score but based on randomly sampled gene expressions. If the difference between the original score and the random permutation score is larger than a chosen threshold value, the corresponding change in gene expression is claimed to be significant. Each threshold value corresponds to a false discovery rate (FDR), indicating the percentage of genes identified as being significant by chance alone. Thus, increasing the threshold value decreases the number of claimed significant genes but also decreases the FDR, yielding a greater degree of confidence. Here, we have used SAM's multi-class analysis function, with the threshold value chosen so that the corresponding FDR was 0. The multi-class function is used to identify genes undergoing significant changes in expression between multiple time points.

Western blotting and GFP-Atg8p processing assay:

GFP-Atg8p transport and processing was monitored by microscopy and biochemical means as previously described (Kim et al. 2001; Shintani and Klionsky 2004b). The plasmid expressing the GFP-Atg8p fluorescent chimera (pCU-GFP-ATG8) (Kim et al. 2001) was introduced by standard DNA transformation into the nonfilamentous yeast strain BY4743 (Winzeler et al. 1999) and into the filamentous Y825/6 strain. Fluorescent images of GFP-Atg8p were acquired using the DeltaVision Spectris inverted epifluorescence microscope (Applied Precision, Issaquah, WA).

For Western blot analysis, strains carrying pCU-GFP-ATG8 were grown in SC-Ura medium with 50 μm of CuSO4 to an OD600 of 0.8. Two OD600 equivalents of cells were transferred into SLAD medium with 0, 50, or 100 μm ammonium sulfate and were incubated at the same temperature for an additional 3 hr. Cells were successively collected by centrifugation before precipitating proteins with 10% trichloroacetic acid (TCA) followed by two washings with 100% acetone. Finally, proteins were resuspended in 80 μl of SDS-PAGE sample buffer (72 μl of Laemmli sample buffer and 8 μl of 1 m dithiothreitol) by sonication and vortexing in the presence of glass beads. Samples were incubated at 75° for 10 min and 0.5 OD600 equivalents of cells were resolved by SDS-PAGE. After Western blotting, membranes were probed with both anti-GFP (Covance Research Products, Berkeley, CA) and anti-Pgk1p (Invitrogen, Carlsbad, CA).

Cell survival assays:

Nitrogen starvation experiments were performed essentially as described previously (Scott et al. 1996). Briefly, wild-type and deletion strains were grown in 5 ml YPD to 0.6 OD600. Cells were collected and washed twice before being transferred to SLAD medium. After growth for the indicated periods of time, samples were collected and diluted 10,000-fold. Then 100 μl of each diluted culture was spread on a YPD plate. Viable colonies were counted after 2 days' growth at 30°. All platings were performed in triplicate.


Increased transcription of the autophagy pathway during early PHG:

In S. cerevisiae, the transition to filamentous growth is striking, encompassing morphological and cellular changes driven, in part, by an extensive transcriptional regulatory network. By microarray-based expression profiling of the yeast genome, we investigated the scope of genes and cell processes transcriptionally regulated during PHG. As previous studies suggest that the transcriptional program of PHG is initiated quickly within the first few hours upon nitrogen limitation (Prinz et al. 2004), we specifically chose to profile the early onset of PHG, identifying genes differentially expressed after 20 min, 1 hr, and 2 hr (approximately one generation) of nitrogen deprivation in a filamentous strain of budding yeast. Note that the Σ1278b strain serves as the genetic background for our studies; unlike most laboratory strains of S. cerevisiae, Σ1278b undergoes an extensive and easily controlled transition to PHG and, as a result, is the preferred background for studies of yeast filamentous growth.

