(Received for publication, January 6, 1997, and in revised form, January 30, 1997)
From the Laboratory of Kidney and Electrolyte Metabolism, NHLBI, National Institutes of Health, Bethesda, Maryland 20892-0951
In yeast glycerol-3-phosphate dehydrogenase 1 is
essential for synthesis of the osmoprotectant glycerol and is
osmotically regulated via the high osmolarity glycerol (HOG1) kinase
pathway. Homologous protein kinases, p38, and stress-activated protein kinase/Jun N-terminal kinase (SAPK/JNK) are hyperosmotically activated in some mammalian cell lines and complement HOG1 in yeast. In the
present study we asked whether p38 or SAPK/JNK signal synthesis of the
osmoprotectant sorbitol in rabbit renal medullary cells (PAP-HT25),
analogous to the glycerol system in yeast. Sorbitol synthesis is
catalyzed by aldose reductase (AR). Hyperosmolality increases
AR transcription through an osmotic response element (ORE)
in the 5-flanking region of the AR gene, resulting in
elevated sorbitol. We tested if AR-ORE is targeted by p38 or SAPK/JNK
pathways in PAP-HT25 cells. Hyperosmolality (adding 150 mM
NaCl) strongly induces phosphorylation of p38 and of c-Jun, a specific
target of SAPK/JNK. Transient lipofection of a dominant negative mutant of SAPK kinase, SEK1-AL, into PAP-HT25 cells specifically inhibits hyperosmotically induced c-Jun phosphorylation. Transient lipofection of a dominant negative p38 kinase mutant, MKK3-AL, into PAP-HT25 cells
specifically suppresses hyperosmotic induction of p38 phosphorylation. We cotransfected either one of these mutants or their empty vector with
an AR-ORE luciferase reporter construct and compared the hyperosmotically induced increase in luciferase activity with that in
cells lipofected with only the AR-ORE luciferase construct. Hyperosmolality increased luciferase activity equally (5-7-fold) under
all conditions. We conclude that hyperosmolality induces p38 and
SAPK/JNK cascades in mammalian renal cells, analogous to inducing the
HOG1 cascade in yeast. However, activation of p38 or SAPK/JNK pathways
is not necessary for transcriptional regulation of AR
through the ORE. This finding stands in contrast to the requirement for
the HOG1 pathway for hyperosmotically induced activation of yeast
GPD1.
Homeostasis of intracellular inorganic ion concentrations and cell volume is essential for proper function of all cells. This is achieved by selective, energy-consuming electrolyte transport processes at the cell membrane (1) and by metabolic regulation of organic osmolyte levels in the cytoplasm (2). In situations of osmotic stress, cells adaptively osmoregulate by means of adjusting membrane water/ion fluxes and organic osmolyte levels. Such adaptive osmoregulation presumably is preceded by sensory and signaling events that monitor changes in osmotic strength and transduce this information to target genes and proteins of osmoprotective value.
Recent work employing yeast genetics provides a wealth of information
about the nature of osmosensing signal transduction in these primitive
eukaryotic cells. The yeast Saccharomyces cerevisiae senses
hyperosmolality through autophosphorylation of a sensor histidine
kinase, SLN1, that is homologous to two-component sensor kinases common
in bacteria (3). A second osmosensor protein, SHO1, has also been
identified in this species (4). Both osmosensors regulate the high
osmolarity glycerol (HOG1)1 response
pathway, which is a kinase cascade that induces a stress-response element (STRE) 5 upstream of a number of yeast genes (5). One target
of the HOG1 pathway is the glycerol-3-phosphate dehydrogenase gene
(GPD1) (6). GPD1 is essential for production of glycerol, which is a major organic osmolyte in yeast, and is accumulated in
response to hyperosmotic stress to restore cell volume and intracellular inorganic ion concentrations to norm values. The HOG1
pathway is necessary for this response.
