Activation of extracellular signal-regulated kinases during dehydration in the African clawed frog, Xenopus laevis

The African clawed frog, Xenopus laevis, is native to southern Africa where it lives in swamps, lakes and rivers as well as man-made bodies of water (Tinsley and Kobel,1996). The frog is largely aquatic and rarely ventures onto land. However, seasonal changes in water levels can make it hard for the frogs to survive through the dry season and they have options including overland migration to new ponds or digging into the mud of evaporating ponds and entering a period of aerobic torpor known as estivation(Tinsley and Kobel, 1996). Both of these can place the frogs under desiccation stress. Xenopussp. have been widely studied since the 1950s as laboratory model animals, most extensively as a model of embryonic development(Beck and Slack, 2001). However, less is known about the responses of this species to environmental stresses, although some studies have looked at nitrogen metabolism and changes in the activities of urea cycle enzymes during dehydration and estivation(Balinsky et al., 1961; Balinsky et al., 1967). The molecular signaling mechanisms that allow adult Xenopus to deal with environmental dehydration stress have not been explored. The present study investigated one of the signal transduction cascades that mediate external stress signals to coordinate metabolic and gene expression responses by cells. This is the extracellular signal-regulated kinase (ERK) cascade, one family of the mitogen-activated protein kinase (MAPK) superfamily.

The MAPK superfamily encompasses major conserved signaling cascades that lead to the control of diverse physiological processes(Roux and Blenis, 2004; Krishna and Narang, 2008). MAPKs have been classified into six distinct pathways in mammals based on their preferential response to selected signals(Krishna and Narang, 2008). The three most studied groups of vertebrate MAPKs are: (1) the p38 MAPKs,which were originally described as a 38 kDa protein that is phosphorylated under endotoxin treatment or osmotic shock; (2) the stress-activated protein kinases (SAPKs), also known as JNKs, because they were first discovered though their role in the phosphorylation of the c-Jun transcription factor, and (3)the ERKs (Cowan and Storey,2003; Krishna and Narang,2008). All MAPK cascades begin at the cell membrane where an external signal causes the activation of a receptor kinase, typically a receptor protein tyrosine kinase (RPTK). This then triggers a three-tier regulatory cascade of kinases that provide a huge amplification of signal through the serial activation of the MAPK kinase kinase (MAPKKK), MAPK kinase(MAPKK) and finally the MAPK.

The regulation of the three main MAPK members in response to different stresses has been well characterized in mammalian cell lines but until recently the role of this superfamily in regulating adaptive responses in stress-tolerant animals has received little attention(Cowan and Storey, 2003). Recent work in our lab has examined the responses of the three MAPK families to whole body freezing in wood frogs, anoxia exposure in turtles and snails,and hibernation in ground squirrels(Greenway and Storey, 2000a; Greenway and Storey, 2000b; Cowan and Storey, 2003; MacDonald and Storey, 2005; Larade and Storey, 2006). In wood frogs (Rana sylvatica), for example, natural freezing survival was associated with the activation of p38 and JNK in different tissues,whereas the ERK pathway was unaffected(Greenway and Storey, 2000a). Studies have also explored the activation of MAPK family members in another frog species (Rana ridibunda) by diverse stimuli, including oxidative stress (Aggeli et al., 2001; Gaitanaki et al., 2003). In the present study we hypothesized that the ERK2 pathway would be activated in response to dehydration stress leading to subsequent activation of downstream ERK targets. Among the targets of ERK are various transcription factors and protein kinases that could play roles in cellular responses to dehydration,including enhanced expression of genes encoding protective proteins as well as protein kinase-mediated readjustment and/or suppression of energy-expensive cellular processes under stress.

This study examines the effect of dehydration on the expression of the ERK2 pathway in X. laevis including the upstream regulators of ERK2 and some of its downstream targets. The data show a strong coordinated activation of the entire cascade, providing good evidence of a significant role for ERK signaling in regulating both metabolic and gene expression responses to dehydration in the frogs.