In total, this microarray analysis reveals an extensive transcriptional program encompassing a wide variety of genes and cell pathways; a full listing of genes differentially expressed in at least one sampled time point is provided as supplemental data (at http://www.genetics.org/supplemental/). In particular, this transcriptional profile reveals an interesting and previously undocumented point: The pathway mediating autophagy is extensively upregulated during early PHG (Figure 1, A and B). In yeast, the process of nonselective bulk autophagy requires 19 genes (Nair and Klionsky 2005); 11 of these genes were transcriptionally induced during PHG (ATG1, ATG3, ATG4, ATG5, ATG6, ATG7, ATG8, ATG9, ATG14, ATG17, and ATG22). Specifically, mRNAs for these genes were identified as being differentially abundant in at least one of the time points examined, with increased abundance evident upon 1- to 2-hr nitrogen stress in filamentous yeast. Microarray results were confirmed by real-time PCR (supplemental Table S1). Formally, this reflects either increased transcription of a given gene or decreased RNA turnover, and we use the general terms “induction” or “upregulation” to indicate this point.

Figure 1.—

Differential expression of autophagy-related genes during early PHG. (A) Red-green color scale indicating the mean change in expression (log scale) for a given gene relative to the standard deviation of repeated measurements for that gene; all experiments were performed with four biological replicates at each time point. In this heat map, the red coloring indicates increased mRNA abundance at the given time point. Statistical significance was determined by SAM (as described in materials and methods). The score provided alongside each gene is a statistical measure describing the degree of the observed change in gene expression; the higher the score, the greater the change in mRNA abundance for a given gene in the time course studied. Note that all autophagy-related genes shown here exhibit statistically significant changes in gene expression, as determined using a threshold cut-off score such that the FDR is 0. (B) Listing of the autophagy-related process and specific function for each protein encoded by the ATG genes identified in this microarray study.

In addition to the genes responsible for bulk autophagy, we also find three autophagy-related genes specific for the cytoplasm to vacuole targeting (Cvt) pathway (ATG19, ATG20, and ATG21) upregulated during our microarray time course analysis of early PHG (Figure 1, A and B). The Cvt pathway is a type of selective autophagy in which oligomers formed by the resident vacuolar protease Lap4p/Ape1p are transported directly from the cytoplasm into the vacuole lumen (Reggiori and Klionsky 2002). Thus, we identify transcriptional upregulation of genes encoding components of both yeast trafficking pathways mediating protein delivery directly from the cytoplasm to the vacuole. In total, 29 autophagy-related genes have been identified in yeast (He and Klionsky 2006; Klionsky and Kumar 2006) and we find 14 of these genes transcriptionally induced during early filamentous growth in S. cerevisiae. It is noteworthy that these 14 autophagy-related genes were not identified as being differentially regulated in previous microarray-based studies of filamentous growth (Prinz et al. 2004), possibly due to the different time frames, growth conditions, and statistical measures employed in the respective analyses.

Induction of autophagic activity during filamentous growth:

Although approximately one-half of all known autophagy-related genes are transcriptionally upregulated during PHG, it is possible that autophagy itself may not be active during filamentous growth due to post-transcriptional regulatory mechanisms. To consider this possibility more explicitly, we used the GFP-Atg8p processing assay developed by Shintani and Klionsky (Shintani and Klionsky 2004b) as an indication of autophagic induction. Atg8p is a ubiquitin-like protein essential for autophagy that is unconventionally linked to the lipid phosphatidylethanolamine (Ichimura et al. 2000). Part of Atg8p remains associated with autophagosomal structures from the stage of initial formation to complete breakdown in the vacuole; therefore, GFP-Atg8p provides a marker to follow the itinerary of double membrane vesicles (Kirisako et al. 1999). As a result, the delivery of Atg8p to the vacuole serves as a useful measure of autophagosome formation and autophagic induction. To visualize this process, we use a GFP-Atg8p chimera. Upon delivery to the vacuole, Atg8p is degraded, while GFP, which is relatively stable in the presence of vacuolar hydrolases, accumulates in the vacuolar lumen. Therefore, the presence of the free GFP moiety in the vacuole indicates Atg8p delivery and autophagic induction (Shintani and Klionsky 2004b).

