Transcriptional Response to Hypoxia in the Aquatic Fungus Blastocladiella emersonii^,

Abstract

Global gene expression analysis was carried out with Blastocladiella emersonii cells subjected to oxygen deprivation (hypoxia) using cDNA microarrays. In experiments of gradual hypoxia (gradual decrease in dissolved oxygen) and direct hypoxia (direct decrease in dissolved oxygen), about 650 differentially expressed genes were observed. A total of 534 genes were affected directly or indirectly by oxygen availability, as they showed recovery to normal expression levels or a tendency to recover when cells were reoxygenated. In addition to modulating many genes with no putative assigned function, B. emersonii cells respond to hypoxia by readjusting the expression levels of genes responsible for energy production and consumption. At least transcriptionally, this fungus seems to favor anaerobic metabolism through the upregulation of genes encoding glycolytic enzymes and lactate dehydrogenase and the downregulation of most genes coding for tricarboxylic acid (TCA) cycle enzymes. Furthermore, genes involved in energy-costly processes, like protein synthesis, amino acid biosynthesis, protein folding, and transport, had their expression profiles predominantly downregulated during oxygen deprivation, indicating an energy-saving effort. Data also revealed similarities between the transcriptional profiles of cells under hypoxia and under iron(II) deprivation, suggesting that Fe²⁺ ion could have a role in oxygen sensing and/or response to hypoxia in B. emersonii. Additionally, treatment of fungal cells prior to hypoxia with the antibiotic geldanamycin, which negatively affects the stability of mammalian hypoxia transcription factor HIF-1α, caused a significant decrease in the levels of certain upregulated hypoxic genes.

Oxygen is essential for the survival of most eukaryotic organisms due to its crucial role in important biochemical and physiological processes, including energy production (in the form of ATP) from glucose by aerobic metabolism. In nature, oxygen deprivation challenges can occur with a significantly high frequency, depending on the organism lifestyle. Accordingly, organisms have evolved a complex set of cellular and molecular adaptive responses activated by decreases in oxygen availability (hypoxia or anoxia). These responses are adaptive because they can lead to changes in the physiological and metabolic status of the cells, which will allow them to cope with the stress associated with oxygen limitation and ultimately enhance cellular survival. Especially in aquatic habitats, hypoxia can be an important evolutionary driving force, resulting in both convergent and divergent physiological strategies for survival under low-oxygen conditions.

The unicellular eukaryote Saccharomyces cerevisiae has been used as a model system to understand many basic cellular functions and biochemical processes in eukaryotic systems. Hypoxia response studies are not an exception. Yeast responds to changes in O₂ availability by altering the expression of a number of oxygen-responsive genes, consisting of hypoxic genes that are transcriptionally activated during hypoxia and aerobic genes that are transcribed only during normoxia, when oxygen is plentiful (26, 37). The majority of yeast hypoxic and aerobic genes are controlled by transcription factors Rox1p and Hap1p, the former repressing hypoxic genes and the later activating aerobic genes, when oxygen levels are high (19, 26). Both transcription factors are dependent upon heme biosynthesis, which is an oxygen-dependent process. Mga2p is another putative oxygen-sensitive transcriptional regulator, which has been considered to be related to the mammalian transcription factor HIF-1 (hypoxia-inducible factor 1), the central regulator of hypoxic gene expression in metazoans, due to its similarity regarding activation by reactive oxygen species generated in the mitochondria (16) and to the mimetization of the hypoxic transcriptional response by cobalt chloride and iron chelator treatment (48), which are known for stabilizing HIF-1, as will be discussed below. However, the usefulness of S. cerevisiae as a model system for the study of higher eukaryotes' response to extreme hypoxia and anoxia is virtually limited by the fact that this yeast is a facultative anaerobe.

Transcription factor HIF-1 is known as the main regulator of oxygen homeostasis in metazoans, and up to now, a putative homologue of this transcription factor has not been described for fungi. In higher eukaryotes, HIF-1 mediates developmental and physiological pathways that either deliver O₂ to the cells or allow them to survive under O₂ deprivation. HIF-1 functions as a heterodimer composed of an O₂-regulated HIF-1α subunit and a constitutively expressed HIF-1β subunit. Under normoxic conditions, HIF-1α is continuously synthesized and degraded by the binding of the von Hippel-Lindau tumor suppressor protein (VHL), which targets HIF-1α to ubiquitination and proteasomal degradation. However, when oxygen becomes limiting (hypoxia), HIF-1α degradation is inhibited and the protein accumulates, forming a dimer with HIF-1β that is capable of binding to cis-acting hypoxia response elements in target genes and recruiting coactivator proteins, which leads to increased transcription of these genes. In normoxia, binding of VHL to HIF-1α is dependent upon its hydroxylation of residue Pro⁴⁰² or Pro⁵⁶⁴ or both by prolyl hydroxylase domain protein 2 (PHD2), a dioxygenase that utilizes O₂ and α-ketoglutarate as substrates while generating CO₂ and succinate. FIH-1 (factor inhibiting HIF-1) is another α-ketoglutarate-dependent dioxygenase that hydroxylates Asn⁸⁰³ and thereby blocks the interaction of HIF-1α with the coactivators p300 and CBP (42). The catalytic center of both PHD2 and FIH proteins contains Fe²⁺ (42), and for this reason, iron(II) deprivation is a known hypoxia-mimicking stimulus of HIF-1 stabilization.

