Phenomic and transcriptomic analyses reveal that autophagy plays a major role in desiccation tolerance in Saccharomyces cerevisiae

Abstract

Saccharomyces cerevisiae can survive extreme desiccation, but the molecular mechanisms are poorly understood. To define genes involved in desiccation tolerance, two complementary genome-wide approaches, phenomics and transcriptomics, have been used, together with a targeted analysis of gene deletion mutants tested individually for their ability to survive drying. Genome-wide phenotypic analyses carried out on a pooled library of single-gene deletion mutants subjected to three cycles of desiccation and re-growth to post-diauxic phase identified about 650 genes that contributed to strain survival in the drying process. Air-drying desiccation-tolerant post-diauxic phase cells significantly altered transcription in 12% of the yeast genome, activating expression of over 450 genes and down-regulating 330. Autophagy processes were significantly over-represented in both the phenomics study and the genes up-regulated on drying, indicating the importance of the clearance of protein aggregates/damaged organelles and the recycling of nutrients for the survival of desiccation in yeast. Functional carbon source sensing networks governed by the PKA, Tor and Snf1 protein kinase complexes were important for the survival of desiccation, as indicated by phenomics, transcriptomics, and individual analyses of mutant strains. Changes in nitrogen metabolism were evident during the drying process and parts of the environmental stress response were activated, repressing ribosome production and inducing genes for coping with oxidative and osmotic stress.

---

Introduction

Organisms capable of surviving desiccation and the dry state are found across a range of taxa, including bacteria, yeasts, plants and invertebrates, but the mechanisms of desiccation tolerance are far from fully understood.^1,2 One barrier to a more complete understanding of the phenomenon is that many of the organisms studied are poorly characterised, with limited development of genetic tools or genome sequences. This is not the case for the yeast Saccharomyces cerevisiae, however, which is the only well-characterised model organism capable of extreme desiccation tolerance. Although sensitive to desiccation during its exponential phase of growth, yeast becomes tolerant after diauxie and in stationary phase, when typically 70% of cells will survive air drying to a residual water content of less than 10% by mass.^3,4 Therefore, the budding yeast, which is arguably the best-characterized eukaryote,⁵ represents the ideal organism for determination of the set of genes implicated in desiccation tolerance. Furthermore, yeast is especially amenable to genome-wide studies of transcription (transcriptomics) or phenotype (phenomics), and the responses to and tolerance of various stresses, chemicals, and toxins have been characterised.⁶

General stress resistance is a well-known stationary phase phenotype,⁷ while the importance of diauxie may be more specific.⁴ In neither case is the mechanism of desiccation tolerance fully understood. Comparison of the genomic expression programs triggered by diverse environmental stresses has revealed many condition-specific features in the responses, while also indicating that the majority of the changes form part of a general environmental stress response (ESR) common to all treatments, although governed by condition-specific regulatory mechanisms.^8,9 Many genes involved in the ESR have been correlated with growth rate, indicating a relationship between the expression of stress-response genes and adverse growth due to environmental stresses.^10,11 The rationale for surveys of this type, many of which are accessible through the Saccharomyces Genome Database (www.yeastgenome.org), is that at least a proportion of genes induced in response to a particular stress are stress-combative, i.e. that they alleviate stress damage or are involved in physiological adjustments required for adaptation to the changed environment. Intriguingly, however, an increasing number of studies are showing that there is only a limited correlation between stress-induced genes and those genes required for management of the stress condition. For example, deletion of genes up-regulated in minimal medium was no more likely to affect fitness in this medium than deletion of genes whose transcript levels were unchanged.¹² Similarly, very few genes required for DNA repair were induced by DNA-damaging agents.¹³ Other researchers report similar conclusions for yeast subjected to osmotic stress,¹⁴ oxidative stress,¹⁵ and anaerobic growth conditions.¹⁶

As a consequence, phenomics, the second approach for the identification of key stress-combative genes, is gaining prominence and the availability of tagged deletion mutants of all non-essential genes of S. cerevisiae has been crucial in this regard.^14,17–20 For example, in response to oxidative stress, 456 genes, a majority of them involved in transcription, protein trafficking, and vacuolar function, were found to increase sensitivity upon deletion.¹⁵ A smaller-scale study involving 600 mutants identified about 14% of them as sensitive to oxidizing agents in the functional categories of stress response, heavy metal homeostasis and cell wall proteins.²⁰ A genome-wide analysis of hyperosmotic stress yielded 488 genes whose deletion increased sensitivity, including those responsible for glycerol production, ion homeostasis, cytoskeleton organization, signaling pathways and vacuolar protein transport.¹⁴

For desiccation tolerance, a previous study attempted to identify global transcriptional changes associated with extreme water loss.²¹ However, there are concerns over the experimental strategy adopted, which involved drying exponential phase cells over 72 h, resulting in the death of more than 99.9% of the cell population, i.e. there was a marked lack of tolerance to desiccation. Furthermore, the cells were dried from 1 ml medium with reduced glucose concentration, which is likely to have subjected the cells initially to starvation and osmotic stresses over a long period of time before finally resulting in desiccation. The extremely low survival (0.1%) after rehydration suggests that the cellular response to these stresses was unsuccessful, making interpretation of the results unclear. Earlier work on yeast desiccation tolerance emphasised a role for the non-reducing disaccharide of glucose, trehalose,²² but mutants unable to produce trehalose were also found to be capable of surviving desiccation.⁴ Therefore, there is a pressing need for determination of the genes involved in the response to, and survival of, desiccation stress in yeast.

In this paper, we report the findings of a whole-genome deletion analysis of genes implicated in desiccation tolerance in the budding yeast, together with a genome-wide transcriptional study. Individual desiccation-survival analysis of selected gene deletion mutants was also performed for verification. Combining the results leads to the definition of genes making a significant contribution to survival of desiccation and suggests important pathways involved in the process.

Results

Combining phenomics with transcriptomics to identify genes that contribute to desiccation tolerance: an overview of the datasets