As indicated in Figure 2A, using the GFP-Atg8p processing assay, we detect induction of autophagy during filamentous growth in S. cerevisiae. By Western blot analysis, we detect free GFP resulting from cleavage of GFP-Atg8p in the vacuole during nitrogen stress (concentrations of 0, 50, and 100 μm ammonium sulfate) in the filamentous yeast strain Σ1278b. For comparison, we repeated this analysis in the standard nonfilamentous S288c-derived genetic background BY4743; autophagy is known to occur in nonfilamentous yeast during nitrogen stress, and we observe comparable levels of autophagic induction in filamentous yeast as compared to nonfilamentous yeast. These results are confirmed by fluorescence microscopy of yeast cells carrying GFP-Atg8p under conditions of nitrogen stress and sufficiency (Figure 2B). Under normal growth conditions, GFP-Atg8p is localized to a perivacuolar punctate spot, indicating the preautophagosomal structure (PAS). The PAS is believed to be the site where double-membrane vesicles are formed (Suzuki et al. 2001; Kim et al. 2002). Under conditions of nitrogen deprivation, however, GFP staining is evident in the vacuole in filamentous and nonfilamentous strains of yeast, indicative of Atg8p transport to the vacuole and autophagic activity. Thus, we conclude that autophagy is induced during nitrogen stress in filamentous yeast.

Figure 2.—

GFP-Atg8p processing assay of autophagic induction during nitrogen stress in filamentous and nonfilamentous yeast strains. (A) Western blot of GFP-Atg8p processing; full-length GFP-Atg8p fusion and cleaved GFP are indicated. Pgk1p serves as a loading control. From these blots, autophagic induction is detected in both filamentous (Σ1278b) and nonfilamentous (BY4743) yeast strains. (B) Analysis of GFP-Atg8p processing by fluorescence microscopy. In both filamentous and nonfilamentous strains, staining of the perivacuolar PAS is evident under conditions of nitrogen sufficiency, and vacuolar staining is evident under conditions of nitrogen deprivation. By this assay, vacuolar GFP staining is not observed under conditions of nitrogen stress in yeast cells deficient in autophagy (data not shown). Bar, 5 μm.

Phenotypic analysis of PHG in autophagy-impaired mutants:

Since PHG and autophagy are active stress responses in a filamentous strain of yeast, with nitrogen deprivation acting as a common stimulus, we sought to further investigate a relationship between these processes. For this purpose, we generated a homozygous diploid strain of the Σ1278b genetic background deleted for ATG1. Atg1p is a serine/threonine kinase essential for autophagy (Matsuura et al. 1997; Stephan and Herman 2006). Atg1p is required for the induction of autophagy, and it is thought to function as part of a protein complex with several other components of the autophagy pathway (Reggiori et al. 2004; Klionsky 2005).

Using the standard assay of Gimeno et al. (1992), we examined homozygous diploid atg1Δ mutants for filamentous growth under conditions inducing PHG. Interestingly, we found increased PHG in atg1Δ relative to the filamentous wild-type strain (Figure 3). Conversely, inhibition of filamentous growth does not affect autophagy appreciably (data not shown); this is consistent with the volume of studies characterizing active autophagy in strains of budding yeast deficient in filamentous growth.

Figure 3.—

Colony and cell morphology of autophagy mutants in a filamentous strain of budding yeast. (A) Low-magnification images of yeast colonies from wild-type, autophagy-deficient, and ATG1 overexpression strains in the filamentous Σ1278b genetic background. All strains were grown for 6 days at 30° in SLAD medium (materials and methods) supplemented with 1% ethanol; this analysis was repeated using nitrogen deprivation (SLAD medium) alone to induce PHG, and observed results were identical. Bar, 1 mm. (B) Cell morphology of wild-type, autophagy-deficient, and ATG1 overexpression strains of the filamentous Σ1278b genetic background as imaged by differential interference contrast (DIC) microscopy. Strains were cultured as above; colonies were scraped into a solution for DIC microscopy. Bar, 5 μm. (C) Pie charts indicating the observed cell length:width ratios of each strain. The cell sample number is indicated in the center of each chart.