HIF-1α can also be regulated in an O₂-independent manner by the receptor of activated protein kinase C (RACK1). It promotes the O₂/PHD/VHL-independent and proteasome-dependent degradation of HIF-1α by competing with heat shock protein HSP90, which has been found to play a role in protecting HIF-1α from proteasomal degradation by binding to the PAS-A domain of HIF-1α. In addition, inhibitors of HSP90 such as geldanamycin are known to induce HIF-1α degradation independently of oxygen concentration (29).

In the present report, the model of study is Blastocladiella emersonii, an aquatic fungus that belongs to the Blastocladiomycetes class, being located at the base of the fungal phylogenetic tree (18, 44). The life cycle of B. emersonii starts with the zoospore, a nongrowing motile cell that is responsible for the dispersal of the fungus. In the presence of appropriate stimuli, zoospores start the germination process, which is characterized by a number of drastic morphological and biochemical changes (31). The morphological events include retraction of the zoospore polar flagellum, construction a cell wall rich in chitin, fragmentation of the single giant mitochondrion into normal-sized ones, and formation of a germ tube whose ramification will give rise to a rhizoidal system through which the cell adheres to the substrate and nutrients are absorbed, among many other changes (31). At the end of this cell differentiation stage, the fungus enters the vegetative growth phase, characterized by nuclear division that is not accompanied by cell division, originating a multinucleated cell named zoosporangium. Nutrient starvation at any time during vegetative growth can induce the sporulation stage, another cell differentiation process, which culminates with the production and liberation to the medium of a number of zoospores, completing the cycle. During both germination and sporulation, a large proportion of B. emersonii genes have been shown to be differentially expressed, indicating the presence of important transcriptional control mechanisms at these stages, making this fungus a good model for gene regulation studies (40, 41, 52).

Although oxygen supply is essential for survival of B. emersonii, due to its aquatic saprophytic lifestyle the fungus seems to be adapted to low-aeration environments and thus can be an interesting system for studying the hypoxic stress response. In this sense, changes in gene expression in B. emersonii cells in response to different dissolved oxygen (DO) concentrations were analyzed using cDNA microarrays. Additionally, the presence of a possible HIF-1-like mechanism of oxygen sensing was investigated by comparing the transcriptional changes in B. emersonii cells subjected to hypoxia and iron deprivation as well as in cells treated with the antibiotic geldanamycin prior to exposure to hypoxia.

Data obtained indicated that B. emersonii responds to hypoxia by readjusting the expression levels of genes responsible for energy production and consumption. At least transcriptionally, this fungus seems to favor anaerobic metabolism through the upregulation of genes encoding glycolytic enzymes and lactate dehydrogenase, while most genes from tricarboxylic acid (TCA) cycle enzymes were downregulated or unchanged. Microarray experiments also revealed similarities in the transcriptional profiles of cells under hypoxia and under iron(II) deprivation, suggesting that these stresses are somehow related and that Fe²⁺ ion could have a role in the mechanism of oxygen sensing and/or response to hypoxia in B. emersonii. Furthermore, pretreatment of cells with the antibiotic geldanamycin prior to exposure to hypoxia caused a significant decrease in the expression levels of certain upregulated hypoxic genes, suggesting that this fungus could have a mechanism similar to that of the mammalian hypoxia transcription factor HIF-1α, whose stability is negatively affected by this antibiotic.

MATERIALS AND METHODS

Culture conditions of B. emersonii.

Growth of B. emersonii was performed at 28°C in DM3 medium (33) in a bioreactor equipped with pH and DO electrodes (Rosemount Analytical). The pH was measured online and kept constant by the automatic addition of 0.5 M NaOH or HCl using two peristaltic pumps (Watson-Marlow). A blend of air and nitrogen was sparged into the bioreactor at a fixed total flow rate (350 ml·min⁻¹). To control DO value, two mass flow controllers (Aalborg) were used to adjust the flow of each gas according to the measured DO value. To maintain both pH and DO value, we used a programmable logic controller (PLC; Unitronics), which operates automatically the peristaltic pumps and mass flow controllers to keep pH and DO value under the programmed conditions. Aliquots removed from the medium were filtered with 3-mm paper (Whatman), immediately frozen in liquid nitrogen, and stored at −70°C. Cells were never exposed to ambient air for more than 5 min before freezing. Glucose concentration in the filtrate was measured with an YSI 2700 Select biochemistry analyzer (Yellow Springs Instruments Co.), according to the manufacturer's instructions.

Gradual-hypoxia experiments.

The reactor containing 400 ml of DM3 medium was inoculated with 6·10⁵ zoospores·ml⁻¹ in normoxia (70% of oxygen saturation) at pH 6.8. The cells were grown for 6.5 h to reach the vegetative growth stage. After this, oxygen concentration in the culture was decreased from its initial value to 35 (C35), 17.5 (C17.5), 5 (C5), 1 (C1), and 0% (C0) of oxygen saturation in 1-h steps. In the 0% step (C0), the culture was sparged with pure nitrogen for 1 h and then the oxygen level was restored to its initial value, 70% saturation (C70r). At the end of each 1-hour step, aliquots were withdrawn from the culture vessel for isolation of total RNA. Samples were also taken at the different DO concentration steps for determination of extracellular glucose concentrations. The reference RNA used in microarrays was always at the initial 70% saturation (C70). Hybridizations were performed under all different oxygen conditions (C35, C17.5, C5, C1, C0, and C70r).

Direct-hypoxia experiments.