To identify genes with a role in enabling yeast to survive desiccation, a phenomics screen was undertaken by exposing the library of homozygous diploid single-gene deletion mutants²³ to air-drying. The whole library of approximately 4800 mutants was pooled and subjected to three cycles of growth to late post-diauxic phase, followed by desiccation and re-growth. A pilot-scale experiment with 12 deletants of varying desiccation tolerance had shown that those strains with poor tolerance were under-represented in the pooled culture after such treatment (data not shown). Cells from the initial culture prior to drying, and after the third cycle of desiccation treatment, were used as control and test samples, respectively. Sampling was performed from post-diauxic growth phase cultures because of their inherent desiccation tolerance: wild-type haploid and diploid strains in the BY background showed 70% ± 6% survival on drying at this time, compared to 3% ± 2% during exponential phase. Genomic DNA was extracted from the samples and their “molecular barcodes” amplified by PCR before hybridization to microarrays. Genes were regarded as implicated in desiccation tolerance if their deletion led to a 2-fold or greater decrease in representation after the treatment (ESI†, S1). Decreased desiccation tolerance was found in 653 mutants, constituting about 14% of non-essential genes in yeast and about 10% of its genome. To validate the dataset, a panel of 29 of these mutants was subjected to further tests, culturing haploid deletants individually and determining their susceptibility to desiccation stress (ESI†, Table S1). Although the mutants displayed a range of survival values, 22 (75%) were clearly adversely affected in their ability to survive drying, and 16 (55%) were judged to be severely affected or intolerant. Although predictive of a strain's ability to cope with desiccation, the magnitude of the fold-change decrease in representation measured in the phenomics experiment was not generally predictive of the extent of the survival when cultured and dried in isolation. The majority of the 653 gene products identified by the phenomics approach can nonetheless be considered genuinely “beneficial” to desiccation tolerance. To complement the phenomics data, a transcriptomics study was undertaken to obtain a global picture of changes in gene transcription of post-diauxic cultures in response to desiccation. Cells from duplicate cultures of the wild-type yeast strain BY4741 in the post-diauxic growth phase were air dried for 0, 30, 60 and 120 min and gene expression determined using Affymetrix S98 microarrays. Analysis of data acquired using technical replicates revealed probe sets representing a total of 814 ORFs that were significantly differentially expressed at the 5% probability level relative to the 0 min sample (ESI†, S2). Approximately 12% of the yeast genome therefore responded to the desiccation treatment, transcription of 484 ORFs (about 7% of the genome) being up-regulated and 330 (5%) down-regulated (Fig. 1). The transcriptional response was relatively slow, with the majority of significant changes observed after 60 and 120 min, and very few after 30 min. Of the genes identified as being beneficial to survival of desiccation in the phenomics experiment, 83 were also found to be significantly differentially expressed in the 120 min following the onset of drying, 47 up-regulated and 36 down-regulated (marked in ESI†, S2).

Fig. 1 Desiccation of yeast cells in the post-diauxic phase of growth induces changes in transcription of 12% of the genome. Strain BY4741 was grown in YEPD and cells air dried as detailed in the Experimental. RNA was isolated from samples dried for 0, 30, 60 and 120 min and gene expression determined using Affymetrix S98 microarrays. The changes in transcription shown are relative to the 0 min sample (significant to the 5% probability level), and are presented to summarise whether individual genes were significantly differently expressed at one, two or all time points. Detailed lists are presented in ESI†, S2.

Autophagy is important for the survival of desiccation

To obtain an overview of the core biological processes represented by the 653 genes identified in the phenomics analysis, we applied the Bayesian modeling approach MGSA (model-based gene set analysis)²⁴ to search for over-representation of associated GO terms (Table 1). Of the top four most significant terms identified, two were associated with mitochondrial functions (GO:0032543 and :0033108) and two with autophagy (GO:0041362 and :0016237). Hypergeometric GO analysis similarly identified these functions as being the predominant biological processes significantly over-represented in the data (ESI†, S3). A corresponding analysis of the genes up-regulated in response to desiccation identified the GO term for vacuolar protein catabolism (GO:0007039) as the most significant, and the process of membrane invagination for vesicle formation (GO:0010324) was also prominent (Table 2 and ESI†, S4). Taken together, these results indicate an important role for autophagy in desiccation tolerance. The removal of damaged proteins or organelles, and the consequent release and recycling of nutrients are both likely to be beneficial under desiccation stress. This is further supported by the fifth most significant GO term in Table 1, chaperone-mediated protein complex assembly, which includes the genes encoding Irc25p and Poc4p involved in the assembly of a mature, active proteasome complex. Transcripts of several proteasome subunit or assembly genes (ECM29, BLM10, RPN10, HSM3) were also significantly up-regulated 1.5- to 3-fold following the start of desiccation (see ESI†, S2).

Table 1 GO terms significantly over-represented in the genes identified by phenomics as being beneficial for surviving desiccation stress^a

ID	Name	Marginal^b	Population total	Population term	Study total	Study term
a The analysis was performed using the Ontologizer tool^24 interrogating only the Biological Process subontology using the MGSA testing option set on 10^6 MCMC steps. b The posterior marginal probability: higher values in the 0–1 range indicate stronger support. All GO terms with values reproducibly greater than 0.5 are shown. c BRO1, DID4, DOA4, SNF7, SNF8, SRN2, STP22, UBP2, VPS24, VPS25, VPS28, VPS36. d ATG1, ATG11, ATG12, ATG13, ATG15, ATG16, ATG17, ATG18, ATG2, ATG21, ATG3, ATG5, ATG7, ATG9, CIS1, GSG1, NVJ1, PEP4, VAM3, VAM6, VAM7, VPS30, VPS41, VTC1, VTC4. e BCS1, IRC25, PIM1, POC4.
GO:0032543	Mitochondrial translation	1.00	5558	102	596	37
GO:0043162	Ubiquitin-dependent protein catabolic process via the multivesicular body sorting pathway	0.97	5558	20	596	12^c
GO:0016237	Microautophagy	0.95	5558	43	596	25^d
GO:0033108	Mitochondrial respiratory chain complex assembly	0.87	5558	21	596	16
GO:0051131	Chaperone-mediated protein complex assembly	0.71	5558	5	596	4^e
GO:0009415	Response to water	0.53	5558	34	596	11
GO:0051668	Localization within membrane	0.52	5558	12	596	7
GO:0015711	Organic anion transport	0.51	5558	10	596	5

---

Table 2 GO terms significantly over-represented in the genes identified by transcriptomics as being significantly up-regulated (A) or down-regulated (B) immediately following desiccation stress^a

ID	Name	Marginal^b	Population total	Population term	Study total	Study term
a The analysis was performed using the Ontologizer tool^24 interrogating only the Biological Process subontology using the MGSA testing option set on 10^6 MCMC steps. b The posterior marginal probability: higher values in the 0–1 range indicate stronger support. All GO terms with values reproducibly greater than 0.5 are shown. c ACS1, ALD3, AMS1, APE3, ARA1, ATG19, BCY1, BDH2, CMK1, CTT1, DCS1, DCS2, DDR48, DUG1, ECM21, EDC2, GDB1, GGA1, GID7, GLC3, GLK1, GRE3, HBT1, KIN1, LAP4, MDS3, MYO3, PGM2, PMC1, PRB1, RIM11, RIM15, SOL4, TFS1, TPS1, TPS2, TPS3, UBP15, UGP1, VID28, VPS13, XKS1. d AKL1, ALY2, ATG14, ATG2, ATG7, CHC1, CSR2, DNF2, ECM21, EDE1, LSB5, LSP1, MON2, MYO3, NEO1, OSH2, ROD1, ROM2, RVS167, SLA1, VAM6, VTC2, VTC3, YPK1.
(A) Up-regulated genes
GO:0007039	Vacuolar protein catabolic process	1.00	5558	117	468	42^c
GO:0000122	Negative regulation of transcription from RNA polymerase II promoter	0.99	5558	75	468	16
GO:0010324	Membrane invagination	0.93	5558	137	468	24^d
GO:0007568	Aging	0.85	5558	74	468	16
GO:0032787	Monocarboxylic acid metabolic process	0.84	5558	151	468	35
(B) Down-regulated genes
GO:0042254	Ribosome biogenesis	1.00	5558	415	302	81
GO:0009451	RNA modification	0.90	5558	171	302	31

---

The prominence in the phenomics data of gene products associated with mitochondria is suggestive of a role for these organelles in the response to desiccation stress. However, mitochondria are essential for aerobic respiration during post-diauxic growth, and the experimental design could also therefore contribute to these results. The latter is supported by the observation that mutants lacking HAP2, HAP4 and HAP5, which are responsible for the up-regulation of several genes important for respiration,^25–27 also exhibited decreased survival under the conditions used (see ESI†, S1).

