Autophosphorylation Of Jnk2

[Cell Cycle 6:14, e1-e10, EPUB Ahead of Print: http://www.landesbioscience.com/journals/cc/abstract.php?id=4434; 15 July 2007]; ©2007 Landes Bioscience

Report

Autophosphorylation Properties of Inactive and Active JNK2 Genaro Pimienta1 Scott B. Ficarro3 Gustavo J. Gutierrez2 Anindita Bhoumik2 Eric C. Peters3 Ze’ev Ronai2 Jaime Pascual1,*

Abstract

.

UT E

Introduction

The c‑Jun N‑terminal kinases (JNKs) are a subgroup of the mitogen activated protein kinases (MAPKs), cloned initially from cDNA libraries of human liver1 and rat brain.2 These protein homologues were shown to phosphorylate the transcription factor c‑Jun on its N‑terminal region, upon activation by TNFa (ref. 1 and ultraviolet (UV) radiation.2‑4 JNKs relay cellular signals conveyed by the MAPK cascade, when stimulated by physical stress, inflammatory cytokines, such as IL‑1 and TNF, or DNA‑damaging agents in the form of UV or ionizing radiation.5 The JNK signalling cascade relies on protein scaffolds to achieve specificity and spatial localization of the transferred signal.6 It also functions within the context of other signalling systems, such as the NFB cascade7‑9 and those orchestrated by the PKC.10,11 A consequence is that JNK signalling can initiate events of cell cycle progression or apoptosis, depending on the cell‑type and the nature of a given stimuli.12 Current models suggest that a transient stimulation of JNK generally leads to cell proliferation, whereas a persistent activation is conducive to apoptosis.12,13 Three genes are found in humans and mice that code for three JNK protein homologues giving rise to 10 differentially spliced isoforms.14 JNK1 and JNK2 are ubiquitous, whereas JNK3 is tissue‑specific and found mainly in the brain.14,15 Like other MAPKs, JNKs have a conserved protein kinase fold and a typical activation loop. This loop has the sequence signature of Thr (T) ‑x‑ Tyr (Y) motif, where x is a Pro (P) in the case of JNKs.16 A hallmark in JNK activation is its induction by the dual phosphorylation (on T and Y) of the aforementioned motif.17,18 This event is orchestrated by two upstream dual‑specificity (S/T and Y) MAPK kinases, the MKK419,20 and MKK7,21,22 which in turn are activated by a large number of upstream S/T MKK kinases (MAP3K).17,18 JNKs are fully active only when MKK4 and MKK7 act synergistically, to achieve the dual phosphorylation of its activation loop.23‑25 The bulk of endogenous JNK is represented by two JNK homologues (p46 and p55) that derive from the genes jnk1 and jnk2 respectively. Their amino acid sequences are about 80% identical, differing mainly by a C‑terminal extension that makes JNK2 (p55) longer than JNK1 (p46).14 Despite having similar gene expression patterns, their apparent redundancy is controversial and their individual biological relevance still not deciphered.

IEN

This manuscript has been published online, prior to printing for Cell Cycle, Volume 6, Issue 14. Definitive page numbers have not been assigned. The current citation is: Cell Cycle 2007; 6(14): http://www.landesbioscience.com/journals/cc/abstract.php?id=4434 Once the issue is complete and page numbers have been assigned, the citation will change accordingly.

IST

RIB

Original manuscript submitted: 04/23/07 Manuscript accepted: 05/10/07

OT D

*Correspondence to: Jaime Pascual; Burnham Institute for Medical Research; 10901 North Torrey Pines Road; La Jolla, California 92037 USA; Tel.: 858.646.3100; Fax: 858.646.3195; Email: [email protected]

ON

Institute of the Novartis Research Foundation; San Diego, California

USA

.D

for Medical Research; La Jolla, California USA

CE

1Inflammation and Infectious Diseases Center; 2Cancer Center; Burnham Institute

3Genomics

The c‑Jun N‑terminal kinases (JNKs) are ubiquitous proteins that phosphorylate their substrates, such as transcription factors, in response to physical stress, cytokines or UV radiation. This leads to changes in gene expression, ensuing either cell cycle progression or apoptosis. Active phospho JNK1 is the main in vivo kinase component of the JNK cascade, whereas JNK2 is presumed not to participate as a kinase during JNK signal‑ ling. However, there is evidence that JNK isoforms interact functionally in vivo. Also, a recent chemical genetics investigation has confirmed that JNK transient activation leads to cellular proliferation, whereas a sustained one is pro‑apoptotic. Here we investigate the phosphorylation pattern of JNK2, with protein biochemistry tools and tandem mass spectrometry. We choose to focus on JNK2 because of its reported constitutive activity in glioma cells. Our results indicate that purified JNK2 from transfected nonstressed 293T cells is a mixture of the mono‑sites pThr183 and pTyr185 of its activation loop and of pThr386 along its unique C‑terminal region. Upon UV stimulation, its phosphorylation stoichiometry is upregulated on the activation loop, generating a mixture of mono‑pTyr185 and the expected dual‑pThr183/pTyr185 species, with the pThr386 specie present but unaltered respect to the basal conditions.

SC

Key words

Acknowledgements

ES

BIO

JNK2/tandem mass spectrometry/autophosphorylation/time-dependent phosphorylation pattern

Note

LA

ND

This work was partially funded by grant P01CA102583 to J.P.

©

20

07

Supplementary material can be found at: www,.landesbioscience.com/supplement/ pimientaCC6-14-sup.pdf