The culture conditions were the same as those for gradual hypoxia, but instead of a gradual decrease in dissolved oxygen concentration, the concentration was decreased directly from normoxia (70% saturation) to hypoxia (1% saturation). The culture was kept under hypoxia for 5 h, and aliquots were withdrawn after 1 h (C1_1h) and 5 h (C1_5h). After 5 h, oxygen levels were restored to 70% saturation for 1 h and an aliquot was also taken (C70r_1h). The reference RNA used in microarrays hybridizations was always the initial 70% saturation (C70). Hybridizations were performed using RNA from cells isolated at the C1_1h, C1_5h, and C70r_1h time points.

Iron(II) deprivation experiments.

B. emersonii was cultivated in the same bioreactor as that described for the oxygen deprivation experiments. Zoospores (6 × 10⁵ per ml) were inoculated in 400 ml DM3 medium under normoxia (70% oxygen saturation) at 28°C and pH 6.8. After 6.5 h of growth, a subinhibitory concentration (150 μM) of 2,2′-dipyridyl was added to the culture, which was then maintained for 3 h under this condition. Aliquots of cells were withdrawn for RNA isolation in 1-h steps, including time zero (before 2,2′-dipyridyl addition). For microarray hybridizations, we used RNA samples from time zero (reference) against RNA from cells incubated for 1 h with 2,2′-dipyridyl.

Geldanamycin experiments.

Zoospores (6 × 10⁵ per ml) were inoculated in 140 ml DM3 medium under normoxia (70% oxygen saturation) at 28°C and pH 6.8. After 5.5 h of growth, 2 μM (final concentration) of geldanamycin was added and the culture was kept for 1 h under normoxia. Then, the oxygen concentration was decreased to 1% saturation, and after 1 h, aliquots of the cell culture were withdrawn for RNA isolation. For microarray hybridizations, we used RNA from cells kept under hypoxia for 1 h with geldanamycin pretreatment against RNA from cells kept under hypoxia for 1 h without geldanamycin pretreatment.

RNA isolation.

Frozen cells were crushed to a very fine powder using a chilled mortar and pestle in liquid nitrogen. Total RNA was isolated using TRIzol (Invitrogen) and quantified by direct spectrophotometry, and its integrity was checked through agarose-2.2 M formaldehyde gel electrophoresis, followed by ethidium bromide staining and RNA visualization under UV light.

cDNA microarray hybridization and analysis.

Ten micrograms of total RNA was reversed transcribed and labeled using a Superscript Plus indirect cDNA labeling system and Alexa Fluor dyes 555 and 647 (Invitrogen). The microarray chips were designed as described previously (14, 41) and contained 3,773 distinct expressed sequence tag sequences obtained from cDNA libraries constructed with RNA from cells at different stages of the life cycle of the fungus (40) and cells exposed to heat shock or cadmium stress (14). Based on gene content data from other fungi, our arrays probably encompass about 45 to 50% of the complete set of B. emersonii genes. Slides were cohybridized with the fluorescence-labeled probes overnight at 42°C. They were then washed in 1× SSC (0.15 M NaCl plus 0.015 M sodium citrate) and 0.2% sodium dodecyl sulfate (10 min at 55°C), twice in 0.1× SSC and 0.2% sodium dodecyl sulfate (10 min at 55°C), and in 0.1× SSC (1 min at room temperature). The slides were then rinsed briefly in Milli-Q water and dried under a nitrogen stream. Each experimental condition was analyzed using three independent biological experiments. Since each slide carried two replicates of the arrayed genes, a total of six intensity readings were generated for each gene in the microarray. Slides were scanned with a Generation III scanner (Molecular Dynamics), with the photomultiplier tube adjusted to 750 for both channels. The fluorescence mean intensity and surrounding median background from each spot were obtained with Array-Vision, version 6.0 (Imaging Research, Inc.). Data from clones that generated poor-quality spots (visually inspected) were excluded. Normalization was carried out by locally weighted regression (LOWESS) fitting on an M-versus-A plot, where M is the fluorescence log ratio of the test sample relative to the control condition [M = log₂(test/control)] and S is the log mean fluorescence intensity {S = log₂[1/2 (test) + 1/2 (control)]} (25).

Determination of differentially expressed genes.

We used intensity-dependent cutoff values for classifying a gene as differentially expressed based on the results of self-self hybridization experiments (49). In this type of hybridization, the same cDNA sample was labeled independently with both 555 and 647 dyes to estimate the experimental noise. In our case, we used in the self-self hybridization experiments at the 70% oxygen saturation point, which was the reference for all hybridizations. The HTself program, available on the web (http://blasto.iq.usp.br/∼rvencio/HTself/), was used to determine the intensity-dependent cutoff curves. These curves delimit the boundaries of intrinsic experimental noise and, therefore, genes which do not show statistically significant variation in their expression levels. Using these intensity-dependent cutoff values, we were able to determine which genes were differentially expressed during hypoxia. A gene was classified as differentially expressed at a given oxygen concentration point if more than 66% (at least two in three) of its replicates were outside the intensity-dependent cutoff curves.

Clustering analysis.

Differentially expressed genes in gradual-hypoxia experiments were clustered in 8 groups according to their expression patterns by using the K means algorithm implemented in the SpotWhatR software program (25). To characterize overrepresented functional gene categories among the differentially expressed genes, we measured the level of statistical association between “being in a given group” and “belonging to a functional category” by using the BayGO method (50). We considered that a gene category was overrepresented if the value of the statistical significance was smaller than 0.05.

Quantitative real-time RT-PCR.