At present, two mammalian protein kinases have been described that are functionally homologous to yeast HOG1. These are p38 (CSBP=RK=MXI2=MPK2) and SAPK/JNK. p38 appears to be the closest mammalian homologue of HOG1 based on functional complementation, overall sequence similarity, and a shared glycine residue in the conserved dual phosphorylation motif TGY (7). Both p38 and SAPK/JNK are osmotically regulated in one or another mammalian cell line, e.g. p38 in murine cell lines of monocytic origin (7) and SAPK/JNK in Chinese hamster ovary cells (8). In addition, other environmentally regulated kinases (ERKs) are activated upon osmotic stress in certain mammalian cell lines, e.g. ERK1 in rat 3Y1 fibroblasts (9). However, only p38 and SAPK/JNK functionally complement HOG1 in yeast. HOG1 deletion mutants are rescued from lethality in hyperosmotic medium when expressing either mammalian p38 (7) or SAPK/JNK (8). In mammalian cells, important targets of p38 include mitogen-activated protein kinase-activated protein kinases 2 and 3 that phosphorylate the small heat shock protein HSP27 (10). SAPK/JNK targets include transcription factors that are products of immediate early genes (e.g. c-jun) (11). However, unlike in yeast, no target genes or proteins of known osmoprotective value have thus far been shown to be osmotically regulated via p38, SAPK/JNK, or ERKs in mammalian cells.
Osmoregulated genes in mammalian cells include those encoding
transporters for compatible osmolytes (betaine, taurine, and inositol)
and the enzyme aldose reductase (AR), which catalyzes production of the
compatible osmolyte sorbitol from glucose (12). These four organic
osmolytes and glycerophosphocholine (but not glycerol) predominate in
renal medullary cells. Analogous to the GPD1/glycerol system in yeast,
AR transcription is up-regulated in rabbit renal medullary
(PAP-HT25) cells exposed to hyperosmolality, resulting in elevated
sorbitol (13). Hyperosmolality increases AR transcription
through an osmotic response element (ORE) in the 5-flanking region of
the AR gene (14).
In the present study we investigated whether mammalian p38 or SAPK/JNK pathways are activated by hyperosmolality in PAP-HT25 cells. In addition, we asked if increased synthesis of the osmoprotectant sorbitol via hyperosmotic induction through AR-ORE is signaled via p38 or SAPK/JNK cascades, analogous to the glycerol system in yeast. The results of this study are discussed in a comparative context and with emphasis on the role and specificity of osmosensitive kinase cascades as part of the general cellular defense machinery against osmotic stress.
Cell culture media were composed as described previously (15). The protease inhibitors Pefabloc® and leupeptin, as well as Tween 20, bovine serum albumin, and protein A-alkaline phosphatase were provided by Boehringer Mannheim. EDTA, EGTA, sodium pyrophosphate, sodium metavanadate, potassium ferricyanide, and potassium ferrocyanide were purchased from Aldrich. Microcystin LR was obtained from Calbiochem. Sodium fluoride, sodium azide, glycine, and thimerosal were from Sigma. Dithiothreitol, Tris base, and the protease inhibitors pepstatin A and aprotinin were purchased from ICN. SDS was from Digene, and the enhanced luciferase assay kit was from Analytical Luminescence Laboratory. Glycerol, methanol, and sodium chloride were obtained from J. T. Baker, and bromphenol blue, Blotto®, and prestained SDS-PAGE broad range standards were from Bio-Rad. PhosphoPlusTM p38 (P-Tyr182 #9211S) and PhosphoPlusTM c-Jun (P-Ser73 #9164S) phosphospecific antibodies and regular p38 (#9212) and c-Jun (#9162) antibodies as well as Phototope® Star ECL Western detection reagents were provided by New England Biolabs. LipofectamineTM, OPTI-MEM® reduced serum medium, and X-gal were purchased from Life Technologies, Inc. All other reagents were purchased from Mallinckrodt.
Cell Culture, Hyperosmotic Stimulation, and Extract PreparationPAP-HT25 cells between passages 63-76 were used for
all experiments. Hyperosmotic stimulation of cells was achieved by
substitution of hyperosmotic medium (600 mosmol/kg H2O) for
isosmotic medium (300 H2O). Hyperosmotic medium was
prepared by the addition of NaCl to regular growth medium. For assaying
p38 and c-jun phosphorylation, cells were kept up to 60 min
in hyperosmotic medium before extracts were prepared. For assaying
luciferase activity, cells were treated for 24 h under
hyperosmotic conditions before extract preparation. Controls in which
cells were treated with isosmotic instead of hyperosmotic medium were
always included. At the end of treatment, cells were lysed on ice with
cell lysis buffer (Analytical Luminescence Laboratory catalog number
1820 plus the following additives: 5 µM microcystin LR, 5 mM EDTA, 2 mM EGTA, 10 mM sodium
fluoride, 5 mM sodium pyrophosphate, 5 mM
sodium metavanadate, 4 mM Pefabloc® SC, 10 µg/ml
leupeptin, 10 µg/ml pepstatin A, and 1 µg/ml aprotinin). The
soluble components were separated from cellular debris by centrifugation at 15,000 × g and 4 °C for 15 min.