e1

Cell Cycle

2007; Vol. 6 Issue 14

Autophosphorylation Properties of Inactive and Active JNK2

with 20 mL of either Nickel‑NTA or Glutathione Sepharose High Performance resin, for His or GST fusion proteins, respectively. After extensive washing (5 column volumes) with equilibration buffer (25 mM Tris‑HCl, pH 8.0, 500 mM NaCl, supplemented with 2.5 mM imidazole in the case of Nickel column), the bound proteins were eluted with a linear gradient of either 2.5–200 mM imidazole for His fusion proteins or 0–50 mM reduced glutathione in the case of GST fusion polypeptides. The elution peak was collected and dialyzed over‑night against 2 L of 25 mM HEPES, pH 7.6, 50 mM NaCl and 10 mM DTT. Proteins were purified further by anion exchange, with a manually packed column containing 20 mL of Mono Q Sepharose resin, using a linear NaCl gradient 0‑1M, with the same buffer conditions used to dialyze overnight. The pure proteins were concentrated by centrifugation with an Amicon Ultra concentrator (10,000 Da molecular weight cut off, MILLIPORE), to about 1 mg/mL and used immediately or flash frozen with liquid nitrogen and stored at ~80˚C. SDS‑PAGE and Western blotting protocols. SDS‑PAGE and Western blot (WB) transfers were performed on preassembled Bio‑Rad systems, using standard protocols, with precast 4–20% Tris‑Glycine minigels and polyvinylidene difluoride (PVDF) membranes, respectively (Invitrogen). The samples were loaded with reducing buffer. SDS‑PAGE gels were stained and imaged with Pro‑Q Diamond and Experimental Procedures Sypro Ruby as indicated by the manufacturer (Invitrogen). PVDF Protein expression constructs. The DNA expression constructs membranes were probed sequentially with primary rabbit anti‑JNK encoding GST‑c‑Jun N‑terminal fragment (a.a. 1‑89), Flag‑JNK2, total, mouse anti‑pTP motif and mouse anti‑pY monoclonal antiHis‑JNK1 and His‑JNK2 are those described previously.10 Mutations bodies (Cell Signaling). For imaging, Alexa Fluor700‑conjugated on His‑JNK2 were introduced by using the Quick Change anti‑mouse/rabbit secondary antibodies were used (Cell Signaling). Site‑Directed Mutagenesis Kit (Stratagene) and confirmed by DNA An ODYSSEY infrared imaging system (LI‑COR) was utilized for sequencing. Four single point mutants (K55R, T183D, Y185E and blot documentation and analysis. T243A) were used in this study, one double phosphomimic point In vitro kinase assays. Kinase assays were performed at 30˚C for mutant (T183D/Y185E) and a truncated form (a.a. 1‑387) of JNK2 an hour. Pure recombinant His‑JNK and its substrate GST‑c‑Jun that overlaps with JNK1. N‑terminal were mixed at various ratios in reaction buffer (25 mM Protein expression in E. coli. The GST and 5xHis fusion HEPES, pH 7.4, 1 mM DTT, 25 mM MgCl2, 150 mM NaCl proteins were produced growing bacteria in Luria‑Bertani Broth and 50 (M ATP). Kinase reactions were separated by SDS‑PAGE (LB) supplemented with ampicillin at 37˚C. Typically, a fresh and documented with ProQ‑Diamond/Sypro Ruby staining or by colony transformed with a particular DNA construct, was picked exposure of the dried SDS‑PAGE on an X‑ray film (Fuji). When up from a solid LB plate and used to inoculate 1–2 mL of liquid radioactivity was used, 5 (M of (32P)(‑ATP was added per reaction. LB and incubated overnight. Next, this preculture was diluted to Cell culture and transient transfection. Human embryonic 1 L of LB in a 2 L flask and cultured until the optical density of the kidney (HEK) 293T cells were maintained at 37˚C and 5% CO2, in media measured at 600 nm reached a value of 0.6–0.9. Recombinant Dulbeccoís modified Eagle’s medium (DMEM), supplemented with protein over‑expression was induced with 0.25 mM of IPTG at this bovine serum (10%) and penicillin/streptomycin 1% (v/v). Cells were point. Protein expression experiments lasted either 4 or 12 hours and transfected at 30–35% confluent with a Flag‑JNK2 DNA construct. were performed at 15˚C or 37˚C, depending on the experiment. See This was done by calcium phosphate or with the Lipofectamine the results sections for details. Plus Reagent (Invitrogen), following the manufacturerís protocol. At Protein purification from E. coli. All recombinant proteins were 18 hours post‑transfection, the medium was aspirated and replaced prepared in a similar way. Only the affinity column used to enrich with fresh one. them was different, as specified below. For the liquid chromatogUV‑stimulation of cells and time course analysis. At near 100% raphy (LC) protocols described here, we have used a ƒKTA‑prime confluent, cells were either exposed to UV (45J) and harvested or fast performance liquid chromatography (FPLC) system, with chro- harvested nonstimulated. The medium from each plate was aspirated matographic columns (XK 16), affinity resins and other LC utilities, and the cells washed twice with ice‑cold phosphate buffered saline from GE‑Amersham Biosciences. E. coli expression milieu was (PBS). The open plates where exposed to 45J of UV and immediately harvested at 4˚C by centrifugation (4500 rpm) during 30 minutes, incubated with fresh medium for a fixed amount of time (0.5/1/4/8 using a Sorvall SLA30 rotor. Cells were resuspended in ice‑cold lysis hours). Each UV‑stimulation time point was harvested and lysed, buffer (25 mM Tris‑HCl, pH 8.0, 500 mM NaCl and 1 tablet/L of as indicated below. A control time point (time zero) of cells not an EDTA‑free protease inhibitors cocktail from Roche) and lysed treated with UV was also prepared. For each time point, 5–10 plates by sonication. Next, the lysate was centrifuged with a Sorvall SS34 (100 mL) were cultured and transfected to obtain sufficient Flag‑JNK2 rotator at 12,000 rpm/4˚C, to remove the insoluble cell debris. for proteomic analysis. Purification of flag‑tagged JNK2 from 293T cells. Confluent The resulting supernatant was loaded onto a 50 mL super‑loop and injected on a preequilibrated affinity column, manually packed ± UV treated cells were washed twice with ice‑cold PBS. The cells

©

20

07

LA

ND

ES

BIO

SC

IEN

CE

.D

ON

OT D

IST

RIB

UT E

.