The increased growth of the homozygous diploid atg1Δ strain is evident in its colony morphology (Figure 3A) and at the cellular level as well (Figure 3B). Yeast cells undergoing PHG are characteristically elongated and can be distinguished from cells undergoing normal vegetative growth by this fact; however, a colony is a heterogeneous cell population, and even during PHG, not all cells within a colony will be elongated. The relative fraction of elongated cells, though, does provide a confirming measure of the extent of PHG in a given strain under given growth conditions. To assess this more quantitatively, we measured the length and width of cells from atg1Δ and wild-type colonies under PHG-inducing growth conditions; this analysis is indicated in Figure 3, B and C. PHG cells exhibit a length:width ratio of ∼1.5–2.0 or greater, and we find a larger fraction of these elongated cells in atg1Δ relative to wild type. Specifically, 4 times as many atg1Δ cells as compared to wild-type cells exhibit a length:width ratio of ≥2.0 under identical PHG-inducing conditions. Other than the increased fraction of elongated cells, we do not detect significant differences in cell morphology between the atg1Δ mutants and wild-type Σ1278b strains; both strains exhibit large vacuoles indicative of growth under conditions of nutrient stress (Figure 3B).

The phenotype observed upon deletion of ATG1 may be specific to this gene or may result from general inhibition of the autophagy pathway. To distinguish between these possibilities, we generated a homozygous diploid strain of the Σ1278b background deleted for ATG7. ATG7 encodes an activating enzyme (E1) that is part of two ubiquitin-like systems essential for autophagy (Mizushima et al. 1998; Ichimura et al. 2000). Atg7p is required for vesicle expansion and completion; it is not thought to function in complex with Atg1p. As indicated in Figure 3, homozygous diploid atg7Δ mutants also exhibited increased filamentous growth under PHG-inducing conditions. Colonies of atg7Δ mutants display increased surface-spread filamentation, and approximately five times as many cells from atg7Δ colonies are elongated (length:width ratio >2.0) relative to wild-type Σ1278b cells. Thus, we observed exaggerated PHG in both atg1Δ and atg7Δ mutants.

Phenotypic analysis of PHG upon ATG1 overexpression:

In complement to phenotypic studies of atg deletion mutants, we also overexpressed ATG1 and assessed PHG. For this study, we expressed ATG1 from the copper-inducible CUP1 promoter carried on a low-copy yeast shuttle vector derived from pRS416 (Scott et al. 2007). Even in the absence of copper sulfate, expression of Pcup1-ATG1 yields 2- to 3-fold overexpression of ATG1 (as confirmed by Western blot analysis). It is important to note that overexpression of ATG1 is insufficient to activate autophagy under noninducing conditions in yeast; however, it is difficult to quantify autophagic activity, and, thus, it is difficult to assess whether the process occurs more aggressively upon overexpression of ATG1 under conditions of nitrogen stress. Qualitatively, by the GFP-Atg8p assay described previously, autophagy is strongly activated by ATG1 overexpression under conditions of nitrogen stress. Also consistent with increased autophagic induction, a yeast strain of the Σ1278b background overexpressing ATG1 exhibits smaller colony size on low-nitrogen medium than a corresponding wild-type strain.

To assess PHG upon ATG1 overexpression, we assayed the strain described above for surface-spread filamentation at the colony level and for cell elongation/clustering at the single-cell level. As indicated in Figure 3, ATG1 overexpression is sufficient to markedly decrease PHG relative to a wild-type strain grown under identical conditions of nitrogen stress. This phenotype is consistent with results from the converse experiment in which hyperactive filamentous growth was observed upon deletion of ATG1.