Elements of the Snf1 network required for transcription of glucose-repressed genes are beneficial for surviving desiccation

The phenomics analysis identified 31 transcription factors among the products of genes potentially beneficial for surviving desiccation, together with 17 protein kinases (see ESI†, S1). Significantly, genes encoding two transcription factors activated by the Snf1 protein kinase, Cat8p and Sip3p, together with Sak1p (one of the protein kinases responsible for activating Snf1p), figured among the beneficial genes (Fig. 2a). Snf1p is required for growth in the absence of glucose, regulating transcriptional changes associated with glucose de-repression (reviewed in ref. 28). The identification of genes from this regulatory pathway whose deletion significantly reduced survival on drying also suggests a role in the desiccation stress response. This is supported by the results of the transcriptome analysis where a 1.6-fold activation of SNF1 transcription was observed 60 and 120 min into the drying process. Analysis of the promoter sequences of the up-regulated genes to search for statistically over-represented transcription factor binding sites using the Ceres software tool²⁹ also identified an over-representation of sites for Sip4p (p-value < 10⁻⁶), a Snf1 kinase-dependent transcriptional activator (Table 3A).

Fig. 2 The Snf1p regulatory network is important for the survival of desiccation. (a) Members of the network identified in a phenomics screen for mutants with reduced capability to survive desiccation stress. (b) Changes in desiccation tolerance following mutation in selected members of the network. Strains derived from BY4741 were grown individually to the post-diauxic phase of growth in YEPD media before drying. The proportion of cells surviving desiccation was measured as detailed in the Experimental, and the mean of three replicate experiments is shown. Standard deviations are indicated by the error bars. Snf1p and Snf4p are the catalytic (α-) and regulatory (γ-) subunits of the protein kinase complex, respectively, while Sip1p, Sip2p and Gal83p are alternate β-subunits. Sip3p, Sip4p, Cat8p and Mig1p are transcriptional regulatory targets of the Snf1 kinase complex. Sak1p, Elm1p and Tos3p are kinases capable of activating the Snf1 complex, while Reg1p forms part of a phosphatase complex involved in its deactivation (reviewed in ref. 28).

Table 3 Statistically over-represented transcription factor binding sites identified in the promoters of genes significantly differentially expressed in response to desiccation^a

Binding site	p-value	Number of promoters	Number of sites
a Lists of genes identified as significantly up-regulated (A) or down-regulated (B) following desiccation (see ESI^1, S2) were analysed using Ceres.^29
(A) Up-regulated genes
UME6	<10^−10	49	6
MSN4	<10^−3	18	26
SUT1	<10^−4	14	18
SIP4	<10^−6	8	20
CIN5	0.0011	31	43
STB4	0.0011	5	6
SKN7	0.0018	26	44
INO4	0.0026	13	18
MSN2	0.0038	16	19
SKO1	0.0094	7	7
(B) Down-regulated genes
RAP1	<10^−12	42	74
FHL1	<10^−10	41	69
SFP1	<10^−4	17	31
LEU3	0.0049	6	6

---

To validate these results, an analysis of desiccation survival in haploid strains carrying gene deletions in selected members of the Snf1p regulatory network was undertaken (Fig. 2b and ESI†, Table S2). Deletion of SNF1 or SNF4 resulted in strains that did not survive the drying process, while sip1Δ, sip4Δ and sak1Δ mutants showed a marked decrease in survival. Deletion of MIG1 or CAT8, encoding transcriptional regulatory targets of Snf1p phosphorylation, similarly increased sensitivity to desiccation relative to the parent strain BY4741, while disruption of SIP2, GAL83, ELM1 or TOS3 had no appreciable effect. To assess the effect of increasing Snf1 kinase activity, a strain carrying a deletion in REG1 was analysed. Reg1p promotes inactivation of Snf1p by the Glc7p phosphatase.³⁰ Interestingly, the reg1Δ mutant strain showed an improved survival of cells from the post-diauxic cultures during the desiccation process, 95% ± 7% compared to 70% ± 6% for the parent strain. This beneficial effect was also observed when drying cells from exponential phase culture, producing a 4-fold improvement in survival from 3% ± 2% in BY4741 to 12.7% ± 0.4%.

The Ras/PKA and Tor nutritional signaling pathways participate in desiccation stress survival

Genes involved in the regulation of gene expression in response to nutritional signals sensed via the Ras/PKA and Tor protein kinases were also found to be beneficial for surviving desiccation (Fig. 3). Mutation in the stress response transcription factor genes MSN2 and MSN4, and in GIS1 whose gene product regulates the expression of stationary phase-specific genes,^31–33 resulted in a 2.9-, 3.9- and 7.6-fold decrease, respectively, in representation of the mutant strains in the library pool following three rounds of desiccation. Binding sites for Msn2p and Msn4p were also both statistically over-represented in the promoter sequences of genes whose expression was up-regulated during desiccation (Table 3A), and transcription of MSN2 was itself 1.9-fold up-regulated during this time (see ESI†, S2). Expression of MSN4, however, was down-regulated by a similar amount. The activities of Msn2p, Msn4p and Gis1p are regulated by the Rim15p kinase, and ultimately via activity of the Ras/PKA and Tor protein kinases.²⁸RIM15, together with YAK1 encoding a kinase negatively regulated by PKA that phosphorylates Msn2p,³⁴ were both massively under-represented in the desiccated mutant library pool, by 264- and 166-fold, respectively. Transcription of both these genes was also up-regulated about 1.8-fold upon desiccation (see ESI†, S2). These results were confirmed by analysis of the survival of mutant strains grown and desiccated individually; all strains tested displayed a marked reduction in the proportion of cells surviving (Fig. 3b). Mutants in TPK1 and BCY1, encoding catalytic and regulatory subunits of the PKA kinase complex, similarly showed an increased sensitivity to desiccation, and expression of these genes, along with the adenylate cyclase gene CYR1, was induced about 1.7-fold following the start of drying (see ESI†, S2). In contrast, RGS2 encoding a negative regulator of glucose-induced cAMP signaling was down-regulated upon drying.

Fig. 3 Elements of the Ras/PKA and Tor nutrient sensing pathways are important for the survival of desiccation. (a) Members of the network identified in a phenomics screen for mutants with reduced capability to survive desiccation stress. (b) Changes in desiccation tolerance following mutation in selected members of the network. Strains derived from BY4741 were grown individually to the post-diauxic phase of growth in YEPD media before drying. The proportion of cells surviving desiccation was measured as detailed in the Experimental, and the mean of three replicate experiments is shown. Standard deviations are indicated by the error bars. Yak1p and Rim15p are protein kinases whose activity is regulated by the PKA and/or Tor kinases. Msn2p, Msn4p and Gis1p are transcription factors controlled by phosphorylation by Rim15p. Tpk1p is a catalytic subunit of the PKA kinase complex, while Bcy1p is a regulatory subunit.

