1
MK2 phosphorylates tristetraprolin on in vivo sites including S178, a site required for 14-3-3 binding
Carol A. Chrestensen1, Melanie J. Schroeder2,*, Jeffrey Shabanowitz2, Donald F. Hunt2, Jared W.
Pelo3, Mark T. Worthington3, and Thomas W. Sturgill1,+
From the Departments of Pharmacology1, Departments of Chemistry and Pathology2, and the Digestive Health Center of Excellence3, University of Virginia Health Sciences Center,
Charlottesville, VA 22908
Running Title: MK2-induced phosphorylations of TTP +Corresponding author *Completed in partial fulfillment of the requirements for the Ph.D. degree, University of Virginia, Charlottesville, VA Address for correspondence: Thomas Sturgill M.D.-Ph.D., Box 800735, Department of Pharmacology, University of Virginia, Charlottesville, VA 22908 Telephone: 434-924-8659 Facsimile: 434-924-5207 eMail: [email protected]
Copyright 2003 by The American Society for Biochemistry and Molecular Biology, Inc.
JBC Papers in Press. Published on December 19, 2003 as Manuscript M310486200 by guest on February 18, 2018
http://ww
w.jbc.org/
Dow
nloaded from
2
Summary MK2 (MAPKAP kinase 2) is required for TNF synthesis. Tristetraprolin (TTP) binds to the 3’
UTR of TNF mRNA and regulates its fate. We identified in vitro and in vivo phosphorylation
sites in TTP using nHPLC-µESI mass spectrometry and novel methods for direct digestion of
TTP bound to affinity matrices (GSH-beads or anti-myc linked to magnetic beads). MK2∆3B,
activated in E. coli by p38α, phosphorylates TTP in vitro at major sites S52 and S178 (>10 fold
in abundance) as well as at several minor sites that were detected after enriching for
phosphopeptides with IMAC. MK2 phosphorylation of TTP creates a functional 14-3-3 binding
site. In cells, TTP was phosphorylated at S52, S178, T250, S316 and at SP sites in a cluster
(S80/82/85). Anisomycin treatment of NIH 3T3 cells increased phosphorylation of S52 and
S178. Over-expression of MK2 sufficed to increase phosphorylation of S52 and S178, but not
S80/82/85 or T250. Thus, S52 and S178 are putative MK2 sites in vivo. Identified
phosphosite(s) may be biologic switches controlling mRNA stability and translation.
Key words: MAPKAP kinase 2/mRNA stability/Tristetraprolin
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
3
Introduction
MK2 mediates several p38α,β MAP kinase dependent processes (for review see (1)),
demonstrated most clearly by results from targeted disruption of the MK2 gene in mice (2).
MK2 (-,-) mice have suppressed stress responses. Cellular studies show deficits in motility,
chemotaxis, and cytokine production. Macrophages taken from MK2 (-,-) mice exhibit normal
TNF mRNA induction in response to endotoxin but do not release TNF protein. Cellular TNF
protein was markedly decreased in MK2 (-,-) macrophages, suggesting a block in production of
TNF from TNF mRNA (2). TNF expression is regulated both via mRNA stability and
translation (3,4), but is not completely understood.
The p38α,β MAP kinase pathway regulates stability of mRNAs that contain AU-rich
elements (AREs) in their 3’ untranslated regions. Examples include TNF, COX-2, IL-6, and IL-
1β (4-6). Evidence chiefly comes from mRNA stabilization caused by transfection of
mutationally activated MEK3/MEK6 or by addition of agents that activate p38 MAPK, and
conversely from destablization caused by addition of a p38α,β MAP kinase inhibitor. Since
p38α,β is required to activate MK2, these experiments do not dissect contributions from MK2.
