Báo cáo Y học: Atlantic salmon possess three mitogen activated protein kinase kinase 6 paralogs responding differently to stress pot

In cells exposed to sorbitol, sodium arsenite and UV radiation, the different salmon MKK6s were shown to be selectively activated.. Phylogenetic analysis of MKK6 and MKK3 sequences from

kinase kinase 6 paralogs responding differently to stress Tom E Hansen1,2, Pa˚l Puntervoll3, Ole Morten Seternes4 and Jorunn B Jørgensen1,2

1 Department of Marine Biotechnology, Norwegian College of Fishery Science, University of Tromsø, Norway

2 The Norwegian Structural Biology Centre (NorStruct), University of Tromsø, Norway

3 Computational Biology Unit, Bergen Centre for Computational Science, Norway

4 Department of Pharmacology, Institute of Medical Biology, University of Tromsø, Norway

The p38 group of mitogen-activated protein kinases

(MAPKs) is activated by pro-inflammatory cytokines

and environmental stress [1] We have recently cloned

three different Atlantic salmon (Salmo salar) p38

cDNA variants (As-p38a, b1 and b2) [2] All three

variants were phosphorylated after treatment of cells

with the stressors sodium arsenite and sorbitol

Addi-tionally, the pro-inflammatory stimulants bacterial

lipopolysaccharide (LPS), CpG oligonucleotides and

the cytokine interleukin (IL)-1 were shown to activate the p38 signalling pathway in salmon macrophages The inhibition of tumor necrosis factor (TNF)-2 and IL-1b expression in LPS stimulated salmon macro-phages by the p38 specific inhibitor SB203580, high-lights the importance of p38 in the regulation of cytokine expression also in fish

The activation of the MAP kinase pathway is achieved through a three component protein kinase

Keywords

MAP kinase; MKK3; MKK6; p38; salmon

Correspondence

J B Jørgensen, Department of Marine

Biotechnology, Norwegian College of

Fishery Science, University of Tromsø,

N-9037 Tromsø, Norway

Fax: +47 77 64 60 20

Tel: +47 77 64 67 16

E-mail: jorunn.jorgensen@nfh.uit.no

(Received 8 May 2008, revised 28 July

2008, accepted 4 August 2008)

doi:10.1111/j.1742-4658.2008.06628.x

Mitogen activated protein kinase kinase (MKK) 3 and 6 are the main p38 mitogen-activated protein kinase activators in mammals In the present study, three Atlantic salmon MKK6 orthologs were identified The deduced amino acid sequences of the salmon MKK6 proteins were highly similar to mammalian MKK6 sequences, and they were ubiquitously expressed All three were shown to be upstream activators of salmon p38 In cells exposed

to sorbitol, sodium arsenite and UV radiation, the different salmon MKK6s were shown to be selectively activated Thus, our results suggest a specific function of the three salmon MKK6s depending on which stress stimuli the cells are exposed to Phylogenetic analysis of MKK6 and MKK3 sequences from different species indicate that salmon is unique in having three MKK6 gene copies, whereas other fish species possess one or two MKK6 genes Interestingly, in contrast to mammals, fish do not have

an MKK3 gene We propose that two major duplication events have occurred for the ancestral MKK3⁄ 6 gene: one in tetrapods yielding MKK3 and MKK6, and another one in fish yielding two MKK6 paralogs The third MKK6 copy found in salmon is probably the result of the salmonid-specific tetraploidization event In conclusion, we report for the first time

in any species the existence of three MKK6 genes displaying distinct expres-sion and activation patterns Furthermore, MKK3 is dispensable in some vertebrates because it is absent from fish genomes despite being present in chicken and all mammals sequenced so far

Abbreviations

As, Atlantic salmon; CHSE, Chinook salmon embryo; eEF2, eukaryotic elongation factor 2; EST, expressed sequence tag; GFP, green fluorescent protein; GST, glutathione S-transferase; IL, interleukin; LPS, lipopolysaccharide; MAP3K, MAPK kinase kinase; MAPK, mitogen-activated protein kinase; MKK, MAP kinase kinase; TNF, tumor necrosis factor.

cascade consisting of the MAPK kinase kinase

(MAP3K), the MAPK kinase (MKK or MAP2K) and

the MAPK The MAPKs are activated upon

phosphor-ylation of Thr and Tyr in the activation loop by specific

MKKs [1,3,4], whereas the MKKs are activated by

phosphorylation of their Ser and Thr residues in the

activation loop by MAP3Ks [5] Once activated, the

MAPKs may translocate into the nucleus and

phos-phorylate specific target molecules on Ser or Thr

resi-dues Activated MAPKs phosphorylate a wide array of

targets localized both in the cytoplasm and the nucleus,

including transcription factors and other kinases that

facilitate the transcription of MAPK regulated genes [6]