The hypothesis that activity of the Msn2p/4p transcriptional regulators is important for desiccation tolerance is also supported by the effect of deleting REG2, or the 14-3-3 protein-encoding gene, BMH1. Both mutations are predicted to increase the activity of Msn2/4p, and both mutants displayed an increased desiccation tolerance; 81% (±1%) and 86% (±1%), respectively, compared to 70% (±6%) for the parental strain (ESI†, Table S2). The bmh1 deletion also markedly improved survival of exponential-phase cells when subjected to desiccation, raising it from 3% ± 2% in BY4741 to 17% ± 2% in the mutant strain.

Nitrogen metabolism is reorganized during desiccation stress

Several genes with important functions related to nitrogen metabolism (CAR1, MEP1, GAP1, PUT4) were strongly and consistently up-regulated in the 120 min following the onset of desiccation (see ESI†, S2). Closer inspection revealed many more such genes that were significantly differentially expressed during the start of the drying process (Fig. 4), indicating a reorganization of nitrogen metabolism in the cultures during this time. Transcription of the general amino acid permease gene GAP1 was about 2.5-fold up-regulated in both the 60 and 120 min samples, while the proline permease PUT4 was similarly 4-fold up-regulated, and the AGP1 and AGP3 genes, which encode low-affinity amino acid permeases, 1.6- to 2.3-fold up-regulated. Transcription of TAT1 (encoding a transporter capable of importing valine, leucine, isoleucine, tryptophan, tyrosine, and histidine) was also activated, but expression of AQR1 (with a role in excreting excess amino acids) was down-regulated. NPR1, a gene encoding a protein kinase that stabilizes amino acid transport proteins, was also among the up-regulated genes. Clearly, amino acids are being actively scavenged by the drying cells. Polyamine transport processes were similarly up-regulated with transcripts of the TPO4 transporter gene increased 1.7-fold, along with PTK1 (2.5-fold) encoding a serine/threonine protein kinase which activates spermine uptake. The ammonia permease gene MEP1 was also more than 3-fold up-regulated in the 60 and 120 min samples, although (paradoxically) the ATO3 gene encoding a transport protein with a putative ammonia export function was also 2.5- to 3-fold induced in each of the 30, 60 and 120 min samples. Expression of four genes encoding enzymes for amino acid catabolism was up-regulated, including the ammoniagenic Gdh1p and Dsd1p, while the amino acid biosynthesis genes, SER2, THR1, HOM3, were repressed. Interestingly however, the arginine biosynthesis genes ARG3 (4-fold) and ARG1 (2-fold) were both up-regulated, as was arginase CAR1 (about 2-fold) and the PRO3 gene required for the final step in proline biosynthesis (1.6-fold). Both Car1p and Pro3p participate in the conversion of arginine to proline via ornithine, while Arg1p, Arg3p and Car1p form a cycle involving ornithine, arginine and citrulline that consumes glutamine and generates ammonia via urea (Fig. 4b).

Fig. 4 Changes in gene expression suggest a reorganisation of cellular nitrogen metabolism at the onset of desiccation. (a) The genes indicated were significantly differentially expressed and 1.5- to 4.8-fold up-regulated (GAP1, MEP1, PUT4, AGP1, AVT1, TAT1, AGP3, ATO3, ARG3, GDH2, CAR1, DSD1, PDC6, NPR1, ARG82, TPO4, PTK1, PRO3, ARG1), or 1.5- to 2.5-fold down-regulated (AQR1, HOM3, SER2, THR1) in the first 120 min of air drying. (b) Genes encoding key enzymes in arginine, ornithine and proline metabolism were up-regulated during drying. Transcripts of the indicated genes were significantly up-regulated in the 120 min following desiccation, as stated in the text.

The nitrogen discrimination pathway reorients gene expression to enable metabolism of alternate nitrogen sources, and Gln3p is a key transcriptional activator in this pathway, positively regulating genes that are subject to nitrogen catabolite repression.^35,36 Gln3p activity is modulated by interaction with Ure2p. Phenomics analysis indicated that both the GLN3 and URE2 genes are beneficial for the survival of desiccation in the post-diauxic growth phase cells, along with RAS2, which encodes a GTP-binding protein that regulates the nitrogen starvation response via its effect on adenylate cyclase (see ESI†, S1). Analysis of a RAS2 deletion mutant confirmed a severe reduction in its ability to survive desiccation; 10% survival in contrast to 70% in the parental strain (see ESI†, Table S2).

Desiccation stress shows similarity to other stress responses

Down-regulation of the translational machinery is a common feature of the response to many environmental stresses.^8,9 GO analysis of genes significantly down-regulated during desiccation identified ribosome biogenesis (GO:0042254) as the most significantly over-represented term, with 81 associated genes repressed (Table 2). This is also reflected in analysis of the down-regulated set of genes for statistically over-represented transcription factor binding sites which identified sites for Rap1p, Sfp1p and Fhl1p, all regulators of ribosome biogenesis and ribosomal protein gene expression (Table 3B). Gasch et al.⁹ grouped genes responding similarly to a range of environmental challenges (heat shock, salt stress, oxidative stress etc.) into a common category: the environmental stress response (ESR). In the ESR, transcription of approximately 600 genes is repressed following stress challenge, and about 300 genes are induced. Of the 484 genes up-regulated in response to drying during this study 71 belong to the induced ESR gene set, while 36% (128) of the 330 down-regulated genes are common to the repressed ESR response. ESR genes therefore form an appreciable fraction of those regulated by desiccation stress, a fact reflected in the roles of the Msn2p and Msn4p transcription factors which play a major role in mediating the ESR,^37,38 in the survival of desiccation (see above).

Phenomics analysis looking for genes beneficial to desiccation tolerance identified several transcriptional regulators or protein kinases involved in the oxidative (AFT2, SKN7) and osmotic (HAL9, MSN1, MSN2, MSN4, HOG1, PBS2, SSK2) stress responses. Transcription factor binding sites for regulators with roles in modulating gene expression in response to these stresses (Sut1p, Cin5p, Skn7p, and Sko1p) were also significantly over-represented in the list of genes up-regulated in response to desiccation (see Table 3A). However, when strains carrying deletions in the SKO1, SKN7, HOG1, PBS2 and SSK2 genes were tested individually, none showed any significant change in sensitivity to desiccation (see ESI†, Table S2). Other genes with reported roles in osmotic or oxidative stress (RLM1, HOT1, GPD1, GPD2, HOR2) were similarly not required for desiccation survival when tested in isolation, although RHR2 and YAP1 mutants were severely and moderately affected, respectively (see ESI†, Table S2).

Gene expression in response to stress is predominantly SAGA-dependent,^39–41 and genes involved in transcriptional regulation and elongation in the SAGA transcriptional coactivator complex (ADA2, SGF11, SPT8) were identified in the phenomics study, together with genes for the RNA polymerase II mediator complex (SSN2, SIN4, MED1, SRB2, SRB8; see ESI†, S1).