50-µl aliquots of supernatants were stored frozen at 80 °C for
later analysis of p38 and c-Jun phosphorylation or used immediately for
luciferase and protein assays.
SDS-PAGE was performed under standard denaturing conditions in a NOVEX MiniCell using 4-20% polyacrylamide gradient gels and Tris-glycine buffer system. Prestained, broad-range SDS-PAGE standards served as molecular weight markers. Gels were blotted onto polyvinylidene difluoride membrane (Immobilon-P, 0.45 µm, Millipore) by semi-dry transfer in a Pharmacia Multiphor II Novablot unit. We used a continuous buffer system (25 mM Tris base, 200 mM glycine, 20% v/v methanol) for 90 min at 100 mA constant current to achieve optimal transfer of proteins around 40 kDa. For immunodetection, membranes were blocked overnight at 4 °C and processed according to the instructions of the manufacturer of the p38 and c-Jun antibodies and the Phototope® Star ECL reagents (New England Biolabs). Protein A-alkaline phosphatase conjugate was used instead of secondary antibody. Blots were exposed to HyperfilmTM-ECL (Amersham) between 1 and 45 min, and the films were developed using a M35A X-Omat automatic film processor (Kodak). Films were scanned using a Molecular Dynamics densitometer (model PDSI-P90), and the volume of the bands (area × intensity) was quantified in the same area for all bands using ImageQuant software (Molecular Dynamics). The volume of the bands was automatically corrected for film background and expressed as densitometric volume units by the software.
Cell Transfection and Recombinant ProteinsExpression
vectors of dominant negative mutants SEK1-AL (Ser204 Ala and Ser207
Leu) and MKK3-AL (Ser231
Ala and Thr235
Leu) as well as pMT2 plasmid were
generous gifts from Dr. James Woodgett (Ontario Cancer Institute) (16).
pcDNA3 plasmid was purchased from Invitrogen. Luciferase reporter
construct ARL(
1170/
894) carrying a 277-base pair fragment that
contains the aldose reductase osmotic response element (AR-ORE) has
been described previously (14). PAP-HT25 cells were transfected using
LipofectamineTM and OPTI-MEM® medium generally according to the
manufacturer's (Life Technologies, Inc.) instructions. Specifically, 1 µg/ml of each plasmid was transfected at a LipofectamineTM
concentration of 6 µl/ml. During the initial 5 h, lipofected
cells were incubated in the presence of OPTI-MEM® serum-reduced medium
and were incubated thereafter in a 1:1 mixture of OPTI-MEM®
serum-reduced medium and PAP-HT25 growth medium containing 2 × serum. Transfection conditions (five per experiment) were run
simultaneously in six 60-mm dishes per each condition (30 dishes per
experiment). After 24 h the medium in all dishes was exchanged
with isosmotic PAP-HT25 growth medium. After another 24 h the
medium of three dishes per condition was exchanged with the same
isosmotic medium (controls), whereas the remaining three dishes were
subjected to hyperosmotic stress (PAP-HT25 growth medium + 150 mM NaCl). After a final 24-h incubation period under these
conditions, the medium was aspirated, and cells were washed with
phosphate-buffered saline of respective osmolality. Aliquots of cells
that were cotransfected with pcDNA3/His/lacZ plasmid were used for determination of transfection efficiencies by
staining with X-gal solution overnight and counting of blue versus unstained cells. The remaining cells were lysed by
the addition of 300 µl of cell lysis buffer/dish. Cell lysates were centrifuged (15,000 × g, 4 °C, 15 min), and the
supernatants were immediately used for assaying luciferase activity and
protein content.
Samples were assayed in duplicate for luciferase activity using an enhanced luciferase assay kit from Analytical Luminescence Laboratory according to the manufacturer's instructions. Luminescence was measured with a Monolight® 2010 luminometer in dual-ejection mode (Analytical Luminescence Laboratory). For determination of protein content a bicinchoninic acid protein assay kit from Pierce was used. Absorbance of protein samples was measured with a UV/Vis-1601 spectrophotometer (Shimadzu). Calibrations and protein assays were carried out according to the manufacturer's instructions. Luciferase activity is represented as relative light units/µg of protein.