Studies in MK2 (-,-) cells suggest MK2 regulates stability of some cytokine mRNAs (2,4). Lasa
et al. (5) first reported that expression of a mutant of MK2 with constitutive activity stabilized
COX-2 mRNA in the presence of SB2035780 and that expression of a kinase-defective MK2
blocked the stabilization induced by activated MEK6, arguing that MK2 is necessary and
sufficient to induce stabilization of at least the COX-2 mRNA (5). How MK2 regulates cytokine
production post-transcriptionally is unknown. Mahtani, K.R. et al. (7) reported that
tristetraprolin (TTP) is an in vitro substrate for MK2, motivating the detailed studies we describe.
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
4
TTP (for review see (8)) destabilizes class II AREs and is the prototype for a non-zinc-
finger class of nucleic acid binding proteins. Destabilization requires integrity of TTP’s tandem
Cys3His RNA binding domains that coordinate zinc in a disk-like structure (9-11). TTP-null
mice exhibit many defects including inflammatory arthritis and systemic lupus erythematosis-
like symptoms attributed to increased production of TNF (12).
TTP is phosphorylated in cells treated with growth factors or cytokines. Phosphorylation
occurs at more than one site evident by appearance of two distinct slower migrating forms of
TTP on gels after stimulation that are reversed by phosphatase treatment (7,13). TTP undergoes
detectable gel shifts between ~1- 1.5h, and the shifts persists for several hours (7,13). The only
reported TTP phosphorylation site is S220, phosphorylated by ERK2 in vitro (14). p38α also
phosphorylates TTP in vitro, but at unresolved sites (13). S178 mutation in human TTP causes
loss of ability to bind 14-3-3, suggesting S178 is a phosphorylation site (15), but the responsible
kinase(s) have not been identified.
We studied phosphorylation of TTP by potently active, recombinant MK2∆3B produced
in E. coli as well as in situ phosphorylation of myc-TTP expressed in cells. We identified two
major (S52 and S178) and several minor MK2 sites. Description as major sites reflects
abundance of phosphopeptide and estimated stoichiometry in peptides recovered with the
methods we describe. Minor sites detected are not excluded as candidates for functionally
significant sites. We show that the major sites (S52, S178) phosphorylated by MK2 are
phosphorylated in vivo and their phosphorylation can be altered by manipulating the p38
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
5
pathway in vivo. Furthermore, phosphorylation of TTP by MK2 confers ability to bind 14-3-3
proteins. Phosphorylation(s) of TTP may provide biological switches to regulate its dynamic life
cycle and/or its functions to control mRNA stability and translation.
Experimental Procedures
Materials
Several plasmids were generous gifts from colleagues: C. Moroni provided murine TTP in
pCDNA3.1/myc-HisA (16); M. Cobb, pT7.5-mup38-huMEK4 (17); C. Marshall, pEF myc WT
and KD (D207A) huMK2 (18); M. Gaestel, pGEX-muMK2∆3B (19) and pGEX-MK2EE∆3B
(19). pKH3-hu14-3-3β is from our laboratory. Anti-TTP antibodies to the mouse whole protein
(13) and human C-terminal peptide (20) were kindly provided by J. Han and W. Rigby,
respectively. Commercially acquired reagents: horseradish peroxidase (HRP) linked anti-goat
antibodies (Santa Cruz Biotechnology); tosylactivated Dynabeads (Dynal Biotech); γ-bind
Sepharose plus, glutathione Sepharose 4B; HRP linked anti-mouse and anti-rabbit antibodies
(Amersham Biosciences); anisomycin, SB203580, and microcystin LR (Calbiochem); BHK-21
cells, NIH 3T3(ATCC).
Plasmids for expression of active MK2∆3B
A bicistronic plasmid (pTYB4 His6p38α-HA1MEK6EE) was created by moving the coding
region of His6-tagged WT p38α with the ribosomal binding site engineered after the stop codon
(17) into the NcoI site of pTYB4 (New England Biolabs). HA-tagged MKK6EE (excised from
the pcDNA3 parent vector with NotI and Bgl2 followed by Klenow fill-in) was blunt-cloned into
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
6
the p38-pTYB4 construct at the NruI site. PCR was used to create an IPTG inducible MK2
expression vector with a p15A origin of replication (pAC-pET MK2∆3B), similar in overall
strategy to construction of pAC-pET RSK2CTD (21). pAC-pET MK2∆3B expresses a GST-
MK2∆3B fusion protein having an intervening 70 aa leader polypeptide containing His6 and S
tags, in addition to thrombin and enterokinase cleavage sites. All constructs were verified by
sequencing.