Blood was sampled as follows. The aorta was snipped and two 50 μl aliquots of blood were quickly withdrawn into heparinized micropipettes. These were centrifuged for 1 min and then total sample length and packed red blood cell length in the tubes were measured and used to calculate the hematocrit. Remaining blood was pipetted into 2 ml microcentrifuge tubes with an aliquot(∼0.1 vol.) of anticoagulant solution added [10 mmol l^–1EDTA in 0.9% (w/v) NaCl]. Samples were centrifuged for 5 min at 5°C and then plasma was removed and frozen at –80°C. Plasma glucose levels were measured using a commercial glucose analyzer (Freestyle, Alameda, CA,USA). Initial tests with known amounts of glucose added to aliquots of X. laevis plasma showed that there were no detection problems when using frog plasma with this commercial analyzer.

Plasma urea was measured by a colorimetric assay(Price and Harrison, 1987)with some modification. The working reagent was made fresh by combining 2 vol. stock reagent A [600 ml double-distilled water (DDW), 300 ml H₂SO₄, 100 ml 85% H₃PO₄, 0.1 g FeCl₃] with 1 vol. stock reagent B (500 ml DDW, 0.05 g thiosemicarbazide, 2.5 g diacetylmonoxime). Samples (1–50 μl plasma)and standards (from 1 mmol l^–1 urea stock) were added to polypropylene Falcon tubes and adjusted to 1 ml total volume with DWW. A 0.5 ml aliquot of the working reagent was then added, mixed and the tubes were boiled in a covered bath for 10 min. After cooling for a few minutes in a fume hood, 200 μl aliquots of samples were pipetted into wells of a microplate and absorbance at 525 nm was measured using a microplate reader. Urea concentrations in plasma samples were determined by comparison with the standard curve.

Frozen tissue samples were homogenized 1:10 (w/v) in buffer containing 20 mmol l^–1 Hepes, 0.1 mmol l^–1 EDTA, 1 mmol l^–1 Na₃VO₄, 200 mmol l^–1 NaCl, 10 mmol l^–1 NaF, 10 mmol l^–1 β-glycerophosphate, and Sigma protease inhibitor mix. After centrifugation for 15 min at 24,000 g, the supernatant was removed. Soluble protein concentration was measured by the Coomassie Blue dye-binding method using Bio-Rad (Hercules, Ca, USA)-prepared reagent and bovine serum albumin as the standard. Samples were adjusted (if needed) to constant concentrations by the addition of small amounts of buffer and were then mixed 1:1 (v/v) with SDS-PAGE sample buffer [100 mmol l^–1 Tris-HCl, pH 6.8, 4% (w/v) SDS, 20% (v/v) glycerol, 0.2(w/v) Bromophenol Blue, 10% (v/v) 2-mercaptoethanol] and boiled for 5 min. Final soluble protein concentrations for the preparations from different tissues were 5 μg μl^–1 for heart and kidney, 4 μgμl^–1 for skeletal muscle, 2 μg μl^–1for lung and liver and 1 μg μl^–1 for skin.

Aliquots containing equal amounts of protein (10–30 μg depending on the protein being analyzed) were loaded into wells of 10 or 12% SDS polyacrylamide gels. Electrophoresis was carried out on a Bio-Rad mini-gel apparatus at room temperature using 1× running buffer (3.02 g Tris base,18.8 g glycine, 1 g SDS l^–1, pH ∼8.3). Proteins were then transferred onto PVDF membrane by wet transfer with transfer buffer containing 25 mmol l^–1 Tris pH 8.5, 192 mmol l^–1glycine, 20% (v/v) methanol. Membranes were washed in TBST [Tris-buffered saline containing 0.1% (v/v) Tween 20, 20 mmol l^–1 Tris base,140 mmol l^–1 NaCl] for 5 min and were then blocked with 2.5%non-fat dried milk for 10–30 min in TBST before washing twice with TBST. Membranes were then incubated with primary antibodies overnight at 4°C followed by washing for 3× 5 min in TBST and then incubation with secondary antibody (1:1000 or 1:2000 dilution) for 1–2 h at 22°C,and final washes of 3× 5 min with TBST. Immunoreactive bands were visualized by enhanced chemiluminescence (Millipore Corporation, Billerica,MA, USA). PVDF membranes were then stained for 20 min with Coomassie Blue[0.25% (w/v) Coomassie Brilliant Blue R, 50% (v/v) methanol, 7.5% (v/v) acetic acid] and destained for ∼7 min with destaining solution (50 ml distilled water, 50 ml acetic acid, 150 ml methanol). Antibodies were purchased from Cell Signaling Technology (Danvers, MA, USA) except for MEK kinase and MEK1/2,which were from Transduction Laboratories (Lexington, KY, USA); all were used according to manufacturer's instructions.