The principal MKKs for p38 in mammalian cells are

MKK3 and MKK6 and, in some cases, MKK4 [7]

MKK substrate specificity is mediated by the

interac-tion motif located in the N-terminal part of the kinase

[8–12] Four p38 isoforms, a, b, c and d, are found in

mammalian species and a selective activation of each

of the isoforms by MKK3 [13,14] and MKK6 has been

reported [15–18] MKK6 and MKK3b activate all four

p38 members, whereas MKK3 activates all except

p38b [13–16,18–23] MKK4, which primarily activates

c-Jun N-terminal kinase, is shown to participate in the

activation of p38 under certain type of stress [7]

Alto-gether, this selective recognition by different MKKs

and MAPKs highlights some of the complexity of the

mammalian MAPK cascade

Mouse knockout experiments have been useful in

defining the physiological roles for MKK3 and

MKK6 Although mice lacking either MKK3 or

MKK6 are viable, the disruption of both will result in

death during early development [7,24–26] Single

dis-ruption of MKK3 has revealed an essential role for

this kinase in the regulation of TNF-a induced

cyto-kine expression in embryonic fibroblasts and IL-12

production in LPS-stimulated macrophages [24,25]

Targeted deletion of MKK6 in mice shows impaired

deletion of double positive thymocytes [26] Analysis

of fibroblast from mice lacking both MKK3 and

MKK6 demonstrates redundant but also essential roles

for MKK3 and MKK6 in mediating TNF-a stimulated p38 activation By contrast, MKK3 and MKK6 are not essential for UV-induced p38 activation [7] The MAPK signaling pathway is highly conserved through evolution and MKK homologues have been identified in both vertebrates [21,23,27–31] and inverte-brate species [32], and also in yeast [33] In fish, two MKKs from the MKK3⁄ 6 family have been cloned: one from carp (Cyprinus carpio) and one from zebrafish (Danio rerio) [27,29] In the present study, we have identified and characterized three Atlantic salmon MKK cDNAs: Atlantic salmon MKK6a (As-MKK6a), As-MKK6b and As-MKK6c

Results

Cloning of As-MKK6a, As-MKK6b and As-MKK6c With degenerated primers and RACE-PCR, we were able to amplify a cDNA encoding 336 amino acids showing the strongest amino acid similarity to human MKK6 (85% identity) The sequence was given the name As-MKK6a (GenBank accession number AY641477) A blast analysis in the NCBI database with the As-MKK6a sequence revealed the presence of two rainbow trout expressed sequence tag (EST) clones with high similarity to As-MKK6a Based on the sequences of the two ESTs, we were able to clone two other MKK cDNAs that encoded a 357 amino acid protein (As-MKK6b; GenBank accession number EU234532) and a 359 amino acid protein (As-MKK6c; GenBank accession number EU234533) The As-MKK6b and As-MKK6c showed 93% nucleotide sequence identity, respectively The nonmatching nucleotides were spread throughout the whole ORF, which suggests that As-MKK6b and As-MKK6c rep-resent two MKK6 isoforms The predicted sizes of the As-MKK6 proteins were 38 kDa (As-MKK6a) and

40 kDa (As-MKK6b and As-MKK6c) An alignment

of the As-MKK6 sequences and selected MKK6 sequences from other species is shown in Fig 1A Note

Fig 1 Alignment of MKK6 sequences and the phylogenetic tree of MKK3 and MKK6 sequences (A) The multiple sequence alignment of selected MKK6 sequences is shown, emphasizing any differences Identical amino acid residues are denoted by dots, different residues are shown in lowercase, and gaps are shown as hyphens The conserved phosphorylation site residues Ser and Thr lying within subdomain VIII are marked by arrows, and the conserved N-terminal p38 docking motif [(R ⁄ K) 2 -(X)2-6-L ⁄ I-X-L ⁄ I] is framed by a grey box (B) The phylogenetic tree was built from an extended alignment that included additional MKK6 sequences and selected MKK3 sequences Clade credibility values are indicated, and the insect MKK3 ⁄ 6 sequences were used as outgroup For clarity, non-salmon fish sequences occurring in the same clade

as As-MKK6a were named MKK6a and those clustering with As-MKK6b were named MKK6b The accession numbers for the sequences used are: MKK3 ⁄ 6: drosophila, Q9U983; mosquito, Q7PRZ7; ciona, Q4H382 MKK3: chicken, Q5ZL06; cow, A4IFH7; mouse, O09110; human, P46734 MKK6a: tetraodon, Q4SGJ8; fugu, SINFRUP00000137389; stickleback, ENSGACP00000014471; medaka, ENS-ORLP00000009468 MKK6b: tetraodon, Q4S8I5; fugu, SINFRUP00000128869; medaka, ENSORLP00000017352; carp, Q9I959; zebrafish, Q6IQW6 Sequences were retrieved from UniProt (six character long accession numbers) or ENSEMBL