In particular, genetic studies with mice embryonic fibroblasts (MEFs) suggest that active JNK1 is the main in vivo kinase, relaying the JNK cascade signal.26,27 This has lead to the model that JNK1 and JNK2 have distinct, rather opposite functional roles.28 This concept has recently been challenged by a chemical genetics report, which shows that JNK1 and JNK2 display a functional in vivo cross talk and rather have compensatory functional roles.29 A possible explanation is that the differences among JNKs are dictated at the post‑translational level, in the form of a differential phosphorylation pattern. If true, this may underpin their in vivo functional cross‑talk. Here we explore this hypothesis by investigating with biochemical tools and tandem mass spectrometry (MS), the phosphorylation pattern of JNK2 when overexpressed in E. coli and 293T cells. We choose to work with JNK2 (p55), because we find in the literature an emerging notion that JNK2 has autocatalytic properties.30‑32 Our results show that JNK2 but not JNK1 auto‑activates in vitro because it has dual‑specificity autophosphorylation properties. It phosphorylates itself on the T and Y residues of the activation loop. In addition we identify a previously unknown phosphorylation site, T386 that is an amino acid unique to JNK2 located along its C‑terminal extension.

www.landesbioscience.com

Cell Cycle

e2

©

20

07

LA

ND

ES

BIO

SC

IST

IEN

CE

.D

ON

were gently dislodged from the culture plate with a disposable cell scraper. This was done with the cells immersed in 5 mL of ice‑cold PBS, supplemented with phosphatase and protease inhibitors (phosphatase cocktail inhibitors 1 and 2 from Sigma, and protease inhibitors cocktail from Roche). The extract was transferred to a 50 mL Falcon tube (Corning) and harvested by centrifugation at 4000 rpm/4˚C. The cell pellet was either flash frozen in liquid nitrogen and stored at ñ80˚C or resuspended immediately in lysis buffer (25 mM Tris‑HCl, pH 7.5, 150 mM NaCl, 2 mM EDTA, 10% v/v glycerol, 1% v/v Triton X‑100, phosphatase cocktail inhibitors 1 and 2, and protease cocktail inhibitors) and incubated for one hour, with smooth rocking at 4˚C. Next, the lysate was centrifuged at 12,000 rpm/4˚C to remove the insoluble cell debris. The resulting supernatant was incubated for 1.5 hours with anti‑Flag antibody M2 Agarose beads (Roche), at 4˚C with gentle rotation. The Flag‑JNK2/ Agarose‑beads complex was harvested by centrifugation at 12,000 rpm/4˚C and washed 5 times with lysis buffer. Finally, Flag‑JNK2 was eluted by pH disruption, with 100 mM Glycine pH 3.0. The eluted protein was immediately neutralized with 1M Tris‑HCl, pH 8.0 and concentrated to a minimum volume by centrifugation with an Amicon Ultra concentrator (10,000 Da molecular weight cut off, Millipore). The samples were either flash frozen with liquid nitrogen and stored at ~80˚C or immediately analyzed by mass spectrometry (MS). Mass spectrometry experiments. Recombinant JNK2 protein was purified as described above from either E. coli or mammalian 293T cells. Equal amounts of the pure protein were subject to SDS‑PAGE electrophoresis and digested in‑gel with trypsin, chymotrypsin, or Asp‑N using a previously described method.33 Peptides were analyzed by automated nano‑LC/MS using a vented column strategy. Briefly, digests were loaded onto precolumns (360 m x 100 mm I.D. fused silica packed with 4 cm 5 mm Monitor C18 from Column Engineering, Ontario, CA) at a flow rate of 4 mL/min for 10 minutes using an autosampler and HPLC pump (Agilent, Palo Alto, CA). After sample loading, the vent was closed putting the precolumn in‑line with the analytical column (360 mm x 75 mm I.D. fused silica

OT D

Figure 1. JNK2 is a phosphoprotein when expressed in E. coli. The kinase activity of JNK2 was monitored with SDS‑PAGE and stained with Sypro Ruby (upper) and ProQ‑Diamond (lower). The reaction was set at various enzyme/substrate ratios, with values of 1/1; 1/3; 3/3 and 3/1 in mg/ml, as indicated on lanes 1–4 respectively. Lane 5 is the substrate as a negative control. Lanes 6 and 7 are recombinant JNK2 before and after treatment with l‑phosphatase.

RIB

UT E

.

Autophosphorylation Properties of Inactive and Active JNK2

e3

Figure 2. JNK2 unlike JNK1 autophosphorylates in vitro. (A and B) Time course of His‑JNK2 over‑expression in E. coli at 15˚C (A) and 37˚C (B). The SDS‑PAGE gels were stained sequentially with Sypro Ruby (upper) and ProQ‑Diamond (lower). In both cases the first aliquot was taken at time zero after IPTG induction (lane 1). Subsequent time points were sampled every 30 (lanes 2‑5), or 60 minutes (lanes 6 and 7), and after an over‑night incu‑ bation (lane 8). C, Comparing JNK1 and JNK2 expression with SDS‑PAGE, Sypro Ruby (upper) and ProQ‑Diamond (lower). Lanes 9–12 are respectively His‑JNK1, His‑JNK2 WT, a JNK2 kinase dead mutant (K55R) as a negative control, and a construct of JNK2 (1‑387) that lacks its distinctive C‑terminal region. Each lane corresponds to a total protein aliquot taken after four hours of IPTG induction at 37˚C.

packed with 8 cm 5 mm Monitor C18), and peptides were gradient eluted (0–25% B in 30 minutes, 25–90% B in 5 minutes; A = 0.1 M acetic acid in water, B = 0.1 M acetic acid in acetonitrile) into the mass spectrometer at a flow rate of approximately 100 nL/min.† Mapping JNK phosphorylation sites was achieved using two approaches. First, digests were analyzed with a linear ion trap mass spectrometer (LTQ, San Jose, California) operated in data‑dependent mode, where the top 5 most abundant precursor ions were subjected to MS/MS (spray voltage = 2000 V, collision energy= 35%, isolation width = 3 Da). MS/MS spectra were matched to JNK sequences using the SEQUEST algorithm. Next, an equivalent amount of digest was analyzed using a 4000 QTRAP hybrid triple quadrupole/ linear ion trap mass spectrometer (Applied Biosystems, Foster City,

Cell Cycle

2007; Vol. 6 Issue 14

CE

.D

ON

OT D

IST

RIB

UT E

.

Autophosphorylation Properties of Inactive and Active JNK2

IEN

Figure 3A. Ion chromatograms of the tryptic phosphopeptides observed for the TPY motif. Fragmentation data corresponding to the three phospho‑peptides identified for JNK2 activation loop. TACTNFMoxMoxpTPYVVTR. The ion chromatograms of the novel TP‑motifs, T243 and T386 we have identified, are shown in the supplementary material. Asterisk marks refer to water loses.