Both immunoblot and Coomassie-stained bands were visualized using a ChemiGenius Bio-imaging System (Syngene, Frederick, MD, USA) and densitometric analysis was performed using the associated Gene Tools software. A group of three Coomassie-Blue-stained bands that showed constant density in all samples from a given tissue were quantified and immunoblot band densities in each lane were normalized against the summed density of these bands in the same lane in order to correct for any minor variations in sample loading. Mean normalized band densities ± s.e.m. for control and dehydrated samples(N=3–6 independent samples) were calculated and then statistical testing was performed using analysis of variance followed by the Student–Newman–Keuls test.

Initial studies determined the limit of body water loss for X. laevis; frogs died at values just above 30%, the mean percentage of total body water loss for frogs that died being 32.4% ±0.98 s.e.m.(N=7). Based on this value, two groups of frogs were set up and monitored as they underwent experimental dehydration to medium or high levels. Frogs lost water at near-linear rates and the final mean percentages for the medium and high dehydration groups were 16.6±1.6% (N=9) and 28.0±1.6% (N=10), respectively, of total body water lost over 6–7 days (Fig. 1). Hence,the high dehydration group represents frogs whose water loss was close to the lethal limit whereas the medium dehydration group reached only about half of that level. Blood hematocrit values also attested to the loss of body water,the percentage of total blood volume occupied by packed red blood cells rising from 40% in control frogs to 47% in medium and 50% in high dehydration groups(Table 1).

The high cutaneous water loss rates of amphibians can be minimized by behavioral, physiological and biochemical adaptations(Hillman et al., 2009). One method to help retard water loss is to raise the osmolality of body fluids and, among anuran species that estivate or endure hypersalinity, the commonly occurring osmolyte is urea. Indeed, previous studies have shown that under osmotic or water stress, X. laevis reduce ammonia excretion, elevate the activities of urea cycle enzymes and accumulate urea(Balinsky et al., 1961; Janssens, 1964; Seiter et al., 1978). In the present study, plasma urea concentration increased sharply when frogs were dehydrated, rising by 27- and 33-fold, respectively, in the medium and high dehydration groups as compared with control values of 1.66 mmol l^–1 (Table 1). As a consequence of dehydration, amphibians also experience hypoxia as a result of elevated blood viscosity and reduced blood volume that make it harder to fully oxygenate tissues (Hillman et al., 2009). Hypoxia stress typically triggers the mobilization of liver glycogen reserves and elevates plasma glucose in frog species, and in the present study this also occurred. Plasma glucose levels in the high dehydration group were 2.3-fold higher than in controls(Table 1), similar to the effects of dehydration on blood glucose levels in leopard frogs, Rana pipiens (Churchill and Storey,1993). This is consistent with the probable development of hypoxic conditions during dehydration in X. laevis.

The components of the ERK signaling cascade have been well established(Roux and Blenis, 2004; May and Hill, 2008; Ramos, 2008). The classical initiation site is a RPTK in the plasma membrane. Activation of the receptor passes a signal from its phosphotyrosine-docking protein and adaptor growth factor receptor bound proteins (Grb2) to a guanine nucleotide exchange factor called SOS (son of sevenless). SOS facilitates the activation of a small G protein called Ras by catalyzing the exchange of GDP to GTP. Active, GTP-bound Ras then binds to and activates one of the MAPKKKs belonging to the ERK cascade; these include Rafs (A-Raf, B-Raf, and c-Raf), Mos and MEKK. These then activate the MAPKKs (MAPK/ERK kinase or MEK1/2) and finally the MAPKs(the ERK family, ERK1 and ERK2 being the best studied)(Fig. 2). The activated ERK is then capable of phosphorylating cytosolic substrates or transcription factors.