Discussion

A global search for genes involved in the response of yeast to desiccation stress was performed using a combined phenomics and transcriptomics approach. Both methods implicated autophagic processes as being an important part of the response and beneficial to survival of the drying process. Desiccation is highly damaging to macromolecules, causing denaturation of proteins and formation of aggregates which are potentially toxic to the cell.⁴² Clearance of such aggregates, and damaged organelles, is therefore likely to be crucial for prolonged existence under these conditions, and for the resumption of growth following rehydration. The cytoprotective benefits of autophagy have been demonstrated in a recent study where it was found to be central to the enhanced longevity and suppression of necrosis observed following treatment of yeast (and other eukaryotic cells) with the polyamine spermidine.⁴³ Spermidine up-regulates autophagy-related transcripts by altering chromatin acetylation through inhibition of histone acetyltransferase enzymes, but only mediates lifespan extension in autophagy-competent cells. Whether the autophagy induced in this study as a result of desiccation is mediated to any extent by an increase in the intracellular spermidine concentration is unclear, but gross changes in nitrogen metabolism in the drying yeast cells were evident (see below). However, the phenomics study did not identify genes required for spermidine biosynthesis (SPE1, SPE3) as being beneficial for desiccation survival. Strikingly, a significant proportion of genes indicated in the phenomics study as being beneficial to desiccation tolerance relates to mitochondrial structure and function, reflecting the essential role of respiratory metabolism in post-diauxic phase cultures. This may be a result of the experimental design, which involves three cycles of growth to the post-diauxic phase, but an important role in the desiccation response cannot be excluded. A previous study performed using an air-drying protocol similar to that used in industrial processes has also reported that mitochondrial function was important to survival.⁴⁴

Perhaps related to the induction of autophagy at the onset of desiccation, the transcriptome analysis suggested a significant reorganization of cellular nitrogen metabolism. Amino acid, polyamine, and ammonium import systems were induced, coupled with the repression of certain amino acid biosynthesis genes and up-regulation of catabolism genes. This suggests the scavenging of amino acid pools, made available by intracellular protein degradation or cell lysis, and their utilization for processes other than protein synthesis. Indeed, the concomitant reduction in transcripts for constructing the translational machinery indicates a switch away from cellular growth towards processes necessary for stress survival. Interestingly, genes encoding enzymes with functions in the metabolism of ornithine, arginine, and proline were significantly up-regulated during desiccation (see Fig. 4). Arginine has been shown to accumulate during the drying of desiccation-tolerant leaves of the ‘resurrection’ plant Sporobolus stapfianus, where it is proposed to serve as an important nitrogen and carbon reservoir useful during rehydration.⁴⁵ In yeast, arginine and ornithine can accumulate to high concentrations in the vacuole,^46–48 and it is possible they play a similar storage role for use during desiccation or rehydration. Ornithine is also a key precursor to the polyamines spermidine and putrescine. In Arabidopsis thaliana, proline acts as an osmolyte and helps protect against dehydration stress.⁴⁸ The transcriptomics results in this study suggest an activation of transport and metabolic processes to increase the intracellular concentration of proline during desiccation of yeast, perhaps with a similar purpose.

Desiccation is a complex stress involving lack of nutrients, production of reactive oxygen species, high osmolarity (due to concentration of the cytoplasm) and, in a natural environment, increased exposure to UV radiation.^49,50 The transcriptional response of desiccation-tolerant post-diauxic growth phase cells to desiccation stress comprised about 12% of the yeast genome, including about 200 ESR genes previously shown to participate in a common response to a variety of stress conditions such as heat, pH, osmotic and oxidative stresses.⁹ The induction of 71 of the positively regulated ESR genes during desiccation, and the reported down-regulation of ESR genes on rehydration of dried yeast,⁵¹ suggests a role for the general stress transcription factors Msn2p and Msn4p in the desiccation stress response.^37,38,52,53 Indeed, the MSN2 and MSN4 genes were identified as beneficial for the survival of desiccation in the phenomics screen, and in survival studies performed individually on the mutant strains. These results are in contrast to those of a previous study examining the transcriptional response to desiccation stress which found insufficient evidence to support a role for the ESR.²¹ However, that study used desiccation-intolerant exponential phase cells, and the experimental conditions in the current work are expected to more accurately describe the desiccation response. The proposed significance in the survival of desiccation of the hydrophilic protein genes SIP18 and GRE1, up-regulated in the exponential phase study,²¹ is similarly not borne out by our results: neither gene was identified in the phenomics or transcriptomics analyses, and mutant strains did not exhibit increased sensitivity to desiccation (see ESI†, Table S2).

Genes upstream of Msn2p/4p in the regulatory hierarchy were also found to respond to desiccation stress, and to be helpful for the survival of drying, suggesting an important role for the Ras/PKA and Tor signaling pathways in adjusting metabolism to cope with the change in conditions. Snf1 kinase activity, essential for carbon catabolite derepression in yeast (reviewed in refs 28,54), was similarly required for desiccation survival. This perhaps emphasizes the central role of metabolic reprogramming at diauxie which appears to switch yeast from an essentially desiccation-intolerant to a desiccation-tolerant organism. The observed increase in desiccation survival of a mutant strain expected to possess increased Snf1 kinase activity is particularly interesting in this respect. The Snf1 kinase also plays a role in regulation of the environmental stress response, and the responses to salt and heat stresses.⁵⁴ Nrg1p and Nrg2p are two transcriptional repressors that repress many STRE genes during normal growth, many of which are up-regulated in a reg1Δ mutant in a Snf1–Gal83 kinase dependent manner.⁵⁵ The nrg1Δnrg2Δ double mutant is reported to be more tolerant of osmotic, oxidative and pH stresses^55,56 and it would be interesting to also assess its desiccation tolerance. Mutation in REG1 significantly improved desiccation survival in the present study.

Antioxidant activity seems to be a feature of desiccation tolerance in several species^57,58 and is therefore also likely to be so in yeast. Several studies report an increase in reactive oxygen species during dehydration of yeast strains, resulting in denaturation of proteins, nucleic acid damage and lipid peroxidation.^49,59–62 Only a limited number of genes involved in combating oxidative stress were detected in the phenomics and transcriptomics analysis presented here however, and only one of these, that encoding the B-zip transcriptional regulator Yap1p, was verified as being beneficial for the survival of desiccation. Deletion of the catalase gene CTT1 (3-fold up-regulated after the onset of drying) did not hinder desiccation tolerance, nor did removal of either the AFT2 or SKN7 transcription factor genes. This could reflect functional redundancy, or perhaps differences in the drying processes used. Fluidized bed air-drying, similar to that used in the industrial process, is likely to expose cells to greater oxygen damage than drying as a filtered cake, as was used in this study. Overexpression of the superoxide dismutase genes SOD1 and SOD2 has previously been shown to increase desiccation tolerance in yeast,⁶¹ and although neither gene was identified by the genome-wide approaches used here, targeted deletion of either gene did moderately increase sensitivity to desiccation (see ESI†, Table S2), consistent with a protective role for these antioxidant enzymes.

Surprisingly, deletion of the HOG pathway genes (HOG1, PBS2, and HOT1) did not appear to decrease desiccation tolerance (see ESI†, Table S2) in spite of the essential role of this pathway in resistance to osmotic stress, which is arguably similar to desiccation. Interestingly, however, while most of the HOG pathway mutants were not sensitive to desiccation when tested individually, several were indicated by the phenomics experiment. This raises the intriguing possibility that the HOG pathway is important for tolerance in a population of variant strains, perhaps conferring a competitive advantage that is not apparent in pure cultures. Despite this discrepancy, in the limited verification analysis undertaken here, the phenomics approach produced a 75% success rate in identifying genes whose absence adversely affected desiccation tolerance when tested individually. In contrast, similar verification performed on genes suggested by differential regulation in the transcriptomics study produced a 33% success rate (see ESI†, Table S2). This is consistent with the conclusions of previous studies, which observed that relatively few stress-responsive genes are stress-combative,^12–16 and may reflect the effects of functional redundancy, which would require double-mutant strains to reveal a phenotype.