©

20

07

LA

ND

ES

BIO

SC

CA). The instrument was programmed to perform a three second precursor scan of m/z 79 in the negative ion mode from m/z 450 to 1800 (spray voltage = ‑2200 V, curtain gas = 10, declustering potential = 80, mode = peak hopping, step size = 1 Da, Q1 = low resolution, Q3 = unit resolution) using a collision energy ramp of ‑65 to ‑110 V. After the precursor scan, the top three ions above 5000 counts/sec were subjected to an enhanced resolution scan in the positive ion mode (spray voltage = 2400 V, Q1 resolution= open, scan rate = 250 amu/sec, dynamic fill time = on, TIC target = 5 x 1 e6 cps). These ions were then subjected to positive ion MS/MS using a charge state dependent rolling collision energy (scan rate = 4000 amu/sec, Q1 resolution =low, fixed fill time of 40 ms, Q0 trapping = on, Q3 entry barrier = ‑8 V). Peak lists were generated using Analyst software and searched using Mascot (Matrix Science). Once the sites of phosphorylation were identified, specific phosphopeptides were monitored in targeted experiments using one of two approaches. When quantifying more than nine peptides, multiple reaction monitoring (MRM) experiments were performed on the QTRAP instrument (3–5 transitions/peptide) with a dwell time of 50 ms/transition, Q1 resolution set to low, Q3 resolution set to unit, and a collision energy derived from previous MS/MS experiments. When nine or fewer peptides needed to be monitored, targeted MS/MS experiments were performed on the LTQ linear ion trap mass spectrometer with an isolation width of 3 Da and a collision energy of 35%. Peptide and peptide fragment peak areas www.landesbioscience.com

were determined using XCalibur 2.0 or Analyst 1.4.1 software.† Normalization across time course samples was accomplished using signals from nonphosphorylated peptides.†

Results Recombinant JNK2 (p55) but not JNK1 (p46) autophosphory‑ lates during its overexpression in E. coli. Recent work has suggested that JNK2 may be capable of autophosphorylation in vitro,1,25,31 and that this may in part explain why in certain types of cancer, JNK2 is constitutively active.30,32 In accordance, we observe here that E. coli‑produced recombinant JNK2 is considerably active towards its substrate c‑Jun (Fig. 1). To monitor the kinase reactions, we have used the set of chemiluminiscent custom dyes ProQ‑Diamond and Sypro Ruby. Pro‑Q Diamond selectively stains phospho‑proteins on SDS‑PAGE, which once destained and washed, can be restained with Sypro Ruby to observe the total protein content.34 This strategy allowed us to find that E. coli‑produced JNK2 is a phospho‑protein because it is recognized by ProQ‑Diamond, already prior to its incubation in a kinase reaction assay (Fig. 1). As a proof of concept, ProQ‑Diamond fails to recognize JNK2 on an SDS‑PAGE if the protein is treated with (l‑phosphatase, Fig. 1). The autophosphorylation of recombinant MAPKs in E. coli has been previously shown not to be an artifact.35 Therefore, we postulate that JNK2 auto‑ activates in vitro by means of an auto‑phosphorylation event. We also

Cell Cycle

e4

.D

ON

OT D

IST

RIB

UT E

.

Autophosphorylation Properties of Inactive and Active JNK2

IEN

CE

Figure 3B. Ion chromatograms of the tryptic phosphopeptides observed for the TPY motif. Fragmentation data corresponding to the three phospho‑peptides identified for JNK2 activation loop. TACTNFMoxMoxTPpYVVTR. The ion chromatograms of the novel TP‑motifs, T243 and T386 we have identified, are shown in the supplementary material. Asterisk marks refer to water loses.

©

20

07

LA

ND

ES

BIO

SC

reasoned that this happens during the overexpression of JNK2, inside the E.coli milieu. To explore this hypothesis, we traced the autophosphorylation of recombinant JNK2 over‑time, during its overexpression in E. coli. JNK2 expresses well at 15˚C (Fig. 2A) and 37˚C (Fig. 2B), but its autophosphorylation is observed with ProQ‑Diamond only at 37˚C (Fig. 2B). To validate our findings, we compared the expression profile of JNK2 wild type versus poly‑His‑tag fusions of JNK1 wild type, a well‑documented JNK2 kinase dead mutant (K55R), and a truncated JNK2 construct that lacks its unique C‑terminal region (a.a. 1‑387) (Fig. 2C). We found that JNK1 is not phosphorylated. K55R is a known kinase dead mutant that we use here as a negative control. The failure of JNK1 to autoactivate in vitro is in agreement with previous published work regarding the heterologous production of active recombinant MAPKs.36,37 Mass‑spectrometry analysis of phospho‑JNK2 overexpressed in E. coli. We extended our investigation of JNK2 phosphorylation properties using LC/MS. For this purpose, recombinant JNK2 was over‑expressed during 12 hours at 37˚C, to assure maximum autophosphorylation. Pure JNK2 aliquots were digested with three different enzymes, trypsin, Asp‑N or chymotrypsin, to obtain a peptide mixture representing most of the protein sequence. Each protein digest set was then analyzed by LC/MS, using a data dependent tandem MS/MS protocol, and a precursor scanning method capable of selective phosphopeptide detection. The overall analysis resulted in the identification of 17 fragments (Supplementary Table S1), allowing the unequivocal assignment of 12 phospho‑residues. e5

Interestingly, we find that the phosphorylation of the TPY motif (activation loop) is a mixture of one‑site phosphorylated peptides (pT183 or pY185), and the biologically relevant phosphorylation of both sites (pT183/pY185) (Fig. 3). It is well accepted that JNKs phosphorylate their substrates on Ser/Thr residues N‑terminal to a Pro (S/TP‑motifs), and so referred to as Pro‑directed kinases.17 JNK2 has four TP‑motifs (T93, T183, T243 and T386) along its protein sequence. Here we have obtained peptide fragmentation data for the four of them and observe autophosphorylation on T183 from the activation loop (as mentioned above), and on T243 and T386. The other TP‑motif in JNK2 (T93) is not phosphorylated. We also find several phosphorylation‑sites (pT255, pS282, pS377 and pS388) located C‑terminal, rather than N‑terminal, of a Pro residue (PS/T‑motifs). S282 is found along the sequence FPS(282)ES(284)ER and the proximity of another Ser (S284), precluded the precise identification of the phosphorylated residue. Three non-S/TP phosphosites were defined. These are pS87 that is N‑terminal to a Leu, S416 preceding a Thr, and T417 N‑terminal to a Gly. Finally, two ambiguous phosphosites were found within a Ser‑rich sequence, DAAVp(SS)NApT(386)PSQp(S SS)IN, from which pT386 is part. We estimate, based on the specificity of the enzymes used, to have obtained a sequence coverage of 81% along the full protein (343 out of 424 residues), covering a total of 85% of the S/T residues. The activation loop controls the autophosphorylation of JNK2. Our next step was to distinguish the specific autophosphorylation events, from those that may have derived as artifacts from the

Cell Cycle

2007; Vol. 6 Issue 14

.D

ON

OT D

IST

RIB

UT E

.