To evaluate the reliability of the array-based data, seven genes were randomly selected and their expression levels in each experiment were analyzed by quantitative real-time reverse transcription-PCR (qRT-PCR). This analysis showed that the expression profiles of all chosen genes were qualitatively in agreement with microarray data, validating the array expression results quite well. Appropriate primers were designed by using Primer Express 2.0 software (Applied Biosystems). qRT-PCR experiments were performed using GeneAmp 5700 sequence detection system (Applied Biosystems) equipment and a SYBR green PCR master mix (Applied Biosystems). The thermocycling conditions comprised an initial step at 50°C for 2 min, followed by 95°C for 10 min and 40 cycles of 95°C for 15 s and 60°C for 1 min. The specificity of the amplified products was evaluated by analysis of the dissociation curves generated by the equipment. Two independent RNA samples were used for each gene analyzed. The gene encoding a putative heat shock protein 90 protein was used as the calibrator gene in all experiments. The determination of the expression ratios was carried out by using the threshold cycle (ΔΔC_T) method described by Livak and Schmittgen (30). This technique was also used to obtain the expression profiles of some B. emersonii genes that were not spotted on the microarrays.

Extracellular-glucose quantification.

During the gradual-hypoxia experiment, 1-ml aliquots of the culture were withdrawn from the reactor at different incubation times. Aliquots were centrifuged, and the supernatants were frozen at −20°C for future quantification of glucose concentration. The measurements were performed with an YSI 2700 Select (Yellow Springs Instruments Co.) analyzer, according to the manufacturer's specifications. This equipment functions with specific enzymes for the substrate of interest, immobilized between two membranes and electrochemical detection.

Microarray data accession number.

The microarray data discussed in this work have been deposited in NCBI's Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO series accession number GSE17524 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE17524).

RESULTS AND DISCUSSION

Effect of oxygen availability on B. emersonii gene expression.

To investigate the response of B. emersonii cells observed when the cells were exposed to low oxygen concentrations, fungal cells growing under normoxia conditions were subjected to gradual hypoxia, as depicted in Fig. 1. Determination of extracellular concentration of glucose was carried out using aliquots of these cultures, and data indicated that when cells were growing under normoxia, the uptake of glucose from the medium increased very slowly. However, once the cells were subjected to gradual hypoxia, the uptake of glucose increased considerably, increasing even faster during reoxygenation of the culture (Fig. 1).

Fig. 1.

Effect of oxygen availability on glucose uptake in gradual hypoxia. B. emersonii cells were incubated for 6.5 hours in DM3 liquid medium at 70% oxygen saturation (normoxia), and then (arrow) the level of oxygen supplied to the culture was reduced in 1-h steps to 35%, 17.5%, 5%, and 1% (in gradual hypoxia) and to 0% (anoxia). After 1 h of anoxia, in which pure nitrogen was purged for 1 h, the level of oxygen supplied to the culture was restored to the initial 70% value (reoxygenation). Aliquots of the culture (1 ml) were withdrawn from the reactor at different time points following incubation. After centrifugation of the samples, the resulting supernatants were frozen at −20°C for future quantification of glucose concentration (for details, see Materials and Methods). Glucose uptake was evaluated in three independent biological experiments. Aliquots of cells for microarray analysis and optical microscopy (see Fig. S1 in the supplemental material) were withdrawn at the time points indicated by arrowheads.

Morphologically, it was observed by optical microscopy that B. emersonii cells acquire an abnormal elongated form as the oxygen concentration decreased in gradual-hypoxia experiments (see Fig. S1 in the supplemental material). Despite this abnormal appearance, fungal cells seem to continue growing, as observed by their increase in size during hypoxia. After reoxygenation, cells begin to restore their normal morphology, as observed when they were incubated at 19°C under normoxic conditions for 12 h. Almost all cells recovered normal morphology and growth and also entered the sporulation phase (see Fig. S1, panels 8 and 9, in the supplemental material).

The transcriptional changes in B. emersonii cells exposed to low-oxygen conditions were then analyzed using cDNA microarrays. The nucleotide sequences of the cDNAs are available at http://blasto.iq.usp.br/ and were obtained by a high-throughput sequencing program of cDNA libraries constructed using RNA from B. emersonii cells at different stages of the life cycle of the fungus (40) and from cells exposed to heat shock or cadmium stress (14). In gradual-hypoxia experiments, the transcript population from cells isolated under the control condition of 70% oxygen saturation (arrow in Fig. 1) was compared to the transcripts expressed in the presence of various lower levels of dissolved oxygen, under anoxic conditions, and after reoxygenation (arrowheads in Fig. 1). Among the 3,773 genes analyzed, a total of 3,169 (84.0%) could be detected as expressed in the microarrays, and 366 (9.7%) were classified as differentially expressed in hypoxia, anoxia, and reoxygenation.

Since gradual hypoxia could lead to some kind of acclimation, we decided to perform a direct-hypoxia experiment (see Materials and Methods) in which dissolved oxygen in the fungal culture was directly decreased from normoxia (70% saturation) to hypoxia (1% saturation) and then cells were restored to normoxia after a 5-hour hypoxia. In this experiment, among the genes analyzed, a total of 3,379 (88%) could be detected as expressed in the microarrays and 456 (12.1%) were classified as differentially expressed in hypoxia and reoxygenation. Indeed, the direct-hypoxia experiment enriched the gradual-hypoxia data with 284 new differentially expressed genes, besides the 172 genes observed in the two types of experiments (Fig. 2A), with a total of 650 differentially expressed genes in both gradual- and direct-hypoxia experiments (see Table S1 in the supplemental material).

Fig. 2.