Selection of Transiently Transfected CellsTo separate the
transiently transfected cells from nontransfected cells we utilized the
Capture-TecTM pHookTM-1 kit (Invitrogen). PAP-HT25 cells were
cotransfected with 1 µg/ml pHookTM-1 plasmid encoding a single-chain
antibody (sFv) against the hapten
4-ethoxymethylene-2-phenyl-2-oxazolin-5-one (phOx). This antibody is
expressed and displayed on the cell surface of transfected cells and
specifically binds to Capture-TecTM magnetic beads coated with the
hapten phOx. Following transfection, cells were processed according to
the instructions of the Capture-TecTM pHookTM-1 kit (Invitrogen).
Incubations in the presence of Capture-TecTM beads were done either at
isosmotic or at hyperosmotic conditions for 60 min at room temperature
on a LabQuake tube rotator (PGC Scientific). Identical incubations were
done in the absence of Capture-TecTM beads (unselected). Beads covered
with cells were separated using a magnetic stand (Invitrogen). Both
unselected and selected cells were pelleted, lysed in 25 µl of cell
lysis buffer, and stored at 80 °C until used for protein assays
and Western blots. Selection efficiency was determined in separate experiments by cotransfection of pHookTM-1 and
pcDNA3/His/lacZ plasmids followed by X-gal
staining of selected versus unselected cells.
Analysis of the data was done using ANOVA and Student-Newman-Keuls analysis to test for significant differences between group means. The significance threshold was set at p < 0.05. All statistical analysis was carried out using StatMostTM 32 software (DataMostTM Corp.).
Time series experiments during which PAP-HT25
cells were exposed to hyperosmotic medium (600 mosmol/kg
H2O) or isosmotic control conditions (300 mosmol/kg
H2O) were conducted to test for induction of p38 and
SAPK/JNK pathways in mammalian kidney cells. Four identical aliquots
from each cell lysate were processed in parallel by SDS-PAGE and
Western blotting. They were developed using PhosphoPlusTM p38 P-Tyr182 antibody (Fig. 1A),
nonphosphospecific p38 antibody (Fig. 1B), PhosphoPlusTM
c-Jun P-Ser73 antibody (Fig. 1C), or
nonphosphospecific c-Jun antibody (Fig. 1D). Western blots
of cell lysates probed with PhosphoPlusTM p38 P-Tyr182
antibody demonstrated a strong increase in p38 phosphorylation after
hyperosmotic shock (Fig. 1A). This increase in p38
phosphorylation was detectable as early as after 1 min of hyperosmotic
shock, was maximal after 20 min of hyperosmotic shock, and was not
induced under isosmotic control conditions (Fig. 1A).
Increased phosphorylation of c-Jun, a specific target of SAPK/JNK, was
also detectable as early as 1 min after hyperosmotic shock and not
apparent in isosmotic control samples (Fig. 1C). c-Jun
phosphorylation reached a maximum level slightly later (after 40 min)
than p38 phosphorylation (Fig. 1C). Absolute levels of p38
(Fig. 1B) and c-jun (Fig. 1D) did not
change within 60 min of hyperosmotic treatment.
Transfection and Selection Efficiencies Are Unaffected by Hyperosmolality
Lipofection of PAP-HT25 cells using
LipofectamineTM resulted in very high transfection efficiencies. The
efficiency of transfection using this method did not depend on medium
osmolality, at least not for incubations up to 24 h, which were
the longest incubation periods in hyperosmotic medium used in this
study (Fig. 2A). Likewise, transfection
efficiencies did not depend on medium osmolality when cells were
simultaneously transfected with two different kinds of plasmids and
either selected using Capture-TecTM bead technology or just treated
comparably without adding Capture-TecTM beads (unselected, Fig.
2B). Very close to 100% of transfected cells were
successfully selected in both isosmotic and hyperosmotic medium (Fig.
2B).
Dominant Negative MKK3 or SEK1 Mutants Prevent Hyperosmotic Activation of p38 or SAPK/JNK, Respectively
The effectiveness and
specificity of dominant negative MKK3-AL and SEK1-AL mutants was
checked by cotransfecting PAP-HT25 cells with a mutant together with
pHookTM-1 plasmid (whose expression can be used to select transfected
cells) and incubating cells for 60 min in hyperosmotic medium before
lysis. Part of transfected cells were selected during this 60-min
incubation period using the Capture-TecTM system, and the remainder
was incubated in the absence of Capture-TecTM magnetic beads
(unselected cells). Four samples from each cell lysate were processed
in parallel by SDS-PAGE and Western blotting and developed using
PhosphoPlusTM p38 P-Tyr182 antibody (Fig.