Autophosphorylation Properties of Inactive and Active JNK2

IEN

CE

Figure 3C. Ion chromatograms of the tryptic phosphopeptides observed for the TPY motif. Fragmentation data corresponding to the three phospho‑peptides identified for JNK2 activation loop. TACTNFMoxMoxpTPpYVVTR. The ion chromatograms of the novel TP‑motifs, T243 and T386 we have identified, are shown in the supplementary material. Asterisk marks refer to water loses.

©

20

07

LA

ND

ES

BIO

SC

E.coli overexpression conditions. To do so, we sought to establish the timing of appearance of the various autophosphorylation sites during the over‑expression of JNK2. The His‑JNK2 expression levels started to increase drastically at about 90 minutes post‑induction and its phosphorylation is observed with ProQ‑Diamond signal at about 150 minutes after IPTG induction (Fig. 2B). Based on this, we sampled the JNK2 expression at discrete time points after induction from minutes 90 to 240. The samples were immediately harvested without membrane disruption prior to SDS‑PAGE and LC/MS analysis. The reason to analyze nondisrupted E. coli milieu, was based on the assumption that physical and/or chemo‑enzymatic factors that follow cell membrane disruption, could promote undesirable JNK2 autophosphorylation events. By performing a relative quantification analysis of the LC/MS data, we find that pY185 and the pTP motifs (pT183, pT243 and pT386) appear early, upon IPTG induction, and that with the exception of pY185, their stoichiometry increases slowly and with a linear fashion until about minute 150. We observe an exponential increase of T386 between minutes 150‑180, and of T183 and T243 between minutes 210‑240 (Fig. 4). In the case of the Tyr‑containing phosphopeptides, pY185 remains rather constant, with a moderate increase along the time course, whereas the double phosphosite pT183/pY185 is observed only at the last time point, minute 240 (Fig. 4). It is worth mentioning that the non-TP‑motif phosphosites we had identified from a 12 hours expression experiment (Table S1) were not identified this time. We therefore suggest that the crowding conditions to which JNK2 is exposed at some point during its www.landesbioscience.com

overexpression, leads to its unspecific autophosphorylation of nonTP‑moieties. For this reason, we decided to focus further on the 4 phospho‑sites appearing within the time frame 90‑240 mins. To dissect their importance in JNK2 autophorylation, we performed a western blot (WB) experiment with anti‑pTP motif, anti‑pY and anti‑nonphosphorylated JNK monoclonal antibodies. We probed these antibodies against JNK2 (WT), the alanine substitution of the TP‑motif T243A outside the activation loop, and several point mutants that aimed at mimicking the activation loop phosphosites (T183D, Y185E and T183D/Y185E). The mutant T386A was not included in this experiment, because it did not express well in E. coli (data not shown). As summarized in Table 1, we find that the mutants T183D and T183D/Y185E fail to autophosphorylate on the pTP motifs, whereas Y185E and T243A are active enough to do so. Also only T183D/Y185E has a negative anti‑pY signal (Fig. 5A). To extend our investigation, we also tested these JNK2 constructs with ProQ‑Diamond. Here we found that from the two mutants with no anti‑pTP WB signal, T183D gives a positive ProQ‑Diamond signal, whereas T183D/Y185E does not (Fig. 5B). In the case of T183D, the ProQ‑Diamond signal is consistent with its anti‑pY signal due to the presence of pY185. As mentioned above, the heterologous production of MAPKs in their active form is only accomplished if an upstream kinase (i.e., MKK7 for JNK) is coexpressed with/or fused to a given MAPK.36,37 Accordingly, the mutants T183D, Y185E and T183D/Y185E do not phosphorylate the bonafide JNK substrate c‑Jun (Fig. 5C), meaning that they are not properly mimicking the conformation of the

Cell Cycle

e6

ND

ES

BIO

SC

IEN

CE

.D

ON

OT D

IST

RIB

UT E

.

Autophosphorylation Properties of Inactive and Active JNK2

LA

Figure 4. Time‑dependent phosphorylation of JNK2 in E. coli. The time in minutes after IPTG‑induction (horizontal axis) is plotted against the relative amount, in the form of normalized counts (vertical axis) of each pTP motif: pT183 (A), pT243 (B), pT386 (C), the mono‑pY185 (D) and dual‑pT183/pY185 species (E).

©

20

07

activation loop. The Pro‑Q Diamond signal observed for T183D and Y185E suggests that the autophosphorylation and the phosphorylation of the substrate (c‑Jun) are independent. Finally, the remaining phosphorylated TP‑motifs we observed may have a structural role or represent protein‑protein interaction sites. In agreement with this, the mutant T183D prevents the appearance of the other two pTP‑motifs (pT243 and pT386). JNK2 purified from nonstimulated 293T cells is a mixture of the monophosphosites pT183, pY185 and pT386. To corroborate the biological relevance of our findings in E. coli, we cultured 293T cells that were transiently transfected with Flag‑tagged JNK2. For this, JNK2 was immunoprecipitated with agarose‑bound anti‑Flag monoclonal antibodies and subjected to biochemical inspection. From its ProQ‑Diamond signal, we find that Flag‑JNK2 is e7

phosphorylated when isolated from nonstimulated 293T cells (Figure S1, supplementary material). Analysis of the purified samples by LC/MS shows that basal Flag‑JNK2 is phosphorylated on T183, Y185 and T386 (Fig. 6A–C). Previous studies show that in nonstimulated KB and PC12 cells, MKKs are not active.38 Thus, the phospho‑signal we observe for basal Flag‑JNK2 derives most probably from an autophosphorylation event. In addition, the fact that one of the phospho‑sites is Tyr185 implies that JNK2 has dual‑specific (S/T and Y) auto‑phosphorylation properties. This notion is compelling because, as stated above, JNKs had long been conceived as Pro‑directed S/T kinases. Finally, T386 appears here as well, supporting its relevance (Fig. 6C). Mass‑spectrometry analysis of UV‑activated JNK2. From a large‑scale purification protocol, we were able to obtain (mg amounts

Cell Cycle

2007; Vol. 6 Issue 14

Autophosphorylation Properties of Inactive and Active JNK2

Table 1

Specific monoclonal-antibody reactivity and Pro-Q Diamond staining output for His-JNK2 WT and various site-specific mutants