Protein production and purification
BL21 E. coli were transformed with pAC-pET MK2∆3B with or without pTYB4 His638α-
HA1MEK6EE using appropriate antibiotic selections. Transformed cells were grown and
induced, lysates prepared, and proteins purified as described (21). GST-TTP was expressed and
purified as in (22). Purified proteins were quantified with the Bio-Rad Protein Assay™ and with
SDS-PAGE analysis. Phosphocellulose paper assays were used to monitor kinase activity of the
purified GST or His6-tagged proteins. The kinase reactions contained (final, in 40µl) 25mM
HEPES, pH 7.4, 5mM β-glycerophosphate, 100mM NaCl, 2mM DTT, 0.25mg/ml BSA, 10mM
MgCl2, 50µM [γ-32P]-ATP (~4000cpm/pmol), 50µM peptide substrate and were incubated at
30°C for 15 min.
TTP phosporylation and in gel digestion
Reactions to phosphorylate TTP in vitro for site identification by MS/MS (see below) used the
same kinase cocktail minus [γ-32P]ATP and usually omitted BSA. GST-TTP (2µg in 20µl final
volume) was phosphorylated with 60ng of MK2∆3B for 1h (30o). A parallel reaction was always
performed that included [γ-32P]ATP. Initially, in vitro kinase reactions were stopped with
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
7
sample buffer and run on 4-20% gradient mini gels. Phosphorylation of TTP could also be
visualized by gel shift on mini gels when separation times were extended so as to run lower
molecular weight species off the gel. Details are provided in the legends.
Cell culture, transfection, and lysates
NIH 3T3 cells were grown at 37°C with 5% CO2 in Dulbecco's modified Eagle's medium
supplemented with 10% newborn calf serum. BHK cells were grown similarly, substituting 5%
FBS for calf serum. Cells were transfected using Lipofectamine 2000 (Invitrogen) under normal
serum conditions, essentially as directed by the manufacturer using 32µg of DNA per 15cm dish
(16µg of TTP and either 16µg of MK2 or pCDNA3.1 plasmids). Post-transfection (~24hrs) the
cells were treated as described in the legends. Cell lysates were made as described (23) in 50mM
Tris, pH 8, at 4°C, 50mM NaCl, 1% Nonidet P-40, 0.5% Triton X-100, 1.5mM MgCl2, 10µg/ml
each of leupeptin and aprotinin, 1µM PefaBlock (Roche Applied Science), 200µM Na3VO4, and
1µM microcystin LR. The lysates were either processed immediately or quick frozen in liquid
N2 for later use.
Immunoprecipitation using magnetic beads
Tosylated Dynabeads (Dynal Biotech, Norway) were covalently attached to pure monoclonal
myc-antibody as directed by the manufacturer. Conjugated anti-myc IgG beads (1 x 108 of
Dynabeads) were incubated with cleared cellular lysates for 2h. After incubation, the beads
were magnetically concentrated and the lysate removed. The beads were washed 3X PBS
(Invitrogen) or for higher stringency 1X with PBS, 2X with PBS containing 0.1M NaCl, and 1X
with PBS. A portion of the IP was taken for Western blot analysis and the remainder was
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
8
washed 3X with 0.1M NH4HCO3. These NH4HCO3 washed samples were then subjected to
digestion and MS/MS analysis (see below).
Sample preparation for analysis with mass spectrometry
After experimentation with gel based methods that were found to be problematic (see Results
and Discussion), we devised a new scheme to be termed ‘on beads’ digestion. Kinase reactions
were terminated by adding cold GSH-Sepharose beads (20µl of a 1:1 slurry), placing the tubes
on ice, and washing several times with cold PBS to remove ATP (necessary for IMAC). ATP
removal was monitored for sufficiency by following loss of 32P from a parallel hot reaction.