In the present study, immunoblotting was used to measure the levels of phosphorylated c-Raf1 because it has an equal ability to phosphorylate MEK1 and MEK2 (Wu et al., 1996). Different and tissue-specific responses were seen in organs of the medium and high dehydration groups of X. laevis(Fig. 3A). Medium dehydration did not significantly affect phospho-c-Raf^Ser338 levels in any of the five tissues tested but high dehydration did. Phospho-c-Raf^Ser338 levels rose in skeletal muscle and lung by 2.4-to 2.5-fold (P<0.01) compared with control levels in both organs. Similarly, levels of the phosphoprotein rose by 94% in skin(P<0.05) as compared with control levels. Phospho-c-Raf^Ser338 was not detectable in the liver.

Another MAPKKK in the ERK pathway is MEKK(Roux and Blenis, 2004) and analysis of MEKK showed that it also responded to dehydration in X. laevis tissues. MEKK total protein levels were elevated significantly in muscle and liver of the medium dehydration group of frogs by 55% and 4-fold(P<0.01), respectively (Fig. 3B). Levels remained elevated in liver of the high dehydration group of frogs. MEKK protein also responded to high dehydration in lung and kidney, rising by 2- and 2.3-fold. However, in heart and skin, MEKK levels were reduced significantly in dehydrated frogs and MEKK levels also returned to control values in skeletal muscle of the high dehydration group. The activation of c-Raf and MEKK in the present study is also in line with previous findings (Matsuda et al.,1995) on the effect of osmotic shock on these proteins in rat 3Y1 fibroblastic cells. Our data support the proposal that c-Raf and MEKK both play important roles in the response of X. laevis lung and skeletal muscle to dehydration. MEKK is also important in liver and kidney whereas c-Raf responded positively to dehydration in skin.

MEKK, c-Raf, and other activated MAPKKKs phosphorylate the dual specificity kinases MEK1 and 2. These are highly selective for their substrates, which are primarily ERK1 and ERK2 (English et al.,1999). They phosphorylate ERK1/2 at both tyrosine and serine/threonine residues (MacDonald and Storey, 2004), and MEK1/2 have two other unique features that allow them to efficiently activate ERK1/2. The first is the binding site of ERK2, which is an eight residue stretch in MEK1/2 known as the D domain. The second feature is the presence of a proline-rich insert in the C-terminal portion of the catalytic core of MEK1/2, which is important for intracellular stimulation of ERK1/2 (English et al.,1999).

Since upstream regulators of MEK1/2 were activated in response to dehydration in some X. laevis organs, we next measured total protein levels and activation status of MEK1/2. In skin and liver, total MEK1/2 protein decreased by ∼50 and ∼60% (P<0.01), respectively,under both medium and high dehydration conditions(Fig. 4A). MEK1/2 levels were also reduced in heart of the medium dehydration group of frogs by 33%(P<0.01). However, in lung, total MEK1/2 protein content was strongly increased by 2- and 4-fold (P<0.01) under medium and high dehydration, and also increased 2-fold (P<0.01) in the kidney under high dehydration.

It is the active phosphorylated form of MEK that is key to triggering ERK and so the relative levels of phospho-MEK1/2^Ser217/221 are important. Fig. 4B shows that dehydration enhanced phosphorylation of MEK in all tissues, compared with control levels. Under medium dehydration conditions, the amount of phospho-MEK1/2^Ser217/221 increased by 2-fold in lung, 2.3-fold in skin and by 74% in liver (P<0.01). High dehydration further enhanced phospho-MEK1/2^Ser217/221 content in lung and skin to 6-fold and 3-fold above the phosphoprotein content of controls(P<0.01). Phospho-MEK1/2^Ser217/221 levels were also elevated by 1.8-fold in muscle, 2-fold in kidney, and 2.5-fold in heart(P<0.01) under high dehydration conditions.

A strong correlation between c-Raf activation(Fig. 3A) and the phosphorylation of its downstream target MEK1/2(Fig. 4) occurred under high dehydration in muscle, lung and skin. Similarly, there was a good correlation between MEKK activation (Fig. 3B) and elevated levels of phosphorylated MEK1/2 in kidney, lung and liver. However, the elevated phospho-MEK1/2^Ser217/221 levels in heart under high dehydration conditions could not be related to either c-Raf or MEKK so it is possible that other Raf isoforms or Mos could be involved in the pathway in heart (Roux and Blenis,2004; Sridhar et al.,2005).