WT T243A T183D Y185E T183D/Y185E α- pTP-motif

+

+ -

+ -

α- pY

+

+

+

+ -

α-JNK

+

+

+

+

ProQ-Diamond

+

+

+

+ -

Discussion

ND

ES

BIO

SC

UT E RIB IST OT D

IEN

CE

.D

ON

of Flag‑JNK2. This allowed us to investigate how the phosphorylation pattern of active JNK2 fluctuates over time in 293T cells. We chose to irradiate the cells with UV, because this is by far, the physical agent that activates JNKs the most. We found that upon UV‑treatment, active JNK2 is a mixture of four phosphorylated species. These were: mono‑pT183, mono‑pY185, the expected double phospho‑specie pT183/pY185 on the activation loop, and the pT386 signal coming from its unique C‑terminal region (Fig. 6A–D). To establish how the different phospho‑species fluctuate as a function of time, we analyzed the phosphorylation state of active JNK2 at different time points (0.5, 1, 4 and 8 hours), after UV‑stimulation (Fig. 6A–D). As expected from the available literature,17,18 the double‑phospho moiety pT183/pY185 reached a maximum 30–60 minutes after UV‑stimulation and then slowly decayed. This was also the case for pY185, upregulated at about the same time and stoichiometry. In contrast, pT183 decreased, whereas pT386 population remained relatively constant along the time course. The TP‑motif phospho‑site T243 that we found in active JNK2 obtained from bacteria, was not observed here. The appearance of active JNK2 as a mixture of pY185 and pT183/pY185 suggests the existence of a mixed protein population in active JNK2 in vivo. Also, the decay in the values of pT183 in parallel to the appearance of pT183/pY185 indicates that the activation of JNK2 follows an ordered distributive path, in which basal pT183 is converted to the fully active moiety pT183/pY185 upon UV‑light irradiation. As for pT386, a likely explanation is to assume a structural role for this phosphosite, being involved in either inter or intra protein‑protein interactions, and inaccesible to phosphatases. Accordingly, recombinantly produced T386A mutant is unstable when expressed in E. coli.

.

+

©

20

07

LA

In this article we have combined protein biochemistry and tandem MS to investigate the auto‑phosphorylation pattern of recombinant JNK2. With a combination of ProQ‑Diamond and Sypro Ruby staining of reducing electrophoretic gels, the detected JNK2 autophosphorylation in the E. coli milieu happens at 37˚C but not at 15˚C, suggesting that our observations derive from a catalysed chemical process. In agreement, Flag‑JNK2 immuno‑precipitated from nonstimulated 293T cells, is stoichiometrically auto‑phosphorylated on T183, Y185 and T386. It is well established that JNKs become fully active only when phosphorylated on both T183 and Y185 of their TPY motif.17,18 In the case of JNK1, it has been suggested that its full activation requires first the phosphorylation at position Y185 followed by that of the Thr.39 From our LC/MS results on JNK2, the main event that follows UV radiation is the net increase of the pY185 signal as a mixture of mono‑pY185 and the fully active form dual‑pT183/ www.landesbioscience.com

Figure 5. The autophosphorylation of JNK2 in E. coli is controlled by its activation loop. The constructs tested in (A–C) are WT (lane 1), T243A (lane 2), T183D (lane 3), Y185E (lane 4) and the double mutant T183D/Y185E (lane 5). (A) WB analysis with several commercially available monoclo‑ nal antibodies, anti‑pTP motif (upper), anti‑pY (middle) and anti‑total JNK (lower). The same blot was stripped after each antibody reaction. As for the positive anti‑pY signal we see for Y185E, we speculate that this may be either an unspecific recognition by the antibody or the presence of other pY residues besides Y185 that get phosphorylated unspecifically. (B) ProQ‑Diamond (upper) and Sypro Ruby (lower) staining analysis. (C) In vitro kinase assay monitored with 32P‑l ATP (lanes 1 and 3–5 as above) using c‑Jun N‑terminal as the substrate. Top gel is radioactive labeled, bottom gel Coomassie stained.

pY185. We think that the pT183 form is converted to pT183/ pY185, because its stoichiometry decreases during the time frame in which active JNK2 is upregulated and then slowly revertes to its basal levels as active JNK2 decays. Also, the presence of mono‑pY185 in the total amount of post‑activation JNK2, leads to a net low molarity or “dilution” of fully active JNK2 (pT183/pY185). We envisage that in the case of JNK1, the absence in basal conditions of the mono phosphorylated species may in turn lead to a higher net molarity of pT183/pY185 JNK1 relative to active JNK2 (Fig. 7). Regarding the auto‑phosphorylation of JNK2 on T386, this phospho‑site may also have a biological meaning, yet to be elucidated. The reason for this is that TP‑motifs are bona‑fide protein‑ protein interaction sites.40,41 In agreement, the C‑terminal region of JNK2, where T386 is located, has been proposed to function as a substrate‑binding scaffold.4 A plausible scenario would be that

Cell Cycle

e8

Autophosphorylation Properties of Inactive and Active JNK2

ND

ES

BIO

SC

IEN

CE

.D

ON

OT D

IST

RIB

UT E

.

Figure 6. Phosphorylation pattern of Flag‑JNK2 inactive and active isolated from 293T cells. The LC/MS data for each phosphorylation site identified was analyzed by relative quantifica‑ tion. The resulting values plotted as counts normalized per phosphorylation site (vertical axis) versus time after UV‑light irradiation (hori‑ zontal axis) are shown. (A) pT183; (B) pY185; (C) pT386 and (D) pT183/pY185.

07

LA

Figure 7. JNK2 (p55), as opposed to JNK1 (p46), is a phospho‑protein in basal conditions. Based on the observed basal autophosphorylation of JNK2, we propose that during a total JNK response curve, JNK2 is activated first but with an attenuated net activity, whereas JNK1 with no basal autophosphorylation will be activated later during the response curve but with a higher net activity. For simplicity, we assume in our model that the phospho‑ site pT386 has a structural role and is therefore present in all the protein components of the JNK2 pool. See the Discussion section for details.