Several genes highlighted by the deletion library study merit further investigation. TDH1 encodes one of three isoforms of glyceraldehyde-3-phosphate dehydrogenase (GAPDH), but only tdh1 mutants are desiccation intolerant: tdh2Δ and tdh3Δ are unaffected (see ESI†, Table S2). The hypothesis that TDH1 is functionally distinct from TDH2 and TDH3 is supported by previous experimental observations.^63–65 In addition, mammalian GAPDH has been shown to have diverse functions in translational control, microtubule binding, apoptosis, DNA replication and repair, and telomere maintenance.^66,67 In Schizosaccharomyces pombe, GAPDH has been suggested to be involved in sensing and signaling oxidative stress.⁶⁸ Additional functions beyond glycolysis for TDH1 in yeast would therefore not be unusual, and it could play a role in desiccation tolerance or sensing.

Conclusion

Desiccation tolerance in S. cerevisiae is a complex, multi-factorial process with contributions from several hundred genes. We employed both transcriptomics and phenomic analyses in an attempt to unravel this complexity. Although a rigorous statistical comparison of the two methods was beyond the scope of our study, phenomics screening generally proved more effective than transcriptome analysis in homing in on individual genes required for desiccation tolerance. However, combining the approaches afforded the deepest insight into the response of yeast to drying. Autophagy and pathways governed by the Ras/PKA, Tor and Snf1 protein kinases, as well as the general stress response, all appear to play important roles in adjusting metabolism to achieve a cellular state capable of surviving periods of severe dryness. Modulation of nitrogen metabolism away from protein production and towards the accumulation of amino acids that may play a role in desiccation survival also appears to be an important part of the adaptation process.

Experimental

Strains and culture conditions

The strains used in this study are BY4741 (MATa; his3Δ1; leu2Δ0; met15Δ0; ura3Δ0), BY4743 (MATa/MATα; his3Δ1/his3Δ1; leu2Δ0/leu2Δ0; met15Δ0/MET15; LYS2/lys2Δ0; ura3Δ0/ura3Δ0) and their deletion mutants (haploid or homozygous diploid) obtained from EUROSCARF. Some of the mutants were verified by PCR for deletion of the relevant gene using specific primers for the gene and the replacement cassette. Cells were grown in YEPD medium (1% w/v yeast extract, 2% w/v peptone, 2% w/v glucose) at 30 °C with continuous shaking up to mid-exponential phase (A₆₀₀ ≈ 1.5 ± 0.2) or post-diauxic phase (A₆₀₀ ≈ 3 ± 0.2) preceding true stationary phase.

Transcriptomics and phenomics analysis

For transcriptome analysis, BY4741 was grown in two independent replicate cultures up to late post-diauxic phase in YEPD. A 2 ml volume of culture was taken as the control 0 min sample, filtered through 0.45 μm cellulose nitrate membrane filters (Whatman) and dried in an incubator at 30 °C for 30, 60 and 120 min, as described previously.⁴ RNA was prepared using the Ribopure Yeast kit (Ambion) with a phenol–chloroform–acetone mixture according to the manufacturer's protocol, treated with DNase at 37 °C for 1 h. RNA quality was checked by agarose gel electrophoresis before hybridizing to Affymetrix GeneChip Yeast Genome S98 arrays. Technical replicate hybridisations were performed, and raw microarray data are available from ArrayExpress, accession number E-MEXP-2818.

For the phenomics analysis, the homozygous diploid deletion mutant collection was used.²³ The homozygous diploid deletants are the products of two independently isolated haploid parents, thus effects due to extraneous mutations that arose during the deletion process should be minimized. The library was grown in YEPD to late post-diauxic phase in two independent replicate cultures and two samples were taken, one as the untreated control and the other for desiccation. Desiccation was achieved as previously described, drying overnight to reduce the residual moisture content to 6–7% w/w.⁴ The desiccated sample was then subjected to two more rounds of growth and drying before taking the treated sample after the third re-culture at late post-diauxic phase. Genomic DNA was isolated from the samples, and tags amplified and hybridized to tag-3 microarrays as previously described.^69,70 Raw data are available from ArrayExpress, accession number E-MEXP-2395.

Data analysis

For the phenomics analysis, Affymetrix scanner output in standard CEL file format was processed and analysed as previously described.^69,70

For the transcriptomics analysis, CEL files were normalised using RMA as implemented in the R/Bioconductor suite of tools.⁷¹ Normalised values from technical replicate hybridizations were averaged and significantly differentially expressed genes (p < 0.05) identified in the averaged data using LIMMA.⁷² All gene ontology analysis was performed using the Ontologizer software,²⁴ performing MGSA or term-for-term hypergeometric testing as stated in the text. Benjamini and Hochberg multiple testing correction was used in the term-for-term analyses. Over-representation of transcription factor binding sites in differentially expressed gene sets was determined using the Ceres webtool.²⁹

Measurement of tolerance to desiccation

Deletion mutants selected from the EUROSCARF BY4741 haploid deletion mutant library were grown to the stated growth phase in YEPD, and aliquots dried overnight at 30 °C. Viability was assessed in triplicate using flow cytometry analysis of propidium iodide stained cells as previously described.⁴

Abbreviations

ESR	environmental stress response
HOG	high osmolarity glycerol
PKA	protein kinase A
YEPD	yeast extract peptone dextrose

Acknowledgements

SR thanks the Cambridge Commonwealth Trust and St John's College, Cambridge for funding. KG acknowledges a scholarship from the Wellcome Trust. This research was supported by a European Research Council Advanced Investigator award (233232) to AT, by a grant from the Natural Environment Research Council to SGO, and by the COGEME Transcriptome Resource Facility (coordinator SGO) funded by BBSRC under the Investigating Gene Function Initiative.