Activation of MEK1/2, the MAPKK of the ERK pathway, leads to phosphorylation of ERK at a conserved Thr-Glu-Tyr (TEY) motif of the activation loop (MacDonald and Storey,2004). The vertebrate ERK family includes six isozymes (ERK1, 2,3, 5, 7, 8) but ERK1 and ERK2 have received the most study. They share 83%amino acid identity and are structurally and functionally similar(Roux and Blenis, 2004). In general, activation of the ERK pathway mediates cell growth and differentiation in response to a number of different stimuli including growth factors, cytokines, transforming agents, virus infection and carcinogens(May and Hill, 2008; Ramos, 2008). The interaction of ERK2 (the isozyme analyzed in the present study) with its substrates was better understood after using X-ray crystallography. The activation of ERK2 causes the protein to undergo conformational changes, creating a proline-specific surface pocket. When inactive, this pocket is usually occupied by Arg192, so substrate access to the proline catalytic site is inhibited when ERK2 is not phosphorylated on Thr183 and Tyr185. In addition to the proline-specific pockets, ERK2 has the ability to regulate its substrates through direct binding of the C-terminal substrate binding groove of ERK to the activation site of substrates, thereby restricting their interaction with other kinases (Schaeffer and Weber,1999).

The ERK2 antibody crossreacted with a single band in X. laevistissues at the expected molecular mass for ERK2 (∼42 kDa). Under dehydration stress the total protein levels of ERK2 remained stable in muscle and lung of X. laevis (Fig. 5A) but ERK2 levels were suppressed by 30–40%(P<0.05) in heart and skin of both the medium and high dehydration group of frogs. By contrast, total ERK2 protein increased by 35 and 50%(P<0.01) in kidney and liver, respectively, under medium dehydration. The active, phosphorylated form of ERK2 showed very different responses to dehydration, with a significant increase in the amount of active phospho-ERK^{Thr202/Tyr204} detected in all tissues except skin(Fig. 5B), mirroring the activation of MEK1/2, the upstream MAPKK(Fig. 4). In particular,phospho-ERK^{Thr202/Tyr204} content was very strongly enhanced in both muscle and kidney, by 6- and 4.6-fold (P<0.01), respectively,under high dehydration conditions as compared with control levels of the phosphoprotein. Phospho-ERK^{Thr202/Tyr204} levels also increased in lung by 2.8- to 3.3-fold (P<0.01) in dehydrated animals, and heart showed a similar approximate 2-fold increase. This result is consistent with previous work on R. ridibunda heart in which ERK was found to be powerfully activated by stress stimuli(Gaitanaki et al., 2003). A 2.5-fold increase (P<0.01) in the amount of phosphorylated active ERK2 was also seen in liver during medium dehydration.

Overall, then, the above data for the MAPKKK, MAPKK and MAPK components of the ERK cascade are all consistent in documenting an activation of this cascade in response to physiological dehydration stress in most tissues of X. laevis. The exception was skin, which showed elevated levels of phosphorylated c-Raf and MEK1/2 but not of ERK2. However, it is possible that ERK1 might be the target that is activated in skin. Activation of the ERK cascade may be responding directly to dehydration stress or may be responding to hypoxic conditions caused by dehydration effects on the circulatory system. Studies on other species have reported ERK activation by ischemia and/or reperfusion in rabbits (Iliodromitis et al., 2006) or by osmotic stress in animals and yeast(Kultz and Burg, 1998).

The activation of the ERK cascade in X. laevis tissues indicates that activation of some downstream targets of ERK is important for dehydration resistance. ERK substrates are present in all cellular compartments(Roux and Blenis, 2004), but the best-known actions of ERKs are in the regulation of gene expression. ERKs can translocate to the nucleus and influence gene transcription by phosphorylating transcription factors or transcription factor modifying proteins (Krishna and Narang,2008; Ramos,2008). Therefore, we next evaluated the responses of some known ERK targets in X. laevis tissues.