3A), nonphosphospecific p38 antibody (Fig.
3B), PhosphoPlusTM c-Jun P-Ser73 antibody (Fig.
3C), or nonphosphospecific c-Jun antibody (Fig. 3D). Phosphorylation of p38 and c-Jun is significantly
increased by hyperosmolality (p < 0.001) both in
nontransfected cells (Fig. 3, A and C,
lanes b, and Table I) and in cells
transfected with pHookTM-1 only (Fig. 3, A and
C, lanes 1, and Table I). However, in cells
cotransfected with MKK3-AL and pHookTM-1 hyperosmotic p38
phosphorylation is significantly inhibited (p < 0.001)(Fig. 3A, lane 3, and Table I). Under these
conditions, p38 phosphorylation is significantly partially inhibited
even without selection of transfected cells (p < 0.05)(Fig. 3A, lane 2, and Table I). The degree
of inhibition by MKK3-AL in nonselected cells (40.4-46.1%, Table I)
is comparable with the transfection efficiency (45.9-50.1%, Fig. 2).
p38 phosphorylation is not significantly different between isosmotic
controls and hyperosmotically shocked, selected cells that contain
MKK3-AL (Table I). In contrast, MKK3-AL has no effect on hyperosmotic
phosphorylation of c-Jun (Fig. 3C, lanes 2 and 3, and Table I). Conversely, cotransfection of cells with
SEK1-AL and pHookTM-1 prevents hyperosmotic phosphorylation of c-Jun
(Fig. 3C, lanes 4 and 5). In
nonselected cells this inhibition is significant (p < 0.05) and with 45.7-51.6% (Table I) similar to the transfection efficiency (45.9-50.1%, Fig. 2). For selected cells c-Jun
phosphorylation is not significantly different between isosmotic
controls and hyperosmotically shocked cells containing SEK1-AL (Table
I). On the other hand, SEK1-AL does not prevent hyperosmotic
phosphorylation of p38 (Fig. 3A, lanes 4 and
5, and Table I). Absolute levels of p38 and c-Jun did not
differ under any of the conditions tested (Fig. 3, B and
D, and Table I). These data demonstrate that MKK3-AL and
SEK1-AL are very efficient dominant negative mutants in transiently transfected PAP-HT25 cells. Moreover, MKK3-AL specifically blocks hyperosmotic activation of the p38 pathway, whereas SEK1-AL
specifically blocks hyperosmotic activation of the SAPK/JNK
pathway.
|
To examine whether hyperosmotic
induction through the AR-ORE is signaled via p38 or SAPK/JNK
pathways, we transfected PAP-HT25 cells with AR-ORE
luciferase reporter construct alone or together with either MKK3-AL,
SEK1-AL, pcDNA3 (empty MKK3-AL vector), or pMT2 (empty SEK1-AL
vector). Neither MKK3-AL nor SEK1-AL prevented the hyperosmotic
induction through AR-ORE. The increase of luciferase activity was similar under all conditions, ranging approximately between 5- and 7-fold over their respective isosmotic controls (Fig.
4). The experiment shown in Fig. 4 was repeated four
times with independent batches of cells yielding comparable results. The lack of an effect of MKK3-AL and SEK1-AL on hyperosmotic
AR-ORE induction indicates that in mammalian kidney cells
AR-ORE is not targeted by the p38 or by the SAPK/JNK kinase
cascade even though both of these pathways are strongly activated after
hyperosmotic shock.
