Cellular responses to oxidative and osmotic stress

‘Cellular responses to oxidative and osmotic stress; sensing, signaling and gene expression’ was the theme of a novel congress held in Egmond aan Zee, The Netherlands, from August 28 to September 1, 1999. The program, attended by 150 participants, included 14 invited lectures by expert scientists studying these types of stress in organisms as diverse as bacteria, yeast, plants, and mammals, as well as 16 selected talks and 90 poster presentations.

Oxidative and osmotic stress differ significantly while also displaying overlapping responses. In general, oxidative stress is caused by the intracellular accumulation of reactive oxygen species (ROS) (P. Moradas Ferreira, Porto) or a disturbance of the cellular redox state (J. Teixeira de Mattos, Amsterdam). Hence the oxidative defenses encompass both nonenzymatic (glutathione, thioredoxin) and enzymatic (superoxide dismutase, peroxidases, catalase) detoxification mechanisms which destroy ROS or restore the redox balance. Oxidative stress signals may come from the environment, but can also be generated internally and may cause molecular damage to proteins (P. Moradas Ferreira, Porto), DNA, membranes, etc. Osmotic stress leads to efflux or influx of water from or into the cell: hyperosmotic stress causes shrinking, hypoosmotic stress causes swelling. The cellular responses to this type of stress deal with the activity of water channels (aquaporins) and electrolyte transporters, and the accumulation of osmolytes (I. Booth, Aberdeen; S. Hohmann, Gothenburg) as well as the protection of proteins and subcellular structures.

At first glance, therefore, sensing of oxidative stress in principle may include more direct molecular events than sensing of osmotic stress. Focussing on gene expression, stress-induced alterations in transcription may be the immediate consequence of oxidative effects on transcription factors. Such direct sensing of oxidative changes does indeed occur. For instance, D. Touati (Paris) discussed the soxRS regulon in Escherichia coli, which responds to superoxide. Activation of SoxR is controlled by the oxidative state of the iron-sulphur cluster contained in this transcription factor. SoxR activates transcription of SoxS, which, in turn, activates transcription of a large set of genes implicated in the oxidative stress response. Streptomyces contain a specific σ-factor, σ^R, which is active only after oxidative stress (M. Paget, Norwich). The activity of σ^R is regulated by the anti-sigma factor RsrA, which exerts its inhibitory action in vitro only in the absence of thiol reductants like dithiothreitol. Sigma^R and RsrA thus form a redox-sensitive switch. An example from mammalian cells is iκB–NFκB, which also displays intramolecular sensing of redox changes (N. Hunt, Sydney).

Primary changes in the pattern of transcription can be assessed by application of the genome-wide hybridization technique. This method has been extensively used for yeast genes as demonstrated in presentations by J. Labarre (Toulouse) and M. Toledano (Gif-sur-Yvette) on the oxidative stress response in general and the Yap1p and Skn7p regulons, respectively. For example, yeast purine genes belong to a set of genes, that are rapidly downregulated upon oxidative stress. The respective transcription factor Bas1p (homologue of cMyb) was presumed to be a direct target of oxidation because of its in vitro susceptibility to oxidation (E. Brendeford, Oslo). However, in vivo evidence is lacking; Bas2p, a molecular partner of Bas1p, may be a target. In the budding yeast Saccharomyces cerevisiae, the activation of transcription factors Yap1p (AP1-homologue) and Skn7p after an oxidative challenge has not been reported. Skn7p is a response regulator protein; however, this feature is not important for the oxidative stress response but may be implicated in the osmotic stress response. In the fission yeast Schizosaccharomyces pombe, activation of the Pap1p (AP1-homologue) is mediated by the mitogen-activated protein (MAP) kinase pathway (Wsc1—Wis1-Sty1) as explained by N. Jones (London). Remarkably, the same signal transduction route is involved in the activation of the osmostress-induced transcription factor Atf1p. Indeed Sty1 is the homologue of Hog1p in S cerevisiae as described below. Brp1, homologue of Skn7p from budding yeast, is another transcription factor controlling oxidative stress-dependent gene expression in S pombe (N. Jones, London).