The signal transducer and activators of transcription (STAT) proteins are a family of transcription factors that form homo- or heterodimers upon phosphorylation before translocating from the cytoplasm into the nucleus. In mammals there are seven family members (STAT 1, 2, 3, 4, 5a, 5b, 6). Phosphorylation of these transcription factors occurs on a conserved tyrosine residue in the C-terminal domain (Benekli et al., 2003; Ning et al.,2001). STAT3 is a known substrate of ERK1/2(Chung et al., 1997; Krishna and Narang, 2008) and so we analyzed the phosphorylation status of phospho-STAT3^Ser727 as a measure of ERK downstream action. Levels of the phosphoprotein increased significantly under high dehydration conditions in three tissues of X. laevis (Fig. 6A) and in all cases this correlated with elevated phospho-ERK2 in the same tissues(Fig. 5B). In skeletal muscle,phospho-STAT3^Ser727 showed small but significant increases of 24–30% (P<0.05), whereas greater increases of 1.9- to 2.3-fold (P<0.01) occurred in lung. Phospho-STAT3^Ser727levels were also 2-fold higher in kidney of the high dehydration group of frogs (P<0.01). However, phospho-STAT3 levels were suppressed in heart of the medium dehydration group and were unchanged in the other tissues. Activated STAT3 in muscle, lung and kidney can upregulate a number of genes whose protein products assist in cell survival. For example, the anti-apoptotic Bcl2 family members Bcl-x_L and Mcl1 were upregulated in multiple myeloma following STAT3 induction. Other known STAT3 target genes include c-Myc, which plays a role in sensitizing cells to apoptotic stimuli(Benekli et al., 2003). The activation of STAT3 in the muscle and kidney could be a general response to dehydration stress since a significant change in the water contents of individual organs occurs (measured as a change in protein concentration) when frogs dehydrate (Churchill and Storey,1993; Holden and Storey,1997).

The p90 ribosomal S6 kinase (RSK) family of proteins is another known target of ERK action. The protein was originally isolated from Xenopus, and described as a protein kinase with the ability to phosphorylate the S6 protein of the 40S ribosomal subunit. Although S6 is still the best-known RSK target, recent studies have identified other targets including other proteins involved in ribosomal translation, the insulin receptor, the proapoptotic protein BAD and some nuclear targets(Ruvinsky and Meyuhas, 2006). In humans, there are four RSK isoforms; each contains two distinct and functional kinase domains. In the carboxy-terminal domain, RSK autophosphorylation and activation occur and then stimulate the phosphotransferase activity of the amino-terminal kinase domain to allow RSK to phosphorylate its targets. ERK activation of RSK1 requires the phosphorylation of its C-terminal activation loop(Anjum and Blenis, 2008). ERK achieves that by playing one of two roles in the process. Activated ERK can directly phosphorylate RSK1 and ERK can also bring RSK1 closer to membrane-associated kinases that may phosphorylate RSK on specific residues(Richards et al., 2001).

The present study focused on the regulation of one of the essential RSK proteins, RSK1. The RSK1 activation pattern quite closely mirrored the pattern seen for ERK^{Thr202/Tyr204}. High dehydration elevated the amount of p90^RSKSer380, with levels of the phosphoprotein increasing by 7-,5-, 3- and 2-fold (P<0.01) in muscle, lung, kidney and liver,respectively (Fig. 6B). By contrast, in heart, phosphoprotein content decreased by 47% and 69%(P<0.01) during medium and high dehydration, respectively, where the phospho-RSK1 content did not change in skin.

To confirm that RSK activation leads to phosphorylation of downstream targets in tissues of dehydrated frogs, we next measured S6 ribosomal protein(S6RP) levels and its phosphorylation state in X. laevis organs. Phosphorylation sites on S6RP were mapped to five serine residues in the highly conserved carboxyl terminus: Ser235, Ser236, Ser240, Ser244 and Ser 247(Ruvinsky and Meyuhas, 2006). The effects of dehydration on total S6RP protein levels in X. laevistissues are shown in Fig. 7A(note that selected samples were unavailable in three cases). S6 protein levels were 6.9- and 2.4-fold higher (P<0.01) in muscle and liver during high dehydration, as compared with controls. Dehydration also had a very marked effect on lung S6RP protein levels, with mean increases of 23.5-and 17.8-fold (P<0.05) under medium and high dehydration conditions, respectively. However, dehydration had no significant effect on S6RP protein content in kidney, heart or skin. The phosphorylated form of the protein was assessed using an antibody that detects serine 235/236 phosphorylation in the C-terminal region. A strong increase in phospho-S6RP content was detected in most tissues under high dehydration conditions(Fig. 7B); levels were 8.5-,8.5-, 4.7- and 6.8-fold higher in skeletal muscle, lung, kidney and liver,respectively (P<0.01), compared with the control tissues. Phospho-S6RP^Ser235/236 also increased in lung by 70%(P<0.05) under medium dehydration. A good correlation occurred between the increase in phospho-S6RP protein levels and the amount of activated phospho-RSK1 (Fig. 6B) in muscle, lung, kidney and liver of the high dehydration frogs. These findings also agree with previous studies that showed that RSKs are the kinases responsible for phosphorylating the S6RP subunit in X. laevis oocytes (Schwab et al.,1999).