B

that the salmon MKK6 sequences contain both the

N-terminal p38 MAPK docking motif and the

phos-phorylation sites lying within subdomain VIII

Phylogenetic analysis of MKK3 and MKK6

sequences

The initial blast searches performed with the

As-MKK sequences suggested the need for a thorough

phylogenetic analysis of the two closely related MKK3

and MKK6 sequence families from mammals and fish

for two reasons First, although the As-MKK6

sequences are more similar to human MKK6 (80–85%

identity) than MKK3 (73–74% identity), the most

clo-sely related zebrafish sequence (80–87% identity) was

first named MKK3 [27] Second, the initial blast

searches indicated that, in contrast to salmon,

zebra-fish and carp only appear to have one copy of the

MKK3⁄ 6 gene

A phylogenetic tree was constructed for the

MKK3⁄ 6 sequences as described in the Experimental

procedures (Fig 1B) Two equally striking

observa-tions can be made from the tree First, MKK3 does

not appear to be present in fish Second, the MKK6

gene appears to have undergone duplication in fish: all

six fish species analysed have at least one MKK6 gene;

green pufferfish, fugu and medaka have two copies;

and salmon is the only species with three copies

Hence, a second duplication appears to have occurred

in salmon, resulting in As-MKK6c and As-MKK6b

Tissue distribution of As-MKK6a, As-MKK6b and

As-MKK6

By northern blotting, a 4.0 kb transcript representing

As-MKK6a was detected in all the tissues tested with

the highest expression in the ovary Two smaller

tran-scripts of 1.7 kb and 1.4 kb were also found in the

ovary The 1.7 kb transcript was detected as a faint

band in the other organs tested (Fig 2A) By using a

probe encompassing the entire coding region of

As-MKK6b, we revealed only a single transcript of

approximately 1.7 kb showing equal expression levels

in all tissues examined (Fig 2B) Due to the high

sequence similarity between MKK6b and c, it is likely

that the MKK6b probe detects both these transcripts

on the northern blot To distinguish between them,

RT-PCR with primers specific for MKK6a, b and c

were designed and used for expression analysis The

RT-PCR was performed with mRNA from the same

tissues and mRNA from macrophages (Fig 2C) and,

consistent with the northern analysis, all three

As-MKK genes were shown to be expressed in these

tissues The expression of MKK6a was predominant in the liver compared to MKK6 b and c Analysis of MKK6 expression in head kidney macrophages revealed that only As-MKK6a and c were detected in these cells

As-MKK antibody specificity The specificity of the As-MKK6a, b and c antibodies was tested by western blot analysis of immunoprecipi-tated myc-tagged MKK6a, b and c constructs expressed in Chinook salmon embryo (CHSE)-214 cells The purified antiserum raised against the PPPHQSKGEMSQPKG peptide showed specificity to As-MKK6b⁄ c, but not to As-MKK6a (Fig 3A), and

A

B

C

Fig 2 Tissue expression analyses of salmon MKK6a, MKK6b and MKK6c (A) A blot containing poly(A)+ RNA isolated from various salmon tissues was hybridized with probes specific for either As-MKK6a (upper panel) or (B) As-MKK6 b ⁄ c (upper panel) b-actin expression was used as a loading control in (A) and (B) (lower pan-els) (C) Expression of the As-MKK6a, b and c in various tissues and in head kidney macrophages examined by RT-PCR, using MKK6a, b and c specific primers.

was named anti-MKK6b⁄ c By contrast, the purified

antiserum raised against the SQPKGGKRKPGLKLS

peptide recognized all three As-MKKs and was named

anti-pan-MKK6 These antisera were tested on lysate

from primary nonstimulated macrophages (Fig 3B,C)

A band at the predicted size of As-MKK6b⁄ c (40 kDa)

was detected using the As-MKK6b⁄ c antibody in these

cells (Fig 3B) This band most likely represents MKK6c

because RT-PCR analysis revealed that MKK6b was

not expressed in macrophages (Fig.2C) In addition, a

faint band at the same size as As-MKK6a was apparent

in the macrophages Whether this band represents

another salmon MKK6 variant or is due to unspecific

binding in not known The pan-MKK6 antibody

recog-nized two proteins with the predicted sizes of

As-MKK6b⁄ c and 6a (Fig 3C) These results show that

the peptide antibodies raised against the As-MKK6s,

despite some cross-reactivity, can be used to detect

salmon MKKs in tissues and cells

Phosphorylation of As-MKKs by UV irradiation The MKKs are phosphorylated and activated by MAP3Ks [5] The C-terminal part of MKKs contains

a stretch of approximately 20 amino acids immediately

on the C-terminal side of the MKK catalytic domain reported to be important in the docking of the MAP3Ks to the MKKs [34] Similar C-terminal dock-ing sites sequences are present in the three As-MKKs