ROS are actively produced during the response of plant cells against invading pathogens (R. Mittler, Jerusalem). Under these conditions, ROS-scavenging systems are particularly suppressed in order to amplify the so-called hypersensitive response aimed at evoking defense mechanisms like programmed cell death. Suppression occurs at different levels of gene expression. Enhancing the levels of antioxidant enzymes like SOD, ascorbate peroxidase, and catalase in transgenic plants (tobacco, maize) slightly improves stress resistance (D. Inzé, Ghent). By developing transgenic plants deficient in catalase as an experimental system, evidence has been obtained that hydrogen peroxide is able to activate local defense mechanisms against pathogens (D. Inzé, Ghent). ROS also play an important part in the mammalian (inflammatory and immune) response to infection or injury (N. Hunt, Sydney). When, for instance, phagocytes engulf a microorganism, nicotinamide adenine-dinucleotide phosphate (NADPH) oxidase is triggered. The superoxide radicals formed cause destruction of the microorganism. Antioxidant defenses are simultaneously elicited to protect the host.

As mentioned, the production of ROS may damage proteins, DNA, and lipids. Lipid oxidation may generate defense signals. For intance, oxidized lipids like 4 hydroxy-2,3 nonenal (HNE) or oxysterols mediate transcriptional induction of cytokine TGF-β1 mRNA in a human cell line, probably by promoting AP1 binding to the promoter (G. Poli, Turino). These events can be down-regulated by the use of antioxidants, thus preventing fibrotic degeneration of connective tissue.

D. Thiele (Michigan) addressed the important link between oxidative stress responses and copper ion-homeostasis. Copper ions are essential cofactors for enzymes implicated in the oxidative stress response while, on the other hand, excess of Cu ions lead to hydroxyl radical formation. Therefore, Cu ion levels are sensed by metalloregulatory transcription factors to properly regulate the genes involved in its transport and detoxification.

Sensing of osmotic stress is still a mystery, but it is obvious that these types of environmental changes have indirect effects as a consequence of the water influx or efflux. This may, among other factors, lead to detectable changes at the level of the plasma membrane. I. Booth (Aberdeen) explained the components of the osmoregulatory machineries in bacteria. The ‘powerplayers' in these organisms are the aquaporin AqpZ and mechanosensitive channel McsL. B. Poolman (Groningen), on the basis of in vitro reconstitution experiments with Lactobacilli, proposed a model in which regulation of transport systems is a major mechanism in osmostasis. Indeed, accumulation of compatible solutes (preferably glycine betaine) occurs mainly by uptake from the environment. In budding yeast, a hyperosmotic challenge leads to accumulation of the osmolyte glycerol. As a result of the osmoshock, the glycerol channel Fps1p immediately closes (S. Hohmann, Gothenburg). In addition, glycerol biosynthesis is enhanced.

Gowrishankar (Hyderabad) distinguished ion stress, osmostress, and ‘hydrotic’ stress, the latter also being evoked by permeable solutes like glycerol. In mammalian cells, the focus of osmostress response studies is mainly on cell volume regulation as discussed by F. Lang (Tübingen) and B. Tilly (Rotterdam). Among others, Na⁺/H⁺ exchange is regulated by hormones and mitogens (via activation of Ca⁺⁺-channels). Cell volume regulated kinases have been identified, such as the MAP kinases Erk-1 and Erk-2.

Two putative osmosensors, Sho1p and Sln1p, have been identified in budding yeast and are localized to the plasma membrane. The molecular mechanism by which these sensors are activated is unknown; it may be by membrane stretching or loss of turgor pressure. Sln1p is part of a 2-component sensing and signaling system, together with the phosphotransfer protein Ypd1p and the response regulator Ssk1p (F. Posas, Barcelona). Both osmosensors feed into the Hog1p MAP kinase module, at the level of the MAPKK Pbs2p. The Sho1p branch of activation is coupled to Pbs2p via the MAPKKK Ste11p, the Sln1p branch via the abundant MAPKKK's Ssk2p and Ssk22p (F. Posas, Barcelona). Osmostress-induced signal transduction received considerable attention during the meeting.