We show that two mammalian phosphorylation cascades, the p38 and SAPK/JNK pathways, are hyperosmotically activated in mammalian renal medullary cells. Even though this has been shown using p38 and SAPK/JNK expression vectors for other mammalian cells (7, 8), this is the first confirmation of hyperosmotic activation of intrinsic p38 and SAPK/JNK in renal medullary cells. An important unique feature of renal medullary cells is that they are routinely exposed to large osmotic fluctuations in vivo. Hyperosmotic activation of p38 and SAPK/JNK pathways is potentially of great physiological significance for these cells. Yeast studies have demonstrated that p38 and SAPK/JNK, but not other MAPKs, complement mutants that lack the HOG1 gene and rescue mutants from lethality in hyperosmotic medium (7, 8). In addition, p38 and SAPK/JNK are structurally very similar to HOG1 and take the same positions within the homologous mammalian kinase cascades. p38 displays overall amino acid identity of 52% to HOG1 (7). Amino acid identity between SAPK/JNK and HOG1 is 41% (8). The elements of MAPK cascades have been conserved throughout evolution from yeast to mammals and include MAPK (yeast HOG1, mammalian p38, and SAPK/JNK), MAPK kinase (yeast PBS2, mammalian SEK1/MKK4, and MKK3), and MAPK kinase kinase (yeast SSK2/SSK22 and mammalian MEKK1). The closest mammalian homologue of HOG1 is p38, judging by their high sequence similarity and identical dual phosphorylation motif (Thr-Gly-Tyr) (7). In other ERKs amino acids other than Gly separate the two phosphorylation sites. Thus, hyperosmolality activates functionally and structurally homologous protein kinase cascades in yeast and mammalian cells. This finding raises the question whether functionally similar elements upstream and downstream of these pathways are also conserved between yeast and mammals.
Transcriptional Regulation of Osmoprotective Genes in Yeast and Mammalian CellsIn the yeast S. cerevisiae, one target
downstream of the HOG1 pathway is the gene encoding GPD1, which is
necessary for survival of yeast at high osmolality (17). Hyperosmotic
activation of GPD1 depends on the HOG1 pathway and leads to
increased transcription and translation of GPD1 (6). This results in
enhanced synthesis of glycerol-3-phosphate by
NAD+-dependent, GPD1-catalyzed reduction of
dihydroxyacetone phosphate. This reaction is of specific adaptive value
because it is the semifinal and rate-limiting step in the production of
glycerol, which is a major organic osmolyte in yeast. Osmotic stress
has been shown to be mediated via the HOG1 pathway to positively
regulate a STRE in the 5-flanking region of several genes (18) notably also TPS2, which encodes trehalose phosphate phosphatase, an
enzyme that is important for synthesis of the compatible osmolyte
trehalose (5). Among the different MAPK pathways described in yeast
only the HOG1 pathway affects transcription via STRE, and the proteins MSN2 and MSN4 are thought to be the positive transcription factors that
induce STRE in situations of hyperosmotic stress (19).
In the mammalian renal inner medullary cell line, PAP-HT25,
transcription and translation of aldose reductase, an enzyme catalyzing synthesis of the osmoprotectant sorbitol from glucose, is increased by
hyperosmotic stress (20). Regulation of cellular sorbitol levels via
its synthesis is comparable with regulation of yeast glycerol. Glycerol
is not an important osmolyte in mammalian cells, but sorbitol is one of
five major osmolytes in kidney cells, the others being glycine-betaine,
myoinositol, taurine, and glycerophosphocholine. In contrast to
sorbitol, cellular levels of glycine-betaine, myoinositol, and taurine
are regulated via active transport across the plasma membrane, and
glycerophosphocholine concentration is regulated by degradation (12).
In analogy to the positively regulated osmosensitive STRE in yeast, a
positive ORE has been identified in the 5-flanking region of the
mammalian AR gene (14).
In contrast to the extensive knowledge in yeast, little is known about signaling events that mediate hyperosmotic activation of AR-ORE. Because of the similarity between yeast glycerol and mammalian sorbitol systems, it seemed likely that AR-ORE might be targeted by the p38 pathway, but our results are not consistent with this possibility. The second mammalian protein kinase capable of functionally complementing yeast HOG1 deletion mutants, SAPK/JNK, apparently is also not involved in signaling hyperosmotic stress to AR-ORE. Therefore, hyperosmotic activation of AR transcription apparently is regulated via a different pathway than hyperosmotic activation of yeast GPD1. In light of the strong homology between mammalian p38, SAPK/JNK, and yeast HOG1 and the analogy in hyperosmotic regulation and osmoprotective function of the yeast glycerol and the mammalian sorbitol systems, this finding seems somewhat surprising. However, no strictly comparable targets of yeast HOG1 and mammalian p38 or SAPK/JNK pathways have been identified thus far. Similarly, no homologous activators of yeast HOG1 and mammalian p38 or SAPK/JNK pathways upstream of the respective MAPK kinase kinase are reported. The yeast HOG1 pathway is directly regulated by a two-component osmosensor phosphorelay system consisting of the osmosensor SLN1 (21), the phosphorelay protein YPD1 (22), and the response regulator SSK1 (3, 23). Despite considerable efforts, no animal homologues of two-component systems have been reported. A second yeast transmembrane osmosensor, SHO1, activating the HOG1 pathway by direct interaction with PBS2 has no known mammalian homologues (4).