Transcriptome analysis of gradual and direct hypoxia. (A) Diagram showing the numbers of differentially expressed genes in gradual (left circle) and direct (right circle) hypoxia and under both conditions (overlap of the circles). (B and C) K means clustering of the expression profiles of all differentially expressed genes in gradual (B) and direct (C) hypoxia. Vertical grid lines in each cluster indicate oxygen concentration steps (from left to right: C70, C35, C5, C1, C0, and C70r in panel B and C70, C1_1h, C1_5h, and C70r_1h in panel C). Horizontal grid lines indicate values of log₂ expression ratios (fluorescence intensity of test cDNA/fluorescence intensity of reference cDNA).

Differentially expressed genes in gradual and direct hypoxia were clustered according to their expression profiles into 8 groups using the K means algorithm (Fig. 2B and C). For practical reasons, 534 genes were classified according to their upregulation or downregulation behavior under low oxygen concentrations, followed by recovery of the previous transcript levels after reoxygenation. A total of 116 genes without such a remarkable profile, which included genes that were probably affected not by oxygen deprivation but by other factors, such as nutritional changes in the medium or due cell growth during the experiment, were not considered. Figure 3 shows the upregulated and downregulated genes under gradual and/or direct hypoxia, distributed into functional categories, according to their best hits in GenBank. It can be observed that a large number (40%) of genes affected by oxygen deprivation are annotated as “no match,” i.e., with no significant similarity to any gene deposited in GenBank. This is certainly due to the high number (almost 45%) of “no match” genes spotted in the microarrays (14). Nevertheless, the participation of a large number of unknown genes and processes in the response to hypoxia in this fungus is intriguing.

Fig. 3.

Expression profiles of genes upregulated or downregulated in gradual and/or direct hypoxia, distributed into GO functional categories.

Each functional category shown in Fig. 3 will be discussed in detail later, but this transcriptional overview provides us evidence that B. emersonii responds to oxygen deprivation mainly by readjusting the transcript levels of genes required for energy production (energy/metabolism category) and genes involved in protective and structural mechanisms (redox homeostasis, cell wall, and lipid metabolism categories) and by downregulating genes involved in energy-consuming processes (ribosomal/translation, protein folding and modification, and transport). This overview was confirmed by statistical analysis of overrepresented gene categories among the upregulated and downregulated genes in gradual and direct hypoxia, performed by using the BayGO software program (50). As depicted in Table 1, gene categories related to energetic metabolism and protective and structural mechanisms, such as carbohydrate metabolism, glycolysis, carbon utilization, chitin catabolism, chitin binding, structural molecule activity, and pyridoxine biosynthesis, were overrepresented among upregulated genes, while energy-consuming categories related to protein folding, unfolded protein binding, amino acid metabolism, oligopeptide transport, protein targeting to mitochondrion, and ATP binding processes were overrepresented among the downregulated genes.

Table 1.

Overrepresented functional categories from the Gene Ontology Consortium (GO) among the upregulated and downregulated genes in gradual and/or direct hypoxia

Gene group and P valuea	GO no.	Overrepresented GO categoryb
Upregulated
0	GO:0005975	(b) carbohydrate metabolism
0.01	GO:0016868	(m) intramolecular transferase activity, phosphotransferases
0.01	GO:0015976	(b) carbon utilization
0.03	GO:0004733	(m) pyridoxamine-phosphate oxidase activity
0.03	GO:0004034	(m) aldose 1-epimerase activity
0.03	GO:0006807	(b) nitrogen compound metabolism
0.04	GO:0005198	(m) structural molecule activity
0.04	GO:0008061	(m) chitin binding
0.04	GO:0046872	(m) metal ion binding
0.04	GO:0000724	(b) double-strand break repair via homologous recombination
0.04	GO:0031966	(c) mitochondrial membrane
0.04	GO:0004301	(m) epoxide hydrolase activity
0.05	GO:0009116	(b) nucleoside metabolism
0.05	GO:0006096	(b) glycolysis
0.05	GO:0006032	(b) chitin catabolism
Downregulated
0	GO:0006457	(b) protein folding
0.01	GO:0051082	(m) unfolded protein binding
0.02	GO:0005739	(c) mitochondrion
0.02	GO:0006520	(b) amino acid metabolism
0.02	GO:0006626	(b) protein targeting to mitochondrion
0.03	GO:0005515	(m) protein binding
0.03	GO:0004300	(m) enoyl-CoA hydratase activity
0.04	GO:0006857	(b) oligopeptide transport
0.05	GO:0005524	(m) ATP binding
0.05	GO:0004070	(m) aspartate carbamoyltransferase activity

^aA functional category was considered overrepresented if the statistical association with its presence in the cluster was significant (P ≤ 0.05).

^bCategories are divided into three groups: molecular function (m), biological process (b), and cellular component (c).

In the following report, we will consider the differentially expressed genes in gradual and direct hypoxia, according to important selected functional classifications.

Amino acid metabolism.

The majority of genes encoding amino acid metabolism-related proteins showed a downregulated profile in response to hypoxia. Most of these genes encode enzymes related to amino acid biosynthetic pathways, like ornithine-oxo-acid transaminase (proline and arginine), 2-isopropylmalate synthase (leucine), cysteine synthase (cysteine), acetolactate synthase Ilv2 (valine, leucine, and isoleucine), aspartokinase (methionine, lisina, and threonine), histidinol phosphate aminotransferase (histidine), and 5-methyltetrahydropteroyltriglutamate-homocysteine methyltransferase (methionine). Another downregulated gene encodes the glutamate dehydrogenase 3 enzyme. This enzyme converts glutamate into α-ketoglutarate, feeding the TCA cycle. Its downregulation is concordant to the hypoxia condition under which the lack of oxygen impedes the normal TCA cycle function. Two genes were strongly downregulated in this category. They encode methyltetrahydropteroyltriglutamate-homocysteine methyltransferase and S-adenosylmethionine (SAM) synthetase, enzymes which are responsible for SAM biosynthesis. SAM is an important molecule in diverse biosynthetic pathways and acts by methylating nucleic acids, phospholipids, amines, and proteins. Such strong downregulation is probably related to the energy-saving effort necessary for cells under oxygen deprivation, as SAM-dependent processes are considerably energy-expensive.