The budding yeast Hog1p MAP kinase pathway stands as a model for this signalling route. O. van Wuytswinkel (Amsterdam) presented evidence for an alternative mechanism of high osmolarity glycerol (HOG) pathway activation independent of the 2 known input branches mentioned above. Functional homologues of Hog1p are S pombe Sty1 and mammalian JNK and p38. Nucleo-cytoplasmic transport plays an important role in the regulation of MAP kinase activation and subsequent control of gene expression. Activation of Hog1p and Sty1 is down-regulated by the activity of nuclear-localized tyrosine phosphatases which are themselves activated by the MAP kinases. Apart from the transcription factors Atf1 and Pap1, which are activated upon osmotic and oxidative stress, respectively, additional target transcription factors are likely to play a part in fission yeast. Upstream components of the osmostress-activated MAP kinases are also evolutionarily conserved, as was demonstrated by J. Quinn (Newcastle upon Tyne): fission yeast contains a histidine kinase Mak2/3 that activates the response regulator Mcs4 via the phosphotransfer protein Ypd1.

The yeast-plant complementation approach has been rewarding for the isolation of plant homologues of yeast MAP kinase (or upstream) components, as explained by K. Shinozaki (Ibaraki), although functional identity could not be proven in all cases. There are many MAP kinase pathways in plants, which respond to extracellular signals; however, different environmental signals may trigger overlapping pathways, thus leading to an orchestrated response at the level of gene expression (H. Hirt, Vienna).

A remarkable resemblance exists between the components of the response mechanisms in yeast and plants. For instance an Arabidopsis homologue of the budding yeast Sln1p osmosensor, ATHK1, has been isolated (K. Shinozaki, Ibaraki). In addition, Arabidopsis transcription factors that bind to so-called drought responsive promoter elements (DREs) have been identified (K. Shinozaki, Ibaraki). The respective genes are induced upon exposure of cells to drought or high salinity and the transcription factors are probably under the control of specific signal transduction pathways. Overexpression of these genes in transgenic plants was found to improve salt tolerance. In addition, selection for osmotolerant budding yeast strains has led to the identification of plant genes which, upon overexpression, improved plant osmotolerance. N. Verbruggen (Ghent) discussed as an example the Dbf2p protein kinase gene, of which the functional role remains poorly understood. These strategies are promising for the engineering of osmotolerant crops.

S. Hohmann (Gothenburg) showed the transcriptional induction of 285 genes after osmostress exposure, part of which overlaps with the set of genes induced by oxidative stress. Molecular targets of Hog1p in S.cerevisiae have recently been discovered: Hot1p, a putative transcription factor controlling expression of glycerol biosynthesis genes (S. Hohmann, Gothenburg) and Sko1p, a transcriptional repressor negatively regulating transcription of salt-induced genes like the one encoding the Na⁺ pump ENA1, under nonstress conditions (M. Proft, Valencia). In kidney medulla cells, osmotic response elements in the promoters of key genes like those encoding aldose reductase (involved in the biosynthesis of the osmolyte sorbitol) and the transporters of betaine and inositol (other osmolytes) have been identified (M. Burg, Bethesda). These elements represent binding sites for the transcription factor TonEFB, which is upregulated upon increases in osmolality. Acute elevation of NaCl or urea concentrations lead renal medullary cells to arrest growth and induce apoptosis. During these changes many signal transduction events occur, including increased synthesis of the protein factors GADD45 and p53 (M. Burg, Bethesda).

This congress was the first meeting of scientists studying oxidative or osmotic stress in different organisms. The level of information exchange was very high and it was decided that a second international congress on cellular responses to oxidative and osmotic stress will be held in 2001 in Porto (Portugal).