©

20

the pT386 site by interacting with other JNKs or with putative protein scaffolds, becomes solvent inaccessible and thus incapable of downregulation by phosphatases. This argument would explain why the stoichiometry of the pT386 specie does not fluctuate upon UV‑activation and why the mutant T386A is unstable when over‑expressed in E. coli. An emerging concept is that MAPKs convert graded stimuli into switch like responses, because they operate at near saturation stoichiometries of either the activating upstream kinases and/or the inactivating phosphatases.42‑44 It has been shown experimentally that the JNK signaling cascade exerts a positive feedback loop, resulting from a still unidentified autocatalytic behaviour.45,46 These findings e9

corroborate the early prediction that JNKs exhibit an ultra‑sensitive response pattern when activated.42 In addition, while it is well established that active JNK1 is the main in vivo kinase component of the JNK cascade,17,18 JNK2 has been described as having a futile activity perhaps masked by an intracellular inhibitor.26 From our experimental findings, we propose that active JNK2, through its stoichiometrical pY185 component, may be responsible for the saturation conditions in which both the activating kinases and the phosphatases are predicted to operate. Also, since we expect the 10 JNK isoforms to have different kinetic rate constants, a potential compensatory equilibrium between them is likely to occur (Fig. 7). This prediction may underpin the recently proposed

Cell Cycle

2007; Vol. 6 Issue 14

Autophosphorylation Properties of Inactive and Active JNK2

.

UT E

RIB

IST

References

©

20

07

LA

ND

ES

BIO

SC

IEN

CE

.D

ON

1. Kyriakis JM, Avruch J. pp54 microtubule‑associated protein 2 kinase: A novel serine/threonine protein kinase regulated by phosphorylation and stimulated by poly‑L‑lysine. J Biol Chem 1990; 265:17355‑63. 2. Derijard B, Hibi M, Wu IH, Barrett T, Su B, Deng T, Karin M, Davis RJ. JNK1: A protein kinase stimulated by UV light and Ha‑Ras that binds and phosphorylates the c‑Jun activation domain. Cell 1994; 76:1025‑37. 3. Hibi M, Lin A, Smeal T, Minden A, Karin M. Identification of an oncoprotein‑ and UV‑responsive protein kinase that binds and potentiates the c‑Jun activation domain. Genes Dev 1993; 7:2135‑48. 4. Kallunki T, Su B, Tsigelny I, Sluss HK, Derijard B, Moore G, Davis R, Karin M. JNK2 contains a specificity‑determining region responsible for efficient c‑Jun binding and phosphorylation. Genes Dev 1994; 8:2996‑3007. 5. Weston CR, Davis RJ. The JNK signal transduction pathway. Curr Opin Genet Dev 2002; 12:14‑21. 6. Morrison DK, Davis RJ. Regulation of MAP kinase signaling modules by scaffold proteins in mammals. Annu Rev Cell Dev Biol 2003; 19:91‑118. 7. Tang G, Minemoto Y, Dibling B, Purcell NH, Li Z, Karin M, Lin A. Inhibition of JNK activation through NF‑kappaB target genes. Nature 2001; 414:313‑7. 8. Tang F, Tang G, Xiang J, Dai Q, Rosner MR, Lin A. The absence of NF‑kappaB‑mediated inhibition of c‑Jun N‑terminal kinase activation contributes to tumor necrosis factor alpha‑induced apoptosis. Mol Cell Biol 2002; 22:8571‑9. 9. Nakano H, Nakajima A, Sakon‑Komazawa S, Piao JH, Xue X, Okumura K. Reactive oxygen species mediate crosstalk between NF‑kappaB and JNK. Cell Death Differ 2006; 13:730‑7. 10. Lopez‑Bergami P, Habelhah H, Bhoumik A, Zhang W, Wang LH, Ronai Z. RACK1 mediates activation of JNK by protein kinase C [corrected]. Mol Cell 2005; 19:309‑20. 11. Liu J, Yang D, Minemoto Y, Leitges M, Rosner MR, Lin A. NF‑kappaB is required for UV‑induced JNK activation via induction of PKCdelta. Mol Cell 2006; 21:467‑80. 12. Lin A. Activation of the JNK signaling pathway: Breaking the brake on apoptosis. Bioessays 2003; 25:17‑24. 13. Ventura JJ, Hubner A, Zhang C, Flavell RA, Shokat KM, Davis RJ. Chemical genetic analysis of the time course of signal transduction by JNK. Mol Cell 2006; 21:701‑10. 14. Gupta S, Barrett T, Whitmarsh AJ, Cavanagh J, Sluss HK, Derijard B, Davis RJ. Selective interaction of JNK protein kinase isoforms with transcription factors. Embo J 1996; 15:2760‑70. 15. Finch A, Davis W, Carter WG, Saklatvala J. Analysis of mitogen‑activated protein kinase pathways used by interleukin 1 in tissues in vivo: Activation of hepatic c‑Jun N‑terminal kinases 1 and 2, and mitogen‑activated protein kinase kinases 4 and 7. Biochem J 2001; 353:275‑81. 16. Chen Z, Gibson TB, Robinson F, Silvestro L, Pearson G, Xu B, Wright A, Vanderbilt C, Cobb MH. MAP kinases. Chem Rev 2001; 101:2449‑76. 17. Kyriakis JM, Avruch J. Mammalian mitogen‑activated protein kinase signal transduction pathways activated by stress and inflammation. Physiol Rev 2001; 81:807‑69. 18. Karin M, Gallagher E. From JNK to pay dirt: Jun kinases, their biochemistry, physiology and clinical importance. IUBMB Life 2005; 57:283‑95. 19. Derijard B, Raingeaud J, Barrett T, Wu IH, Han J, Ulevitch RJ, Davis RJ. Independent human MAP‑kinase signal transduction pathways defined by MEK and MKK isoforms. Science 1995; 267:682‑5. 20. Lin A, Minden A, Martinetto H, Claret FX, Lange‑Carter C, Mercurio F, Johnson GL, Karin M. Identification of a dual specificity kinase that activates the Jun kinases and p38‑Mpk2. Science 1995; 268:286‑90. 21. Moriguchi T, Toyoshima F, Masuyama N, Hanafusa H, Gotoh Y, Nishida E. A novel SAPK/JNK kinase, MKK7, stimulated by TNFalpha and cellular stresses. Embo J 1997; 16:7045‑53. 22. Tournier C, Dong C, Turner TK, Jones SN, Flavell RA, Davis RJ. MKK7 is an essential component of the JNK signal transduction pathway activated by proinflammatory cytokines. Genes Dev 2001; 15:1419‑26. 23. Lawler S, Fleming Y, Goedert M, Cohen P. Synergistic activation of SAPK1/JNK1 by two MAP kinase kinases in vitro. Curr Biol 1998; 8:1387‑90. 24. Lisnock J, Griffin P, Calaycay J, Frantz B, Parsons J, O’Keefe SJ, LoGrasso P. Activation of JNK3 alpha 1 requires both MKK4 and MKK7: Kinetic characterization of in vitro phosphorylated JNK3 alpha 1. Biochemistry 2000; 39:3141‑8. 25. Fleming Y, Armstrong CG, Morrice N, Paterson A, Goedert M, Cohen P. Synergistic activation of stress‑activated protein kinase 1/c‑Jun N‑terminal kinase (SAPK1/JNK) isoforms by mitogen‑activated protein kinase kinase 4 (MKK4) and MKK7. Biochem J 2000; 352:145‑54. 26. Liu J, Minemoto Y, Lin A. c‑Jun N‑terminal protein kinase 1 (JNK1), but not JNK2, is essential for tumor necrosis factor alpha‑induced c‑Jun kinase activation and apoptosis. Mol Cell Biol 2004; 24:10844‑56. 27. Sabapathy K, Hochedlinger K, Nam SY, Bauer A, Karin M, Wagner EF. Distinct roles for JNK1 and JNK2 in regulating JNK activity and c‑Jun‑dependent cell proliferation. Mol Cell 2004; 15:713‑25.