Cell wall/structure.

It was observed that genes encoding enzymes responsible for chitin degradation (chitinase, chitinase Chi80 precursor, and chitin deacetylase) were upregulated under hypoxia, while the chitin synthase 1 gene was repressed. Genes encoding proteins related to cell structure, such as actin, myosin, tubulin, dynein, calponin, and transgelin, which are involved in morphological changes, growth, motility, secretion, cell division, and differentiation (10, 13), were also regulated by hypoxia. The regulation of genes related to cell wall metabolism and structural proteins is probably related to the morphological modifications observed in B. emersonii cells subjected to hypoxia.

Lipid metabolism.

The lipid metabolism category deserves special attention, as biosynthesis of fatty acids and ergosterol is oxygen dependent. Figure 3 shows that oxygen deprivation led to the induction of a number of genes related to lipid metabolism, as also reported for S. cerevisiae (3, 27, 45), Schizosaccharomyces pombe (47), Candida albicans (43), and Cryptococcus neoformans (9). One of the most strongly upregulated genes encodes a Δ-9 fatty acid desaturase, which corresponds to the OLE1 gene in S. cerevisiae and acts by adding a double bond between the C-9 and C-10 carbons of saturated fatty acids palmitoyl- and stearoyl-coenzyme A (CoA), forming palmitoleic and oleic monounsaturated fatty acids (34), which are essential for maintaining membrane fluidity. The proportion of unsaturated fatty acids incorporated into the cytoplasmic membrane is related to the tolerance of S. cerevisiae to stress situations, like those involving ethanol and heat shock (1, 6). Besides the Δ-9 fatty acid desaturase gene, the gene encoding a cyclopropane-fatty-acyl-phospholipid synthase (CFA synthase) was also strongly upregulated in hypoxia. This enzyme catalyzes the addition of a cyclopropane ring in phospholipid double bonds, favoring membrane fluidity and also protecting the phospholipids from reactive oxygen species attack (12).

The upregulation of genes encoding C-4 methyl sterol oxidase, C-5 sterol desaturase, and fatty acid-2 hydroxylase, enzymes related to ergosterol biosynthesis, which are targets of the mammalian SREBP (sterol response element binding protein) orthologues in C. neoformans (8) and S. pombe (47), was also observed. It has been shown that these orthologues have a role in the adaptation to hypoxia in these fungi (21, 47). However, no putative SREBP orthologue was found in the B. emersonii EST database.

Interestingly, only genes involved in lipid metabolism, among those present in the microarray chips, were observed to be downregulated under hypoxia in B. emersonii. These genes encode two enoyl coenzyme A hydratases, enzymes that act in fatty acid metabolism to produce acetyl-CoA and energy. This is in agreement with the hypoxia condition, under which acetyl-CoA molecules are not efficiently oxidized by the TCA cycle.

Central carbon metabolism.

When the central carbon metabolism is more closely examined (Fig. 4), it can be observed that B. emersonii cells subjected to hypoxia upregulate the expression of several genes encoding enzymes of the glycolytic pathway and the lactate dehydrogenase (strongly upregulated), differently to what is observed with TCA cycle enzyme genes. Interestingly, it was observed that the strongest upregulated genes were those encoding the phosphofructokinase subunits (PFK1 and PFK2), the most important regulation site in glycolysis. Therefore, it seems that at least at the transcriptional level, B. emersonii redirects its aerobic metabolism to anaerobic fermentation, suggesting an example of the “Pasteur effect.”

Fig. 4.

Expression pattern of B. emersonii genes encoding enzymes that participate in central carbon metabolism in cells under hypoxic and anoxic conditions. Red, green, and white boxes indicate those genes whose expression levels increased, decreased, and were unaffected, respectively, by the imposition of hypoxia and anoxia. Gray boxes represent genes that have not been isolated from B. emersonii. Genes marked with an asterisk were evaluated by qRT-PCR. The expression ratios of each gene are given in Table S1 in the supplemental material.

Another interesting gene strongly upregulated under hypoxia was that encoding a putative carbonic anhydrase. These enzymes are responsible for the rapid conversion of carbon dioxide into bicarbonate and protons, and among other physiological functions, they are related carbon dioxide and pH homeostasis. There are no reports about upregulation of carbonic anhydrase genes in other fungi, but in B. emersonii, this observation might be related to the control of a possible acidosis caused by lactic fermentation. In mammals, the carbonic anhydrase IX (CAIX) gene is one of the main targets of HIF-1α, and for this reason, this gene is strongly induced under hypoxia conditions. CAIX has been correlated with tumor invasion and progression, and as the enzyme is membrane bound, its participation in the control of intracellular acidosis and acidification of the extracellular tumor environment has been demonstrated (54).

The upregulation of the genes encoding four subunits of a V-ATPase or (H⁺)ATPase was also observed. These enzymes act as ATP-dependent proton pumps, which can acidify intracellular compartments and transport protons through membranes (22). In B. emersonii, the upregulation of these genes in hypoxia could also be related to the probable acidification caused by lactic acid fermentation.

Ribosomal/translation.