Then the beads were washed 3X with 100mM NH4HCO3 and re-suspended in 50µL. Bound
TTP was reduced with DTT (5mM final, for 1hr at 51°C) followed by iodoacetamide alkylation
(15mM final, 45mins RT, dark). Sequencing grade trypsin (Promega, Madison, WI) was then
added (1:20 ratio) with agitation for 12hrs at RT. The beads were pelleted, and the supernatant
was transferred to a new tube where acetic acid was used to lower the pH to 3.5. To generate
samples for in vivo phosphorylation, the above “on-beads” digestion procedure was repeated for
the TTP immmunoprecipitated with Dynabeads except sample reduction and alkylation was
omitted to minimize antibody digestion. Also, the digestion time was reduced to 6h for the same
reason.
Mass spectrometry
The samples were analyzed by reverse phase nanoflow HPLC micro ESI tandem mass
spectrometry (RP nHPLC µESI MS/MS) on a ThermoFinnigan LCQ Deca or XP ion trap (San
Jose, CA) operating in either data-dependent or targeted mode. Immobilized metal affinity
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
9
chromatography (IMAC) was performed according to (24) except the peptides were not reacted
to form methyl esters, and the elution was done with 7µL of 500mM ascorbic acid. A gradient
consisting of 0–60% B in 50 min, 60–100% B in 10 min (A = 100mM acetic acid (Sigma-
Aldrich) in water, B = 70% acetonitrile (Mallinkrodt), 100mM acetic acid in water) was used to
elute peptides from the reverse-phase column to the mass spectrometer through an integrated
electrospray emitter tip with a flow rate of ~60nl/min. Spectra were acquired with the instrument
operating in the data-dependent mode throughout the HPLC gradient. In this mode, every ∼10
sec, the instrument cycled through acquisition of a full-scan mass spectrum and five MS/MS
spectra (3 Da window; precursor m/z ±1.5 Da, collision energy set to 35%, dynamic exclusion
time of 1 min) recorded sequentially on the five most abundant ions present in the initial MS
scan. For the in vivo KD and WT MK2 samples, analysis was performed in the data-dependent
mode and run in duplicate. Because this sample was more complex than the in vitro sample, the
gradient was changed to 0-15%B in 7 mins, 15-25%B 20 mins, 25-60%B 35 mins and 60-
100%B 10 mins. This allowed for better resolution of the TTP peptides. In addition, these
particular in vivo samples were subjected to a time-segmented targeted analysis in order to obtain
the most data points per peptide peak of interest. Four peptides and their corresponding
phosphorylated forms were chosen for this analysis. In targeted mode, the instrument was
instructed prior to the run to repeat a cycle consisting of a full MS scan followed by a targeted
MS/MS scan of the (M+2H)++ ion for each pair of non- and phosphorylated peptide based on
their retention times (collision energy set to 35%) (see Fig. 6C).
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
10
Peptide Identification
All MS/MS spectra acquired were searched against either a GST- or myc-TTP protein database
by using the SEQUEST algorithm (25). Search parameters included a differential modification
of +80 Da (presence or absence of phosphate) on serine, threonine, or tyrosine and a static
modification of 57 (alkylated Cys for the in vitro analysis only). All possible enzymatic
cleavage sites were considered. Peptide hits were manually confirmed.
Calculations
Estimated stoichiometries were calculated by dividing the sum of the area under the curve from
the phosphopeptide in all of its charge states and modified forms by the total peptide signal. In a
single sample, this is an approximation because phosphopeptides and their nonphosphorylated
forms generally do not ionize with the same efficiency. In addition, we estimated changes in
stoichiometry of site phosphorylation by ratios of the estimated stoichiometries. Provided that
this comparison is confined to the same phospho- and dephosphopeptide in the two conditions,
we can make direct comparisons of the relative change in stoichiometry between two conditions
with greater confidence.