To investigate whether the As-MKKs are phosphor-ylated at Ser and Thr residues within subdomain VIII

of the activation loop, a specific antibody to phosphor-ylated Ser189 of human MKK3 and Ser207 of human MKK6 was used CHSE-214 cells were transfected with myc-tagged MKK wild-type constructs and UV radiated for 30 min, followed by immunoprecipitation

of the tagged MKKs UV irradiation is a cellular stres-sor known to engage multiple signalling pathways end-ing in p38 activation [35] As shown in Fig 4, all three As-MKKs were phosphorylated in UV radiated CHSE-214 cells The amount of phosphorylated As-MKK6b was considerably lower (Fig 4B) than the amount of As-MKK6c (Fig 4C) and 6a (Fig 4A) Phosphorylation of As-MKK6b was not detected in the total lysate, whereas phosphorylated As-MKK6c

A

C

B

Fig 3 Antibody specificity against As-MKKs (A) CHSE-214 cells

were transfected with expression vectors encoding either

As-MKK6a wid-type (wt), 6b wt or 6c wt containing an N-terminal

myc epitope tag After 48 h, the cells were lysed and the

myc-tagged proteins were immunoprecipitated using myc antibody The

immunoblots of precipitated wt myc-MKKs were examined using

pan-MKK6 (recognizing As-MKK6a, b and c, upper panel),

anti-MKK6b ⁄ c (recognizing As-MKK6b and c, middle panel) and

anti-myc sera (lower panel) Primary salmon macrophages were

harvested and As-MKK6a, 6b and 6c were visualized by western

blotting of whole cell extract using anti-As-MKK6b ⁄ c (B) and

anti-pan-MKK6 sera (C) Similar results were obtained in two separate

experiments.

A

B

C

Fig 4 As-MKK6a, b and c are phosphorylated upon stress treat-ment CHSE-214 cells were transfected with myc-MKK6a wild-type (wt) (A), myc-MKK6b wt (B) and myc-MKK6c wt (C) At 48 h post transfection, the cells were treated with 120 mJÆcm)2of UV radia-tion (30 min) or left untreated Cell lysates were harvested and wt myc-MKKs were immunoprecipitated (IP) from the cell lysates The immunoblot of the whole cell extracts (WCE) and IP myc-MKKs wt were analysed by anti-p-MKK3 ⁄ 6 (p-MKK; upper and middle panel) and anti-myc (lower panel) sera.

and 6a were found in the same lysate, and the latter

two were even detected in the nonstimulated cells

These results indicate that As-MKK6b are poorly

phosphorylated in UV radiated CHSE-214 cells

com-pared to As-MKK6c and 6a

Ectopically expressed As-MKK6a, b and c are

differently activated by diverse types of stress

The p38 signalling cascade is activated by diverse

clas-ses of stress [1,28,36] To explore the activation of

salmon MKKs by different stressors, CHSE-214 cells

over-expressing myc-tagged MKK6a, b or c were

trea-ted with sodium arsenite or sorbitol for 30 min

Sodium arsenite is a oxidative stress inducer that

acti-vates p38 [37,38], whereas sorbitol induces osmotic

stress [39] A time course over 60 min with the same

type of treatments and also including UV radiation for

30 min was performed for As-MKK6a transfected

cells The kinase activity of immunoprecipitated

myc-tagged proteins was assayed using recombinant

His-p38a as MKK substrate High MKK6a activity was

detected for all three stress treatments (Fig 5A) The

time course of sodium arsenite treatment revealed no

activity before 15 min post treatment and the activity

remained at the same level to 60 min post treatment

In sorbitol treated cells, As-MKK6a activity was

detected at 5 min post treatment and the activity

remained from 15–60 min post treatment Sorbitol

treatment gave also strong As-MKK6b activation,

whereas a modest change in the As-MKK6b activity

was detected upon sodium arsenite treatment (Fig 5B)

Interestingly, the results for As-MKK6c were opposite,

whereas the addition of sodium arsenite to the cells

induced activation was sorbitol ineffective (Fig 5C)

The latter suggests that sodium arsenite is a poor

MKK6c activator UV radiation was capable of

acti-vating both As-MKK6b and c (results not shown)