Graded PHG during nitrogen stress in filamentous yeast:

Considering the findings presented above, exaggerated filamentous growth during nitrogen deprivation in autophagy-impaired mutants may reflect the worsened state of nitrogen stress in these strains. Autophagy is a recycling process, acting, in part, to mitigate the effects of nitrogen starvation. In the absence of autophagy, nitrogen stress may be significantly increased relative to a wild-type strain grown under identical conditions of nitrogen deprivation; the exaggerated PHG in autophagy-deficient mutants may reflect this condition. If so, PHG must be a graded response, with increased filamentous growth correlated with decreasing available nitrogen. As shown in the lower panel of Figure 4, we find that wild-type Σ1278b exhibits a graded increase in PHG in response to decreasing levels of exogenously supplied ammonium sulfate. Note that the growth medium for this analysis contains some amino acids to complement nutritional auxotrophies in the strain; thus, even the absence of ammonium sulfate generates a state of nitrogen stress rather than nitrogen starvation. We observe a similar graded response in atg7Δ mutants; however, PHG is induced at a higher concentration of exogenously supplied ammonium sulfate (Figure 4, top). In the presence of low-nitrogen growth medium supplemented with 100 μm ammonium sulfate, atg7Δ undergoes PHG, whereas a wild-type strain of the same genetic background does not. We observe identical results in mutants deleted for ATG1 (data not shown). We therefore suggest that exaggerated PHG in autophagy-deficient mutants may be a cellular response to an exacerbated condition of nitrogen stress in these strains.

Figure 4.—

Graded PHG response during nitrogen stress in wild-type and autophagy-deficient filamentous yeast. Low-magnification images of diploid wild-type and homozygous diploid atg7Δ colonies of the Σ1278b genetic background under varying conditions of nitrogen stress. Strains were cultured 6 days at 30° in SLAD medium prepared with the indicated concentrations of ammonium sulfate. Both strains were grown in medium supplemented with uracil and leucine to complement nutritional auxotrophies; thus, the absence of ammonium sulfate does not constitute complete nitrogen starvation. As indicated in the top and bottom panels, atg7Δ initiates PHG at a higher concentration of ammonium sulfate than does the wild-type strain. Bar, 1 mm.

Filamentous growth and autophagy facilitate cell survival during nitrogen stress:

The findings above suggest that both filamentous growth and autophagy act to relieve nitrogen stress, presumably contributing to yeast cell survival. To consider this further, we have assessed the ability of wild-type filamentous (Σ1278b) and nonfilamentous yeast to survive during nitrogen starvation (Figure 5). In contrast to the nonfilamentous yeast strain BY4743, wild-type Σ1278b cells survive nitrogen deprivation fairly well for >9 days, exhibiting only a 25% reduction in colony number. To assess the contributions of PHG and autophagy in a filamentous strain of S. cerevisiae, we generated and assayed a homozygous diploid mutant impaired in autophagy (atg1Δ) and a homozygous mutant impaired in filamentous growth (muc1Δ) for cell viability under conditions of nitrogen starvation in the Σ1278b background. Muc1p is a GPI-anchored cell surface glycoprotein required for diploid PHG under conditions of nitrogen stress (Lambrechts et al. 1996; Rupp et al. 1999; Guo et al. 2000). As shown in Figure 5, the autophagy-defective atg1Δ strain dies rapidly upon nitrogen removal, becoming inviable after 4–5 days. The homozygous diploid muc1Δ mutant survives the course of the assay, exhibiting an efficiency of survival comparable to that of a wild-type nonfilamentous strain, although, after 5 days of starvation, the muc1Δ mutant cells begin to die more rapidly. Thus, both filamentous growth and autophagy contribute to cell survival during nitrogen starvation, but autophagy plays a more critical role.

Figure 5.—

Cell survival of Σ1278b mutants impaired in autophagy and PHG. The diploid wild type Σ1278b strain and diploid nonfilamentous S288c-derivative BY4743 were assayed for cell viability under conditions of nitrogen stress. Briefly, cultures were grown to log-phase in YPD before being transferred to low-nitrogen growth medium. At the indicated days, aliquots of cells were diluted and plated on YPD. Percent survival was determined by counting the number of surviving colonies from triplicate platings and by dividing the number of colonies at each time point by the average number of colonies obtained at day 0. Homozygous diploid atg1Δ and muc1Δ mutants in the filamentous Σ1278b background were also assayed for cell viability as described above.