In contrast to the specific osmosensor pathway in yeast, activation of mammalian p38 and SAPK/JNK pathways apparently is signaled via a less specific mechanism. Recently, it was found that hyperosmolality and UV-B irradiation lead to SAPK/JNK activation that is mediated by unspecific clustering of different kinds of growth factor receptors and cytokine receptors in the absence of ligands (24). The biophysical principles underlying receptor clustering have not yet been uncovered, but it seems possible that rearrangements in membrane lipid order and conformational changes in membrane proteins may be involved. Physical perturbation of the plasma membrane may also induce cytoskeletal rearrangements. These have been shown to activate small RHO-like GTPases RAC and CDC42, which in turn are potent activators of PAK, a MAPK kinase kinase kinase activating p38 and SAPK/JNK pathways (25, 26). Based on our current knowledge it seems that despite high evolutionary conservation of yeast HOG1 and mammalian p38 and SAPK/JNK pathways, elements upstream and downstream of these kinase cascades differ considerably between yeast and mammals.
Physiological Significance of Hyperosmotic Induction of p38 and SAPK/JNK CascadesBecause p38 and SAPK/JNK pathways are hyperosmotically activated in mammalian cells but do not induce AR-ORE, they could have other targets of adaptive value after hyperosmotic shock. These targets may be of general importance for protection against a variety of cellular stresses and not be limited to osmoprotection. Consistent with this idea, p38 and SAPK/JNK pathways are also strongly induced by other environmental stresses and by anti-inflammatory cytokines (11). Heat shock proteins and molecular chaperones are induced by a variety of cellular stresses, including hyperosmolality, and are of particular importance for protecting and refolding proteins during stressful conditions (27). Hyperosmotic activation of HSP25/27 is conferred via phosphorylation in a p38-dependent fashion (28). Heat shock proteins are also osmotically activated via increased transcription (12). Heat shock factor 1 is activated and binds heat shock element in 3T3 cells upon hyperosmotic shock (29).
Another important part of the machinery that protects against stress is the proteolytic system that degrades and removes defective cellular proteins. Internalization and proteolysis of growth factor and cytokine receptors is preceded by clustering of these proteins at the cell surface (24). Perhaps the resultant activation of the SAPK/JNK pathway helps induce the cellular proteolytic system. Also, SAPK/JNK and p38 pathways induce apoptosis of some mammalian cells (30, 31) and cell cycle delay and cellular repair in others (11). It will be interesting to see whether targets of p38 and SAPK/JNK pathways mediate these responses and to compare the response with those in yeast, considering that the yeast HOG1 pathway induces DNA damage-responsive gene (DDR2) and a gene encoding cytoplasmic catalase T (CTT1) via STRE (5).
Possible Mechanisms for Signaling Osmotic Regulation of Gene Expression through the OREOur results are not consistent with the theory that the ORE is targeted by p38 or SAPK/JNK pathways. In principle, AR transcription could be regulated by p38 or SAPK/JNK via a different DNA element. If so, then this would presumably be in cooperation with the ORE, because the ORE is the only element known independently to confer hyperosmotic AR induction (14).
It is not obvious why a signaling pathway proceeding from specific cell surface receptors to the nucleus should be employed by cells for osmotic regulation of genes and proteins. In contrast to growth factors, cytokines, or other small signaling molecules, osmotic stress may change the volume or ionic concentration of all cell compartments. Thus, no specific cell surface receptors may be required, which would eliminate the need for more distant signaling steps. DNA, RNA, and protein structure and interactions of these molecules are directly influenced by osmotic strength (32-35). The question remains, however, as to how such direct effects could signal specific osmoregulatory responses. In conclusion, although homologous MAPK pathways are activated by hyperosmolality in yeast and mammalian cells, there seems to be considerable divergence upstream and downstream of these cascades.
We thank Dr. James Woodgett (Ontario Cancer Institute) for providing us with the dominant negative mutants MKK3-AL and SEK1-AL.