The differences cannot be attributed to variability in

protein expression because western blot analysis of cell

extracts detected equal amounts of total myc-tagged

As-MKK6s The observed difference in As-MKK6a, b

and c activation by different stimuli suggests that these

salmon MKK6s are differentially regulated

Sorbitol induced activation of p38 in salmon TO

cells does not require MKK6a, b or c

The phosphorylation of endogenous MKKs was

exam-ined in TO cells using the commercial MKK3⁄ 6

phos-pho-specific antibody By sodium arsenite stimulation,

one band was detected with a regular substrate,

whereas three different bands were detected with a

ultrasensitive substrate, varying in size in the range 38–44 kDa (Fig 6A) Band 1 in Fig 6A, with an approximately size of 38 kDa, may correspond to MKK6a An increased phosphorylation of the band was detected already after 5 min and the phosphoryla-tion was more or less constant over the whole time course Another band of approximately 40 kDa was detected after 15 min (Fig 6A, band 2) This band corresponded in size to MKK6b or 6c RT-PCR results obtained with primers specific for As-MKK6b and 6c showed that neither MKK6b nor MKK6c was expressed in TO cells (results not shown), which excludes the possibility that band 2 represents these MKK6 variants The phospho-MKK3⁄ 6 antibody is known to weakly cross react with phosphorylated MKK4 [7] A third band (Fig 6A, band 3) was detected in the experiment, which corresponds to the size of MKK4 (approximately 44 kDa) and may repre-sent a salmon MKK4 ortholog In sorbitol stimulated

A

B

C

Fig 5 Activation of As-MKKs by diverse stress (A) CHSE-214 cells were transfected with myc-tagged As-MKK6a wild-type (wt) expression vector and treated with sodium arsenite (SA; 250 l M ), sorbitol (0.3 M ), UV radiation (120 mJÆcm)2) for indicated time points, or left untreated Cells were lysed and myc-tagged proteins were immunoprecipitated from the whole cell extracts (WCE) fol-lowed by an in vitro kinase assay (KA) using salmon His-As-p38a as

a substrate As-MKKs activities were analyzed by western blot anal-ysis detecting phosphorylated His-As-p38a (B) CHSE-214 cells were transfected with myc-tagged As-MKK6b wt expression vector and treated with sodium arsenite (SA; 250 l M ), sorbitol (0.3 M ), or otherwise as in (A) (C) CHSE-214 cells transfected with myc-tagged As-MKK6c wt expression vector, or otherwise as in (A) Phosphorylated His-As-p38a was detected by immunoblotting using anti-phospho-p38 serum (p-p38; upper panel) and exogenously expressed myc-MKKs was detected in cell extracts using anti-myc serum (lower panel) Experiments were performed twice with reproducible results.

TO cells, only the 38 kDa band was detected at levels

comparable with the control, indicating that this

stimulant did not induce MKK6 phosphorylation

(Fig 6A) Interestingly, although we were unable to

detect any MKK6 activation upon sorbitol

treatment, it was shown to phosphorylate salmon p38

(Fig 6A)

Sodium arsenite activation of endogenous MKK6 in

TO cells was further demonstrated by measuring the

ability of MKKs to phosphorylate p38 in vitro The

MKKs were immunoprecipitated by the pan-MKK6

antibody before measuring their ability to

late p38 Figure 6B shows that higher p38 phosphory-lation by salmon MKK6a was observed at 50 min post stimulation compared to the activity at 20 min of stim-ulation A kinase assay of sodium arsenite and sorbitol stimulated TO cells over-expressing As-MKK6a revealed that only sodium arsenite activated As-MKK6a in TO cells (Fig 6C) These results are in agreement with the results shown in Fig 6A, where no endogenous MKK phosphorylation was detected in sorbitol stimulated TO cells

Activation of p38 independently of MKK3⁄ 6, but dependent on p38 autophosphorylation, has been

A

B

D

C

Fig 6 Sorbitol induced activation of p38 MAPK in TO cells does not involve any of the MKK6 paralogs (A) TO cells were treated with sodium arsenite (250 l M ), sorbitol (0.3 M ) at indicated time points, or left untreated Cells were harvested and phosphorylated As-MKK6a, b,

c (p-MKK, first and second panel) and were visualized by western blotting using regular substrate (West Pico) or ultrasensitive substrate (West Femto) respectively Protein loading was verified in the whole cell extracts using the anti-eEF2 serum (lower panel) Phosphorylated As-p38a was detected by immunoblotting using anti-phospho-p38 serum (p-p38; third panel) (B) TO cells were either treated with 250 l M sodium arsenite or left untreated Cells were harvested at indicated time points and endogenous As-MKK6a, b and c were immunoprecipitated with anti-pan-MKK6 from the whole cell extracts (WCE) Activities were detected by kinase assay (KA) using His-As-p38 as substrate Phos-phorylated His-As-p38a was detected by immunoblotting using anti-phospho-p38 serum (p-p38; upper panel) and eEF2 was detected from whole cell extracts (lower panel) (C) TO cells were transfected with myc-tagged As-MKK6a wild-type expression vector After 48 h, the cells were treated with 250 l M sodium arsenite, 0.3 M sorbitol for 30 min, or left untreated Cells were lysed and myc-tagged proteins were im-munoprecipitated from the lysate followed by an in vitro kinase assay Phosphorylated His-As-p38a was detected by immunoblotting using anti-phospho-p38 serum (p-p38; upper panel) To confirm exogenous protein expression, the whole cell extract were blotted and probed with anti-myc serum (lower panel) (D) TO cells were transfected with GFP tagged As-p38a After 48 h, the cells were treated with 10 l M SB203580 for 1 h or left untreated, followed by stimulation with 250 l M sodium arsenite (SA) or 0.3 M sorbitol for 30 min Cells were lysed and phosphorylated p38a were detected by immunoblotting using anti-phospho-p38 serum (p-p38; first panel) The expression of GFP-p38a was verified with anti-GFP (second panel) Phosphorylation of endogenous MK2 was detected with anti-phospho-MK2 (third panel) and anti-eEF2 (fourth panel) was used as a loading control Experiments were performed twice with reproducible results.