28. Sabapathy K, Wagner EF. JNK2: A negative regulator of cellular proliferation. Cell Cycle 2004; 3:1520‑3. 29. Jaeschke A, Karasarides M, Ventura JJ, Ehrhardt A, Zhang C, Flavell RA, Shokat KM, Davis RJ. JNK2 is a positive regulator of the cJun transcription factor. Mol Cell 2006; 23:899‑911. 30. Tsuiki H, Tnani M, Okamoto I, Kenyon LC, Emlet DR, Holgado‑Madruga M, Lanham IS, Joynes CJ, Vo KT, Wong AJ. Constitutively active forms of c‑Jun NH2‑terminal kinase are expressed in primary glial tumors. Cancer Res 2003; 63:250‑5. 31. Cui J, Holgado‑Madruga M, Su W, Tsuiki H, Wedegaertner P, Wong AJ. Identification of a specific domain responsible for JNK2alpha2 autophosphorylation. J Biol Chem 2005; 280:9913‑20. 32. Cui J, Han SY, Wang C, Su W, Harshyne L, Holgado‑Madruga M, Wong AJ. c‑Jun NH(2)‑terminal kinase 2alpha2 promotes the tumorigenicity of human glioblastoma cells. Cancer Res 2006; 66:10024‑31. 33. Shevchenko A, Wilm M, Vorm O, Mann M. Mass spectrometric sequencing of proteins silver‑stained polyacrylamide gels. Anal Chem 1996; 68:850‑8. 34. Steinberg TH, Agnew BJ, Gee KR, Leung WY, Goodman T, Schulenberg B, Hendrickson J, Beechem JM, Haugland RP, Patton WF. Global quantitative phosphoprotein analysis using Multiplexed Proteomics technology. Proteomics 2003; 3:1128‑44. 35. Seger R, Ahn NG, Boulton TG, Yancopoulos GD, Panayotatos N, Radziejewska E, Ericsson L, Bratlien RL, Cobb MH, Krebs EG. Microtubule‑associated protein 2 kinases, ERK1 and ERK2, undergo autophosphorylation on both tyrosine and threonine residues: Implications for their mechanism of activation. Proc Natl Acad Sci USA 1991; 88:6142‑6. 36. Khokhlatchev A, Xu S, English J, Wu P, Schaefer E, Cobb MH. Reconstitution of mitogen‑activated protein kinase phosphorylation cascades in bacteria. Efficient synthesis of active protein kinases. J Biol Chem 1997; 272:11057‑62. 37. Zheng C, Xiang J, Hunter T, Lin A. The JNKK2‑JNK1 fusion protein acts as a constitutively active c‑Jun kinase that stimulates c‑Jun transcription activity. J Biol Chem 1999; 274:28966‑71. 38. Meier R, Rouse J, Cuenda A, Nebreda AR, Cohen P. Cellular stresses and cytokines activate multiple mitogen‑activated‑protein kinase kinase homologues in PC12 and KB cells. Eur J Biochem 1996; 236:796‑805. 39. Kishimoto H, Nakagawa K, Watanabe T, Kitagawa D, Momose H, Seo J, Nishitai G, Shimizu N, Ohata S, Tanemura S, Asaka S, Goto T, Fukushi H, Yoshida H, Suzuki A, Sasaki T, Wada T, Penninger JM, Nishina H, Katada T. Different properties of SEK1 and MKK7 in dual phosphorylation of stress‑induced activated protein kinase SAPK/JNK in embryonic stem cells. J Biol Chem 2003; 278:16595‑601. 40. Yaffe MB, Elia AE. Phosphoserine/threonine‑binding domains. Curr Opin Cell Biol 2001; 13:131‑8. 41. Remenyi A, Good MC, Lim WA. Docking interactions in protein kinase and phosphatase networks. Curr Opin Struct Biol 2006; 16:676‑85. 42. Ferrell Jr JE. Tripping the switch fantastic: How a protein kinase cascade can convert graded inputs into switch‑like outputs. Trends Biochem Sci 1996; 21:460‑6. 43. Markevich NI, Hoek JB, Kholodenko BN. Signaling switches and bistability arising from multisite phosphorylation in protein kinase cascades. J Cell Biol 2004; 164:353‑9. 44. Hornberg JJ, Binder B, Bruggeman FJ, Schoeberl B, Heinrich R, Westerhoff HV. Control of MAPK signalling: From complexity to what really matters. Oncogene 2005; 24:5533‑42. 45. Bagowski CP, Ferrell Jr JE. Bistability in the JNK cascade. Curr Biol 2001; 11:1176‑82. 46. Bagowski CP, Besser J, Frey CR, Ferrell Jr JE. The JNK cascade as a biochemical switch in mammalian cells: Ultrasensitive and all‑or‑none responses. Curr Biol 2003; 13:315‑20.

OT D

functional cross‑talk amongst JNKs,29 where a rapid but attenuated activation of JNK2 may pose an ultra‑sensitive factor to the response curve, while the lagging activation of JNK1 may provide the functional output threshold.

www.landesbioscience.com

Cell Cycle

e10