References

1. J. S. Clegg, Comp. Biochem. Physiol., B: Biochem. Mol. Biol., 2001, 128, 613–624.
2. A. Tunnacliffe and J. Lapinski, Philos. Trans. R. Soc. London, Ser. B, 2003, 358, 1755–1771.
3. G. M. Gadd, K. Chalmers and R. H. Reed, FEMS Microbiol. Lett., 1987, 48, 249–254.
4. S. Ratnakumar and A. Tunnacliffe, FEMS Yeast Res., 2006, 6, 902–913.
5. A. Goffeau, B. G. Barrell, H. Bussey, R. W. Davis, B. Dujon, H. Feldmann, F. Galibert, J. D. Hoheisel, C. Jacq, M. Johnston, E. J. Louis, H. W. Mewes, Y. Murakami, P. Philippsen, H. Tettelin and S. G. Oliver, Science, 1996, 274, 546 ,563–567.
6. M. E. Hillenmeyer, E. Fung, J. Wildenhain, S. E. Pierce, S. Hoon, W. Lee, M. Proctor, R. P. St Onge, M. Tyers, D. Koller, R. B. Altman, R. W. Davis, C. Nislow and G. Giaever, Science, 2008, 320, 362–365.
7. P. K. Herman, Curr. Opin. Microbiol., 2002, 5, 602–607.
8. H. C. Causton, B. Ren, S. S. Koh, C. T. Harbison, E. Kanin, E. G. Jennings, T. I. Lee, H. L. True, E. S. Lander and R. A. Young, Mol. Biol. Cell, 2001, 12, 323–337.
9. A. P. Gasch, P. T. Spellman, C. M. Kao, O. Carmel-Harel, M. B. Eisen, G. Storz, D. Botstein and P. O. Brown, Mol. Biol. Cell, 2000, 11, 4241–4257.
10. M. J. Brauer, C. Huttenhower, E. M. Airoldi, R. Rosenstein, J. C. Matese, D. Gresham, V. M. Boer, O. G. Troyanskaya and D. Botstein, Mol. Biol. Cell, 2008, 19, 352–367.
11. J. I. Castrillo, L. A. Zeef, D. C. Hoyle, N. Zhang, A. Hayes, D. C. Gardner, M. J. Cornell, J. Petty, L. Hakes, L. Wardleworth, B. Rash, M. Brown, W. B. Dunn, D. Broadhurst, K. O'Donoghue, S. S. Hester, T. P. Dunkley, S. R. Hart, N. Swainston, P. Li, S. J. Gaskell, N. W. Paton, K. S. Lilley, D. B. Kell and S. G. Oliver, J. Biol., 2007, 6, 4.
12. E. A. Winzeler, D. D. Shoemaker, A. Astromoff, H. Liang, K. Anderson, B. Andre, R. Bangham, R. Benito, J. D. Boeke, H. Bussey, A. M. Chu, C. Connelly, K. Davis, F. Dietrich, S. W. Dow, M. El Bakkoury, F. Foury, S. H. Friend, E. Gentalen, G. Giaever, J. H. Hegemann, T. Jones, M. Laub, H. Liao, N. Liebundguth, D. J. Lockhart, A. Lucau-Danila, M. Lussier, N. M'Rabet, P. Menard, M. Mittmann, C. Pai, C. Rebischung, J. L. Revuelta, L. Riles, C. J. Roberts, P. Ross-MacDonald, B. Scherens, M. Snyder, S. Sookhai-Mahadeo, R. K. Storms, S. Veronneau, M. Voet, G. Volckaert, T. R. Ward, R. Wysocki, G. S. Yen, K. Yu, K. Zimmermann, P. Philippsen, M. Johnston and R. W. Davis, Science, 1999, 285, 901–906.
13. G. W. Birrell, J. A. Brown, H. I. Wu, G. Giaever, A. M. Chu, R. W. Davis and J. M. Brown, Proc. Natl. Acad. Sci. U. S. A., 2002, 99, 8778–8783.
14. J. Warringer, E. Ericson, L. Fernandez, O. Nerman and A. Blomberg, Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 15724–15729.
15. G. W. Thorpe, C. S. Fong, N. Alic, V. J. Higgins and I. W. Dawes, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 6564–6569.
16. S. L. Tai, I. Snoek, M. A. Luttik, M. J. Almering, M. C. Walsh, J. T. Pronk and J. M. Daran, Microbiology (Reading, U. K.), 2007, 153, 877–886.
17. G. Giaever, D. D. Shoemaker, T. W. Jones, H. Liang, E. A. Winzeler, A. Astromoff and R. W. Davis, Nat. Genet., 1999, 21, 278–283.
18. M. M. Bianchi, S. Ngo, M. Vandenbol, G. Sartori, A. Morlupi, C. Ricci, S. Stefani, G. B. Morlino, F. Hilger, G. Carignani, P. P. Slonimski and L. Frontali, Yeast, 2001, 18, 1397–1412.
19. G. W. Birrell, G. Giaever, A. M. Chu, R. W. Davis and J. M. Brown, Proc. Natl. Acad. Sci. U. S. A., 2001, 98, 12608–12613.
20. V. J. Higgins, N. Alic, G. W. Thorpe, M. Breitenbach, V. Larsson and I. W. Dawes, Yeast, 2002, 19, 203–214.
21. J. Singh, D. Kumar, N. Ramakrishnan, V. Singhal, J. Jervis, J. F. Garst, S. M. Slaughter, A. M. DeSantis, M. Potts and R. F. Helm, Appl. Environ. Microbiol., 2005, 71, 8752–8763.
22. A. Wiemken, Antonie Van Leeuwenhoek, 1990, 58, 209–217.
23. G. Giaever, A. M. Chu, L. Ni, C. Connelly, L. Riles, S. Veronneau, S. Dow, A. Lucau-Danila, K. Anderson, B. Andre, A. P. Arkin, A. Astromoff, M. El-Bakkoury, R. Bangham, R. Benito, S. Brachat, S. Campanaro, M. Curtiss, K. Davis, A. Deutschbauer, K. D. Entian, P. Flaherty, F. Foury, D. J. Garfinkel, M. Gerstein, D. Gotte, U. Guldener, J. H. Hegemann, S. Hempel, Z. Herman, D. F. Jaramillo, D. E. Kelly, S. L. Kelly, P. Kotter, D. LaBonte, D. C. Lamb, N. Lan, H. Liang, H. Liao, L. Liu, C. Luo, M. Lussier, R. Mao, P. Menard, S. L. Ooi, J. L. Revuelta, C. J. Roberts, M. Rose, P. Ross-Macdonald, B. Scherens, G. Schimmack, B. Shafer, D. D. Shoemaker, S. Sookhai-Mahadeo, R. K. Storms, J. N. Strathern, G. Valle, M. Voet, G. Volckaert, C. Y. Wang, T. R. Ward, J. Wilhelmy, E. A. Winzeler, Y. Yang, G. Yen, E. Youngman, K. Yu, H. Bussey, J. D. Boeke, M. Snyder, P. Philippsen, R. W. Davis and M. Johnston, Nature, 2002, 418, 387–391.
24. S. Bauer, J. Gagneur and P. N. Robinson, Nucleic Acids Res., 2010, 38, 3523–3532.
25. S. L. Forsburg and L. Guarente, Genes Dev., 1989, 3, 1166–1178.
26. J. T. Olesen and L. Guarente, Genes Dev., 1990, 4, 1714–1729.
27. M. Rosenkrantz, C. S. Kell, E. A. Pennell and L. J. Devenish, Mol. Microbiol., 1994, 13, 119–131.
28. S. Zaman, S. I. Lippman, X. Zhao and J. R. Broach, Annu. Rev. Genet., 2008, 42, 27–81.
29. R. T. Morris, T. R. O'Connor and J. J. Wyrick, Bioinformatics, 2010, 26, 168–174.
30. K. M. Dombek, N. Kacherovsky and E. T. Young, J. Biol. Chem., 2004, 279, 39165–39174.