reported previously [40,41] Because the p38 specific

inhibitor SB203580 blocks the ability of p38 to be

autophosphorylated [42], we further examined

whether this inhibitor would prevent p38

phosphory-lation in stress-activated TO cells As shown in

Fig 6D, p38 phosphorylation was not affected by

the SB inhibitor, suggesting that p38 activation in

sorbitol stimulated TO cells is not due to p38

auto-phosphorylation

Salmon MKK6a, As-MKK6b and As-MKK6c

activate As-p38a, As-p38b1 and As-p38b2 in

CHSE-214 cells

The p38 MAP kinases are known substrates for

MKK3 and MKK6 in mammalian cells [13,15,16,19–

21,23] We have recently described three p38a

variants in Atlantic salmon, which all possess the

putative dual phosphorylation motif Thr-Glu-Tyr in

the activation loop as well as the docking motifs

reported to be important for docking to activators,

substrate and regulators Moreover, all three As-p38a

variants were shown to be phosphorylated in

CHSE-214 cells stressed with sodium arsenite [2] To

explore the ability of As-MKK6a, As-MKK6b and

As-MKK6c to activate the different As-p38 variants,

constitutively active As-MKKs were constructed

Mammalian constitutive active MKK3⁄ 6 is generated

by replacing the phosphorylation sites Ser and Thr

with the phospho-mimicking glutamic acid (EE) [43]

Constitutive active As-MKKs were generated based

on the same principle Furthermore, catalytic inactive

As-MKKs mutants were constructed by replacing the

aspartic acid in the conserved DFG motif, known to

be essential for catalytic activity [44], with alanine

(DA) CHSE-214 cells were co-transfected with

gluta-thione S-transferase (GST)-MKK EE or DA and

myc-tagged p38 variants, followed by

immunoprecipi-tation and p38 kinase assay using recombinant

ATF-2 as substrate All three As-p38 variants were

activated by constitutive active As-MKK6a,

As-MKK6b and As-MKK6c (Fig 7) As-MKK6b

caused the strongest As-p38a activation among the

three MKKs (Fig 7A, upper panel), whereas

As-MKK6c was the dominant activator of As-p38b1

(Fig 7B, upper panel) In the case of As-p38b2, all

three MKKs showed similar levels of activation

(Fig 7C, upper panel) The ability of the

immuno-precipitated p38 to phosphorylate ATF-2 in vitro

correlated well with results from western blotting

using phospho-p38 antibodies to detect the

exoge-nous and endogeexoge-nous p38 phosphorylation directly

in the lysates of transfected cells (data not shown)

Discussion

In the present study, we report the cloning of three cDNAs encoding different salmon MKK6 sequences The identity between the As-MKK6b and 6c was 94%, whereas their identities to MKK6a were approximately 81% Because the nonmatching nucleotides in the sequences of these MKKs were spread throughout the whole ORF, it is unlikely that the different MKK6

A

B

C

Fig 7 As-MKK6a, b and c are upstream activators of As-p38a, p38b1 and p38b2 CHSE-214 cells were transfected with GST-tagged constitutive active (EE) or catalytic inactive (DA) MKK6a, b

or c expression vectors together with either myc-tagged As-p38a (A), As-p38b1 (B) or As-p38b2 (C) After 24 h, the cells were lysed and myc-tagged p38 was immunoprecipitated from the whole cell extracts (WCE) lysate followed by an in vitro kinase assay (KA) using ATF-2 as p38 substrate The As-p38 activity was analyzed by detecting incorporated phosphate into ATF-2 by autoradiography (first panel) To confirm exogenous protein expression, the whole cell extract were blotted and probed with anti-GST (second panel) and anti-myc (third panel) sera.