It is known that protein synthesis is among the most energy-consuming processes in cells and that the level of utilization of nutrients for energy production is lower under hypoxic conditions than in the presence of oxygen. Thus, one of the adaptations of organisms, especially those tolerant to lack of oxygen, is the energy-saving effort. In B. emersonii, it does not seem to be different, as ribosomal/translation stood out as one of the Gene Ontology (GO) categories with the largest number of downregulated genes in hypoxia (Fig. 3). The results for these downregulated genes can be seen in Table S1 in the supplemental material and are mainly related to ribosomal subunits, translation initiation factors, and aminoacyl-tRNA synthetases. Such downregulation of protein synthesis genes was also observed in human cells (11) and in fungal species such as C. neoformans (9).

Stress response.

In the stress response category, a strong upregulation of the gene encoding B. emersonii mitochondrial heat shock protein Hsp70-9 (15) was observed. For Drosophila melanogaster, it has been reported that constant hypoxia leads to induction of heat shock proteins Hsp70 and Hsp23, and this is correlated with a significant increase in survival under such conditions (2). In mammalian cells, the induction of Hsp70 is HIF-1 dependent, and it was shown that this chaperone has a role in protection against hypoxia injury in tumor, renal, neuronal, and even mesenchymal stem cells (4, 7, 20, 32, 35, 36, 39, 46, 51). Induction of Hsp70 under hypoxia was also reported to occur in C. albicans (43) and C. neoformans (9).

Transport.

Transport though membranes, as mentioned before, is one of the most energy-consuming processes. For this reason, downregulation of genes encoding transporters is completely plausible. Our data show that a large number of transport-related genes were downregulated in hypoxia (see Table S1 in the supplemental material), and among them, we emphasize a mitochondrial ADP/ATP translocase, which transports ADP in and ATP out of mitochondria. Two genes encoding ABC transporters, which are ATP dependent, and a gene for a high-affinity transporter for glutathione were also repressed. Among the upregulated genes, a hexose transporter HXT14, which is concordant with the metabolic shift to anaerobic fermentation mentioned before, was found. Genes encoding transporters of nicotinamide mononucleotide, methionine, and tricarboxylic acids were also found upregulated in hypoxia.

B. emersonii as a promising model of study for hypoxia.

If we compare B. emersonii's transcriptional response to hypoxia with data available for other fungi, it is possible to infer that this aquatic fungus is a more convenient model of study for this kind of stress. In C. neoformans (9) and Trichoderma reesei (5), for instance, despite remarkable similarities to B. emersonii, such as the downregulation of a number of genes related to protein synthesis and the upregulation of genes related to lipid metabolism, sterol synthesis, and stress response, the main difference found concerns the energetic metabolism. As stated above, B. emersonii cells subjected to hypoxia upregulate the expression of most genes encoding enzymes of the glycolytic pathway and the lactate dehydrogenase, differently from what is observed with the majority of TCA cycle enzyme genes, which were downregulated. In contrast, the most-important genes related to energetic metabolism did not seem to be significantly regulated during hypoxia in C. neoformans (9), and a drastic downregulation of most genes encoding glycolytic and TCA cycle enzymes was observed in T. reesei under hypoxia stress (5). Thus, the situation with these fungi is very distinct from what was found in B. emersonii, in which, at least transcriptionally, there is a metabolic shift to anaerobic fermentation. The same effect was demonstrated to occur in S. cerevisiae (28) and in certain mammal cells subjected to hypoxia (17, 53). S. cerevisiae is currently used as a model of study for hypoxia, but the fact that S. cerevisiae is a facultative aerobe limits its usefulness as a model system for mammalian cells. For this reason, we hypothesize that B. emersonii, an obligate aerobe, can be a more convenient model of study for mammalian cells.

Investigation of the oxygen-sensing mechanism in B. emersonii.

As an additional task, we have investigated the possible presence in B. emersonii of a mechanism of oxygen sensing similar to that found in metazoans involving the transcription factor HIF-1 (hypoxia-inducible factor 1), which is considered the main regulator of oxygen homeostasis in higher eukaryotes and mediates developmental and physiological pathways that either deliver O₂ to cells or allow cells to survive under O₂ deprivation. In mammals, the battery of HIF-1 target genes varies considerably among different cell types, but generally it promotes erythropoiesis and angiogenesis to increase oxygen delivery and anaerobic energy production through the upregulation of glycolytic enzymes and glucose transporter (GLUT1 and GLUT3) genes. In addition, the TCA cycle is inhibited by upregulation of lactate dehydrogenase A and pyruvate dehydrogenase kinase 1 genes. The latter enzyme leads to inhibition of pyruvate dehydrogenase, thus preventing entry of pyruvate into the TCA cycle and increasing the conversion of pyruvate to lactate (55). This regulation tends to redirect the metabolic flux away from the mitochondrial TCA cycle and oxidative phosphorylation, which requires molecular oxygen, allowing cells to rely mainly on the anaerobic process of glycolysis for energy production. This transcriptional behavior resembles what was observed in B. emersonii, and for this reason, we decided to investigate a possible HIF1-like mechanism in this fungus by comparing the transcriptional responses of the fungus to hypoxia, iron(II) deprivation, and geldanamycin treatment followed by hypoxia.

Fe²⁺ deprivation.