By microarray analysis of genes differentially expressed during early PHG in the filamentous Σ1278b strain of S. cerevisiae, we detect extensive upregulation of the autophagy pathway. Although both PHG and autophagy are active under conditions of nitrogen deprivation in filamentous strains of yeast, we find that inhibition of autophagy results in precocious and exaggerated filamentous growth. This phenotypic effect is not specific to a single autophagy-related gene; instead, it seems to reflect a requirement for wild-type function of the autophagy pathway as a whole. Collectively, these results suggest a model (Figure 6) in which both autophagy and filamentous growth mutually mitigate the effects of nutrient stress, contributing positively to the available pool of nitrogen in the cell. In particular, the autophagy pathway strongly affects the degree of nitrogen stress in the yeast cell, and the extent of filamentous growth is responsive to the severity of this nitrogen stress. This model explains the exaggerated degree of PHG evident in autophagy-impaired mutants: interruption of autophagy results in a heightened state of nitrogen stress, manifesting itself in the premature initiation of PHG and, thus, hyperactive filamentation.

Figure 6.—

Model of the interrelationship between PHG and autophagy in yeast. In the filamentous strain Σ1278b, nitrogen stress induces both filamentous growth and autophagy, with both processes contributing to cell survival and presumably to the relief of nitrogen stress, although autophagy plays a bigger role (indicated by the larger arrow). In the absence of autophagy, nitrogen stress is exacerbated, and the PHG response is increased. From our studies, autophagy regulates PHG as indicated; however, the molecular mechanism of this regulation remains to be determined.

PHG and autophagy are interconnected stress responses:

Autophagy is involved in many important physiological processes but has been studied most intensely as a cellular adaptation to starvation conditions (Reggiori and Klionsky 2005). As a stress response, therefore, its link with PHG is logical: nitrogen deprivation is a common stimulus inducing PHG and autophagy. Our results indicating transcriptional induction of the autophagy pathway are consistent with specific expression studies of the autophagy genes ATG8 and ATG14 (Kirisako et al. 1999; Chan et al. 2001), reported to be upregulated in nonfilamentous yeast under conditions of nitrogen stress. It is unclear as to why transcriptional induction of autophagy is widespread over the pathway, encompassing the majority of autophagy genes. Possibly, autophagy proteins are required to sustain an intense autophagic activity. In agreement with this hypothesis, it has been shown that the induced expression of ATG8 and its subsequent translation are essential to generate normal-sized autophagosomes (Abeliovich et al. 2000). This extensive, but not comprehensive, complement of PHG-regulated autophagy genes may suggest the presence of many regulatory control points rather than a single focus point. Computational analysis of the ATG promoter sequences does not reveal any enrichment for known transcription factor binding sites, and, in particular, chromatin-immunoprecipitation/microarray studies of PHG transcription factors do not identify autophagy-related genes (Borneman et al. 2006).

Transcriptional induction of both autophagy and PHG correlates with the activity of each process; both processes are active under conditions of nitrogen stress in a strain of yeast capable of filamentous growth, and autophagy is active in nonfilamentous yeast under similar conditions. Therefore, we do not find that autophagy and PHG act as mutually exclusive pathways but rather that both pathways contribute to the cellular response to nutrient stress. The role of autophagy in relieving nutrient stress is well documented (Wang and Klionsky 2003; Klionsky and Kumar 2006; Yang et al. 2006). PHG plays a less substantial, but nonetheless tangible, role in this process as well: A cohort of genes mediating nitrogen utilization is upregulated during PHG (Prinz et al. 2004), and a strain of budding yeast capable of undergoing filamentous growth survives nitrogen stress better than a nonfilamentous wild-type strain (Figure 5). Of course, as the filamentous and nonfilamentous strains are nonisogenic, this observation must be interpreted cautiously.