cDNAs represent different splice variants Thus, our

data suggest that there exist at least three MKK6 genes

in salmon They all contained the phosphorylation

sites Ser and Thr in the activation loop and also the

N- and C-terminal docking domains shown to be

important for MKK activation and substrate

specific-ity Further analysis of the salmon MKK6s revealed

that they all were able to phosphorylate and activate

salmon p38

In higher vertebrates such as man and mouse, the

two genes MKK3 and MKK6 encode proteins that are

the primary p38 activators, whereas in invertebrates

such as Drosophila, Caenorhabditis elegans and in

yeast, there are only a single activator for their p38

orthologs [32,33,45] A single MKK3⁄ 6 ortholog has

been described in the fish species carp and zebrafish

that causes selective activation of p38 in vitro [27,29]

Our phylogenetic analysis suggests that the ancestral

MKK3⁄ 6 gene has undergone two major duplication

events (Fig 1B), one of the events can be observed in

tetrapods, which have MKK3 and MKK6, and the

other can be observed in fish which have one or two

copies of MKK6 Hence, MKK3 does not appear to be

present in fish, and the zebrafish MKK3 sequence

should be renamed to MKK6, which is in agreement

with the name given this sequence in the ZFIN and

UniProt databases Salmon appears to be unique in

having a third MKK6 copy, possibly reflecting the

salmonid specific tetraploidization event [46] and, to

our knowledge, this is the first report of the existence

of three MKK6 isoforms in any species

Inspection of the genomic sequences of five fish

species (zebrafish, green pufferfish, fugu, medaka and

stickleback) revealed evidence that may suggest that

the two MKK6 sequences present in some fish may be

the result of the early ray-finned fish tetraploidization

event (results not shown) In green pufferfish and

medaka, the two fish species that have two copies of

the MKK6 gene and where the chromosomal location

is known, the two MKK6 copies are located on

differ-ent chromosomes This is in contrast to the human

genome where MKK6 and MKK3 are located on the

same chromosome Furthermore, all MKK6 genes, for

which genomic sequence is available, are in synteny

with a gene encoding a protein homologous to the

human G protein-coupled receptor family C protein

(UniProt: Q9NQ84)

All three salmon MKK6 genes showed ubiquitous

tissue distribution and almost similar expression levels

in the different tissues analyzed The transcript length

of MKK6b and 6c was approximately 1.7 kb, whereas

the MKK6a probe revealed an approximately 4.0 kb

transcript and two smaller transcripts of approximately

1.4 kb and 1.7 kb, respectively The latter two were mainly expressed in the ovary However, we cannot exclude the possibility that these transcripts represent other closely related MKKs or are MKK6a splicing variants Due to the high identity between MKK6a and MKK6b⁄ c (approximately 81%), the MKK6a probe may also weakly hybridize to the MKK6b⁄ c transcript The 1.7 kb band seen with the MKK6a probe could therefore represent the MKK6b⁄ c tran-script The distribution of the salmon MKK6s resem-bles the wide tissue distribution of salmon p38 [2]

In addition to its involvement in responses to stress and inflammatory stimuli, the p38 kinase signaling pathway also participates in processes during normal development Zebrafish p38 is involved in the control

of blastomere cleavage during embryogenesis [27,29] and a specific temporal expression pattern is seen in throughout zebrafish embryogenesis [47], suggesting an important role during early development Studies on salmon [2] and carp [29] have demonstrated high expression of both p38 and MKK6 in the ovary The abundance of piscine p38 signalling module members

in this organ may suggest that the p38 pathway aids their survival against environmental stresses during early development Carp MKK6 possess a nuclear export signal sequence that does not exist in the MKK3⁄ 6 families in other species [29] Such a nuclear export signal was not found in the salmon MKK6 sequences reported in the present study

In salmon, three genes encoding p38a isoforms have been identified and all three were activated by stress-inducing and inflammatory stimuli The existence of several p38 genes in salmon may be a way for cells to respond differently to upstream kinases and extra-cellular stimuli, which also has been reported in other studies [8,20] Using constitutively activated MKK6 mutants, we were able to demonstrate that all three MKK6s variant could activate the different salmon p38 variants This was shown by a kinase assay detect-ing ATF-2 phosphorylation Furthermore, the results were verified using a GAL4-responsive and ATF2 dependent luciferase reporter assay, where over-expressing constitutive active As-MKKs mutants increased ATF2-dependent gene expression by six- to 10-fold compared to catalytic inactive As-MKKs mutants (data not shown)