The functional role of S6RP phosphorylation is still debatable. Initial studies proposed that S6RP has a regulatory role in translation initiation but subsequent data was conflicting as to whether a causal link existed between S6RP phosphorylation and the efficiency of protein synthesis(Ruvinsky and Meyuhas, 2006). Indeed, a recent study used S6RP knockin(S6RP^p–/–) mice to assess S6RP function with results that showed that the rate of global protein synthesis in mouse embryo fibroblasts was actually higher than in wild-type cells, and, furthermore, the liver of these mice had the same proportion of ribosomes engaged in polysomes as did the wild-type liver. Other actions for phosphorylated S6RP have been suggested including: (1) specifically increasing the translation efficiency of a subgroup of mRNA species that contain 5′-terminal oligopyrimidine tracts (TOP mRNAs), (2) mediating an inverse regulation of cell size versus cell proliferation, and (3) involvement in glucose homeostasis. Given the profound changes in S6RP levels and phosphorylation status in X. laevis organs, the effects of dehydration stress in this frog model might be an excellent one to examine in order to finally identify a physiological role for S6RP.

As hypothesized, the present study clearly shows that the ERK2 signal transduction cascade is activated in X. laevis tissues in response to dehydration. Activation was particularly pronounced under high dehydration conditions, whereas fewer and less pronounced effects were seen in response to medium dehydration. This suggests that responses regulated by ERKs may be defensive actions to deal with severe dehydration and some of its consequences such as hypoxia caused by elevated blood viscosity. Given that all anurans have wide tolerances to changes in body hydration and osmolality, it appears probable that the stress represented by medium dehydration could be dealt with largely without the intervention of ERK-mediated responses. Activation of the ERK cascade was particularly apparent in lung, which is an organ that is highly susceptible to water loss during breathing. Perhaps ERK signaling might trigger early protective measures by lung to help limit respiratory water loss. In this regard, it is interesting to note that the expression of mucin genes in the mammalian respiratory tract is under ERK1/2 control(Choi et al., 2009), which could suggest that ERK-mediated changes to the amount or composition of mucins in the airway epithelia of X. laevis could be involved in limiting respiratory water loss under desiccating conditions. This remains to be experimentally tested. All organs showed activation of some or all of the components of the ERK2 cascade (c-Raf, MEKK, MEK1/2, ERK1/2) in response to high dehydration, which shows that ERK signaling is important to cellular defense when water loss becomes severe. Two well-known downstream targets,STAT3 and RSK1, were also frequently phosphorylated and activated in X. laevis tissues in response to dehydration, as was the RSK target, the S6 ribosomal protein. This suggests that these targets may contribute to dehydration defense. Some organs (e.g. heart) showed clear activation of the ERK cascade but no change in the downstream targets that were assessed. However, there are a large number of downstream targets of ERK2, so this suggests that dehydration-stimulated ERK signaling is focused on alternative targets in heart. Skin showed a strong increase in c-Raf and phospho-MEK1/2 but not in other cascade components, which suggests that skin responses to dehydration are not directed by ERK2 but are probably accomplished by another ERK member, such as ERK1. Overall, the study provides strong evidence of a role for ERK signaling in dehydration defense in X. laevis organs,documents activation or upregulation of well-known ERK targets, and directs our attention to detailed studies of ERK-mediated responses to uncover all of the metabolic adaptations that are coordinated by ERK signaling in response to dehydration.