31. J. Wu, N. Zhang, A. Hayes, K. Panoutsopoulou and S. G. Oliver, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 3148–3153.
32. A. Reinders, N. Burckert, T. Boller, A. Wiemken and C. De Virgilio, Genes Dev., 1998, 12, 2943–2955.
33. I. Pedruzzi, N. Burckert, P. Egger and C. De Virgilio, EMBO J., 2000, 19, 2569–2579.
34. P. Lee, B. R. Cho, H. S. Joo and J. S. Hahn, Mol. Microbiol., 2008, 70, 882–895.
35. W. E. Courchesne and B. Magasanik, J. Bacteriol., 1988, 170, 708–713.
36. D. Blinder and B. Magasanik, J. Bacteriol., 1995, 177, 4190–4193.
37. M. T. Martinez-Pastor, G. Marchler, C. Schuller, A. Marchler-Bauer, H. Ruis and F. Estruch, EMBO J., 1996, 15, 2227–2235.
38. A. P. Schmitt and K. McEntee, Proc. Natl. Acad. Sci. U. S. A., 1996, 93, 5777–5782.
39. K. L. Huisinga and B. F. Pugh, Mol. Cells, 2004, 13, 573–585.
40. S. J. Zanton and B. F. Pugh, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 16843–16848.
41. M. Zapater, M. Sohrmann, M. Peter, F. Posas and E. de Nadal, Mol. Cell. Biol., 2007, 27, 3900–3910.
42. K. Goyal, L. J. Walton and A. Tunnacliffe, Biochem. J., 2005, 388, 151–157.
43. T. Eisenberg, H. Knauer, A. Schauer, S. Buttner, C. Ruckenstuhl, D. Carmona-Gutierrez, J. Ring, S. Schroeder, C. Magnes, L. Antonacci, H. Fussi, L. Deszcz, R. Hartl, E. Schraml, A. Criollo, E. Megalou, D. Weiskopf, P. Laun, G. Heeren, M. Breitenbach, B. Grubeck-Loebenstein, E. Herker, B. Fahrenkrog, K. U. Frohlich, F. Sinner, N. Tavernarakis, N. Minois, G. Kroemer and F. Madeo, Nat. Cell Biol., 2009, 11, 1305–1314.
44. J. Shima, A. Ando and H. Takagi, Yeast, 2008, 25, 179–190.
45. T. Martinelli, A. Whittaker, A. Bochicchio, C. Vazzana, A. Suzuki and C. Masclaux-Daubresse, J. Exp. Bot., 2007, 58, 3037–3046.
46. K. Kitamoto, K. Yoshizawa, Y. Ohsumi and Y. Anraku, J. Bacteriol., 1988, 170, 2683–2686.
47. R. H. Davis, Microbiol. Rev., 1986, 50, 280–313.
48. T. Nanjo, M. Kobayashi, Y. Yoshiba, Y. Sanada, K. Wada, H. Tsukaya, Y. Kakubari, K. Yamaguchi-Shinozaki and K. Shinozaki, Plant J., 1999, 18, 185–193.
49. M. B. Franca, A. D. Panek and E. C. A. Eleutherio, Comp. Biochem. Physiol., A: Mol. Integr. Physiol., 2007, 146, 621–631.
50. V. Mattimore and J. R. Battista, J. Bacteriol., 1996, 178, 633–637.
51. T. Rossignol, O. Postaire, J. Storai and B. Blondin, Appl. Microbiol. Biotechnol., 2006, 71, 699–712.
52. E. Boy-Marcotte, G. Lagniel, M. Perrot, F. Bussereau, A. Boudsocq, M. Jacquet and J. Labarre, Mol. Microbiol., 1999, 33, 274–283.
53. E. Boy-Marcotte, M. Perrot, F. Bussereau, H. Boucherie and M. Jacquet, J. Bacteriol., 1998, 180, 1044–1052.
54. P. Sanz, Biochem. Soc. Trans., 2003, 31, 178–181.
55. V. K. Vyas, C. D. Berkey, T. Miyao and M. Carlson, Eukaryotic Cell, 2005, 4, 1882–1891.
56. T. M. Lamb and A. P. Mitchell, Mol. Cell. Biol., 2003, 23, 677–686.
57. I. Kranner and S. Birtic, Integr. Comp. Biol., 2005, 45, 734–740.
58. O. Leprince, C. W. Vertucci, G. A. F. Hendry and N. M. Atherton, Physiol. Plant., 1995, 94, 233–240.
59. E. Garre, F. Raginel, A. Palacios, A. Julien and E. Matallana, Int. J. Food Microbiol., 2010, 136, 295–303.
60. M. B. Franca, A. D. Panek and E. C. A. Eleutherio, Cell Stress Chaperones, 2005, 10, 167–170.
61. J. Pereira Ede, A. D. Panek and E. C. Eleutherio, Cell Stress Chaperones, 2003, 8, 120–124.
62. S. Espindola Ade, D. S. Gomes, A. D. Panek and E. C. Eleutherio, Cryobiology, 2003, 47, 236–241.
63. L. McAlister and M. J. Holland, J. Biol. Chem., 1985, 260, 15019–15027.
64. L. McAlister and M. J. Holland, J. Biol. Chem., 1985, 260, 15013–15018.
65. H. Boucherie, N. Bataille, I. T. Fitch, M. Perrot and M. F. Tuite, FEMS Microbiol. Lett., 1995, 125, 127–133.
66. M. A. Sirover, J. Cell. Biochem., 1997, 66, 133–140.
67. M. A. Sirover, J. Cell. Biochem., 2005, 95, 45–52.
68. S. Morigasaki, K. Shimada, A. Ikner, M. Yanagida and K. Shiozaki, Mol. Cells, 2008, 30, 108–113.
69. D. Delneri, D. C. Hoyle, K. Gkargkas, E. J. Cross, B. Rash, L. Zeef, H. S. Leong, H. M. Davey, A. Hayes, D. B. Kell, G. W. Griffith and S. G. Oliver, Nat. Genet., 2008, 40, 113–117.
70. S. Holland, E. Lodwig, T. Sideri, T. Reader, I. Clarke, K. Gkargkas, D. C. Hoyle, D. Delneri, S. G. Oliver and S. V. Avery, Genome Biol., 2007, 8, R268.
71. R. C. Gentleman, V. J. Carey, D. M. Bates, B. Bolstad, M. Dettling, S. Dudoit, B. Ellis, L. Gautier, Y. Ge, J. Gentry, K. Hornik, T. Hothorn, W. Huber, S. Iacus, R. Irizarry, F. Leisch, C. Li, M. Maechler, A. J. Rossini, G. Sawitzki, C. Smith, G. Smyth, L. Tierney, J. Y. Yang and J. Zhang, Genome Biol., 2004, 5, R80.
72. G. K. Smyth, Stat. Appl. Genet. Mol. Biol., 2004, 3 , article3, DOI: 10.2202/1544-6115.1027.

---

Footnotes

† Electronic supplementary information (ESI) available: S1 lists all genes identified in the phenomics analysis as being beneficial for surviving desiccation; S2 details genes significantly differentially expressed in response to drying. S3 and S4 relate to GO analyses of genes from S1 and S2; and Tables S1 and S2 relate to the analysis of individual mutant strains for tolerance to desiccation. See DOI: 10.1039/c0mb00114g

‡ Present address: The Centre for Plant Integrative Biology, School of Biosciences, University of Nottingham, Sutton Bonington Campus, Loughborough, UK.

---