It has been shown that iron(II) chelators, like 2,2′-dipyridyl, have a hypoxia-mimicking effect in that they enhance HIF-1α accumulation (23). Due to this observation, we decided to analyze the expression profile of B. emersonii genes by subjecting the fungal cells to iron deprivation. For that, fungal cells were collected immediately before and 1 h after addition of 150 μM 2,2′-dipyridyl (see Materials and Methods). Among the genes analyzed, a total of 2,435 (64.5%) were found to be expressed and 219 (5.8%) were classified as differentially expressed by comparing cells treated and untreated with the iron chelator (see Table S2 in the supplemental material). Among the differentially expressed genes, 112 (51.1%) were also observed as differentially expressed in the oxygen deprivation experiments (see Table S2 in the supplemental material). Interestingly, the large majority (79%) of the expression profiles of these genes are in agreement in both experiments in regard to their up- or downregulation, which indicates that ferrous iron could play an important role in the hypoxic response in B. emersonii.

Geldanamycin treatment followed by hypoxia.

As mentioned in the Introduction, heat shock protein 90 (HSP90) has been found to have a role in protecting HIF-1α from O2/PHD/VHL-independent proteasomal degradation. Even though no putative HIF-1α homologue has been found in the cDNA libraries sequenced from B. emersonii, ESTs encoding two distinct HSP90 proteins in this fungus were recently characterized (38). In order to look for evidence of a HIF1-like mechanism in B. emersonii, we compared global transcript levels from cells treated or not treated with 2 μM geldanamycin (an inhibitor of HSP90), with both groups of cells subjected to 1 h of hypoxia (see Materials and Methods). Thus, with nontreated hypoxia cells taken as the reference condition, it was expected that genes affected by a putative HIF1α-like factor should be downregulated due to the absence of stable HIF1α even under hypoxic conditions, as HSP90 inhibition could be leading to a constant oxygen-independent degradation of this factor. Indeed, we found 111 downregulated genes in hypoxic cells treated with geldanamycin, compared with the level observed for nontreated cells, but only 12 of them were hypoxia-induced genes found in direct- and/or gradual-hypoxia experiments (see Table S3 in the supplemental material). The other genes were probably downregulated due to hypoxia-independent transcriptional changes caused by geldanamycin.

The genes affected by geldanamycin treatment are three energetic metabolism-related genes (encoding transaldolase, probable ATP-citrate synthase subunit 1, and d-lactate dehydrogenase), three lipid metabolism-related genes (encoding phosphatidylserine decarboxylase, Δ-9 fatty acid desaturase, and leukotriene A4 hydrolase), one structural gene encoding beta-tubulin, one tricarboxylate carrier, and four genes with no putative function assigned (see Table S3 in the supplemental material). Four of these genes, and a probable carbonic anhydrase gene (BeG30N15D05), were also evaluated by qRT-PCR, and data revealed reduction in the upregulated levels for these genes when cells were subjected to hypoxia after geldanamycin treatment (Fig. 5). Carbonic anhydrase, which was discarded from the geldanamycin microarrays for statistical reasons, was also evaluated because it seemed to be a good candidate for HIF1α regulation due to its strong upregulation under oxygen deprivation and because it had been shown that human carbonic anhydrase IX is regulated by HIF1α (24). The observation that these 13 hypoxic genes were affected by geldanamycin treatment is not sufficient to ensure that B. emersonii has a HIF1α-like mechanism but provides intriguing reasons to continue investigating this possibility.

Fig. 5.

Quantitative real-time RT-PCR data showing that levels of some hypoxic genes were decreased by treatment with geldanamycin (GA). Data were obtained from two independent biological experiments.

Final remarks.

The present work has provided important information on global gene expression changes observed during the response to hypoxia stress in the fungus Blastocladiella emersonii. Among the 650 genes differentially expressed under at least one of the conditions of stress, a total of 534 proved to be affected directly or indirectly by the availability of oxygen, once they showed a recovery (or tendency to recover) to normal expression levels when cells were reoxygenated. Data indicate that B. emersonii responds to hypoxia by rearranging the expression of genes responsible for production and consumption of energy. When exposed to hypoxia, this fungus favors anaerobic metabolism, as indicated by the upregulation of genes encoding glycolytic enzymes and lactate dehydrogenase, while most genes involved in the TCA cycle were downregulated or unchanged. This observation is corroborated by the increase in glucose uptake in B. emersonii hypoxia experiments. Furthermore, genes involved in energy-costly processes, such as protein synthesis, amino acid metabolism, protein folding, and transport, have a predominantly downregulated expression profile during oxygen deprivation, showing that the fungus is probably saving energy during hypoxia.

The similarities observed between the transcriptional profiles in hypoxia and Fe(II) deprivation experiments suggest that these stresses are somehow connected, providing evidence that ferrous iron could play an important role in the hypoxic response in B. emersonii. In addition, treatment of cells prior to hypoxia with geldanamycin led to a decrease in the levels of certain upregulated hypoxic genes. Altogether, these data suggest the presence of a hypoxia-sensing mechanism in B. emersonii similar to that involving the mammalian transcription factor HIF1-α, which is also affected by Fe²⁺ ion levels and geldanamycin.

Comparison of the transcriptional response of B. emersonii to hypoxia with the responses of other fungal species, such as T. reesei or C. neoformans, revealed that the main difference consists in the control of energy metabolism. Whereas B. emersonii cells seem to redirect aerobic metabolism to anaerobic fermentation, C. neoformans does not seem to regulate main energy metabolism genes during hypoxia and T. reesei suffers a drastic downregulation of most genes encoding enzymes of both the glycolytic pathway and the TCA cycle. The “Pasteur effect” transcriptionally observed in B. emersonii is more similar to the results found in S. cerevisiae and in mammalian cells subjected to hypoxia. As S. cerevisiae is a facultative anaerobe, we suggest that B.emersonii is a more convenient model of study of hypoxia stress.