The timing and onset of filamentous growth and autophagy:

The activity of both autophagy and PHG during nitrogen stress raises an interesting question as to the timing and onset of each process. This question is difficult to address directly, since we lack a well-defined indicator for the onset of PHG other than cell morphological changes that require a period of 2 hr (one cell cycle) in order to become evident. With that qualification in mind, we speculate that autophagy might be initiated before PHG. The transition to filamentous growth is an extensive morphogenetic process that may be initiated when nutrient stress is sensed as being severe. In essence, autophagy may act to delay the onset of PHG by relieving initial nitrogen stress; note again that PHG occurs at higher concentrations of available nitrogen source in autophagy-deficient yeast (Figure 4). Our ATG1 overexpression results (Figure 3) suggest that the autophagy pathway represses filamentous growth. Again, during nitrogen stress, both autophagy and filamentous growth are active; however, the autophagy pathway may limit filamentous growth, preventing it from being maximally active and thereby contributing to the fine balance between these cellular processes. The fact that 2- to 3-fold overexpression of ATG1 is sufficient to repress filamentous growth suggests that the balance between autophagy and filamentous growth is indeed fine.

A possible direct connection between autophagy and PHG:

Although our model does not necessitate a direct link between autophagy and filamentous growth, it is likely that such a connection exists. Since the autophagy pathway is active under conditions of nitrogen stress irrespective of filamentous growth (Figure 2), but filamentous growth is affected by autophagy (Figure 3), it seems most likely that a component of the autophagy pathway might regulate PHG (this is indicated in Figure 6 by the horizontal arrow from autophagy to PHG). Several PHG or autophagy-related transcription factors and/or kinases may mediate cross talk between the processes; however, at present, chromatin immunoprecipitation/microarray studies of known PHG transcription factors do not indicate extensive regulation of autophagy-related genes (Borneman et al. 2006), and, by the same token, we have yet to uncover an autophagy protein directly regulating a known component of the PHG pathways. In this regard, it is interesting to consider the decreased level of PHG evident upon overexpression of ATG1. By our current understanding of the pathway in yeast and in contrast to the orthologous pathway in Drosophila melanogaster (Scott et al. 2007), overexpression of ATG1 is not sufficient to induce autophagy under repressive growth conditions. While the colony morphology of ATG1 overexpression mutants and results from the GFP-Atg8p processing assay suggest that the rate of autophagy may be slightly increased upon overexpression of ATG1 under conditions of nitrogen stress, the effect is fairly marginal and is unlikely to affect the cellular nitrogen pool sufficiently to account for the marked filamentous growth phenotype. The results presented here indicate that Atg1p nevertheless plays a repressive role in limiting filamentous growth and that Atg1p may directly or indirectly affect activity of the filamentous growth pathways distinct from signals put forth by nitrogen-sensing transducers. Considering this possibility further, however, will be challenging, as many potential Atg1p targets exist in known PHG pathways (Stephan and Herman 2006).

In total, this study describes an interrelationship between autophagy and PHG and therein raises two important points regarding these processes. First, autophagy is a critical mechanism by which the cell can regulate or buffer its nutritional state from environmental stresses of nutrient deprivation. Second, PHG is finely tuned to the nutritional state of the cell, with the extent of PHG reflective of the degree of nitrogen stress to which the cell is subjected.


We thank Daniel Klionsky for providing plasmids pCU-GFP-ATG8 and pCUP1-ATG1; we also thank Damian Krysan, Robert Fuller, Anthony Borneman, and Michael Snyder for providing filamentous yeast strains. This work was supported by grant RSG-06-179-01-MBC from the American Cancer Society, grant DBI 0543017 from the National Science Foundation, and Basil O'Connor Award 5-FY05-1224 from the March of Dimes. F.R. was supported by NIH grant GM53396.


  • Communicating editor: D. F. Voytas

  • Received May 24, 2007.
  • Accepted June 14, 2007.


View Abstract