Analysis of p38 activation in mouse fibroblasts lack-ing MKK3 or MKK6, and stressed with UV radiation, anisomycin or sorbitol, shows that mammalian MKK6 and MKK3 play redundant roles in response to these stressors [7,24,48] Despite a very high identity between the salmon MKK6s, their activation pattern upon exposure to different stressors revealed differences We

found considerably more phosphorylated MKK6a and

6c compared to MKK6b in UV stressed CHSE-214

cells over-expressing the three As-MKK6s Moreover,

exposure to the stressors sorbitol and sodium arsenite

resulted in notable differences in response between the

MKK6 isoforms, as measured by their kinase activity

using recombinant salmon p38 as substrate For

MKK6b, the activation by sorbitol was much more

pronounced compared to sodium arsenite, whereas it

was the opposite for MKK6c Moreover, As-MKK6a

responded equally to these stressors The results

suggest a selective activation of As-MKK6 by

extra-cellular stimuli, and surmise that different MAP3Ks

are involved in MKK6 activation in response to

alter-nate forms of stress It is interesting that p38 is the

only substrate for the MAP2Ks MKK3 and MKK6,

whereas a much wider repertoire of different MAP3Ks

having the ability to phosphorylate and activate

MKK3 and MKK6 exist [35] A C-terminal docking

site called the DVD domain (i.e domain for versatile

docking) consisting of 24 amino acids is essential for

activating mammalian MKKs by specific MAP3Ks

[34] As a consequence, this docking domain may

influ-ence the ability of MKK6 to become phosphorylated

in response to various extracellular stressors The role

of this docking domain in the requirement of MKK3

or MKK6 to be activated by different MAP3Ks is not

known We observed that the corresponding region of

the As-MKK6b and c displayed four amino acids that

are different from MKK6a Whether the divergence in

sequence between the As-MKKs in this region can be

explained by selective substrate specificity of MAP3Ks

in response to different stress needs further

investi-gation

In extracts prepared from TO cells exposed to

cellu-lar stress, we found several bands, representing

puta-tive phosphorylated MKKs, that cross-reacted with

this phospho-antibody Consistent with the results

using ectopically expressed MKK6s, the results

obtained showed that the response was determined by

the extracellular stimuli that were used for activation

In sodium arsenite treated TO cells, two bands

(approximately 38 kDa and 42 kDa, respectively)

showing increased phosphorylation upon activation

were detected, whereas, in sorbitol treated cells, only

basal phospho-MKK6 levels were detected when using

a ultrasensitive substrate The results of a kinase assay

using over-expressed MKK6a verified that only sodium

arsenite and not sorbitol stimulated its activation in

TO cells Despite the inability of sorbitol to induce the

phosphorylation of MKK6a in TO cells,

phosphory-lated p38 was detected in these cells upon both sodium

arsenite and sorbitol treatment The results suggest the

existence of yet other MKK ortholog(s) that phosphor-ylate p38 in sorbitol stimulated TO cells Because p38 phosphorylation was not affected by the SB inhibitor,

it is less likely that p38 activation in sorbitol stimu-lated TO cells is caused by p38 autophosphorylation Interestingly, knockdown of the Drosophila p38 activa-tor D-MKK3 by RNA interference showed a significant, although incomplete, reduction of phosphorylated p38 levels in response to osmotic stress [49], suggesting the existence of another p38 activator in Drosophila By using UV stressed mouse embryonic fibroblast cells lacking both MKK3 and MKK6, it was possible to show that MKK4 participates in the activation of p38 However, the level of activated p38 in these cells was much lower compared to wild-type cells, whereas the level of phosphorylated p38 was not affected in MKK4-single deficient cells [7] This indicates that MKK3 and MKK6 are the main p38 activators, whereas, under certain circumstances, MKK4 partici-pates in the activation of p38 We therefore find it unlikely that MKK4 is the main p38 activator in TO cells stimulated with sorbitol

In conclusion, we have identified three upstream activators of p38 in Atlantic salmon, which all appear

to be MKK6 orthologs Our phylogenetic analysis strongly indicates that MKK3 is not present in fish The ancestral MKK6 gene appears to have undergone duplication in some fish species and our data demon-strate, for the first time, the existence of three MKK6 copies in any species The results obtained from CHSE-214 cells and TO cells suggest a cell type depen-dent expression and activation of the salmon MKK6 variants Thus, in a whole organism, expressing these MKK6genes at different levels may increase the range

of possibilities available to fine tune the strength of p38 signaling in specialized cells

Experimental procedures

Reagents and antibodies Sodium arsenite and sorbitol were obtained from Sigma (St Louis, MO, USA) The p38 inhibitor SB203580 was purchased from Alexis Biochemicals (Lausen, Switzerland) Recombinant ATF-2 and rabbit antibodies against phos-pho-p38 MAPK, phospho-MKK3⁄ 6, phospho-MK2 and eukaryotic elongation factor 2 (eEF2) were obtained from Cell Signaling Technology (Beverly, MA, USA) Rabbit anti-actin serum and mouse anti-GST serum were pur-chased from Sigma and Santa Cruz Biotechnology (Santa Cruz, CA, USA), respectively Mouse anti-myc serum was purified from the 9E10 hybridom, and rabbit anti-green fluorescent protein (GFP) was obtained from Abcam