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
11
Results
Selective activation of GST-MK2∆3B by p38α in E. coli
To activate MK2∆3B, we created a GST-MK2∆3B expression plasmid with a p15A origin of
replication (Methods) that would be compatible with the bicistronic vector expressing both
constitutively active MEK1 and WT ERK2 (17). This co-expression system was used previously
in our lab to activate GST-RSK CTD (21) and ERK2 has been reported to activate MK2 (26).
For this reason, we co-expressed MK2∆3B with active ERK2 but this resulted in no activation of
MK2∆3B. The two proteins also showed no detectable interaction, demonstrated by lack of
ERK2 retention on the affinity column (Figure 1A, panels 1-4).
We designed a new bicistronic plasmid expressing constitutively active HA-MEK6 and
WT HIS-tagged p38α. When MK2∆3B is expressed in E. coli with active p38α produced from
this bicistronic plasmid, the kinase cascade was reconstituted. MK2∆3B was phosphorylated,
demonstrated by reduced mobility of the protein in SDS-PAGE analysis (Fig. 1A & B) and
increased peptide kinase activity (Fig. 1C). In contrast to ERK2, p38α was retained on the
affinity column and binding to GST-MK2∆3B was disrupted by washes with 2M LiCl, consistent
with an electrostatic interaction (Fig. 1A, panels 5-6). Phosphorylated MK2∆3B is dramatically
more active than the mutant MK2EE∆3B (Fig. 1C).
To identify sites in MK2∆3B whose phosphorylation was induced by p38α in E. coli, a
tryptic digest of the activated, purified GST-MK2∆3B was analyzed with LC MS/MS.
Phosphorylation of two of three S/T-Pro sites, at T222, S272 and T334 in human MK2, are
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
12
required for activation (26). T222 is within the activation loop, S272 is located in the C-lobe of
the kinase domain (27), and T334 is adjacent to an autoinhibitory helix. Our analysis showed
that the equivalent three sites are phosphorylated on MK2∆3B activated by p38α in E. coli (data
not shown). Phosphorylation at the equivalent of T338, a presumed autophosphorylation site in
human MK2 (26), was also detected in the purified active enzyme. These same sites were not
detectably phosphorylated in MK2∆3B co-expressed with active ERK2 (data not shown).
MK2∆3B phosphorylates recombinant GST-TTP
GST-TTP protein was phosphorylated by MK2∆3B in a time-dependent manner similarly to the
known in vivo substrate HSP27 while GST alone was either minimally or not at all
phosphorylated (Fig. 2A; data not shown). MK2 phosphorylation sites in TTP were initially
analyzed using peptides recovered from silver-stained SDS-PAGE bands of phosphorylated TTP
(28). An example of a gel run for this analysis is shown in Fig. 2B. Phosphorylation of TTP
causes a mobility shift of TTP that can be resolved with increased electrophoresis. Analyses of
TTP from gels were problematic due to multiple phosphorylated bands, co-migration of the most
gel-shifted band with BSA, and poor recovery that precluded IMAC to identify minor sites.
‘On beads’ digestion of TTP increases both coverage and recovery
To resolve these problems, we devised ‘on beads’ digestion methodologies of general utility for
future MS/MS analyses. We purified the GST-TTP from the kinase reaction with GSH-
Sepharose beads. While still bound, TTP was reduced, alkylated and digested with trypsin. TTP
peptides in the supernatant were analyzed by reverse phase and IMAC MS/MS (Fig. 3). Sample
from ‘on beads’ digestion was found to have increased protein coverage (70%) compared to the
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
13
in-gel digested sample (49%). ‘On beads’ digestion allows for better protein recovery,
eliminates the possibility of selection bias in excision of gel bands, and provides more material
for repeated analyses from a given sample.
S52 and S178 are major in vitro sites for MK2
Reverse phase MS/MS (Fig. 3) allowed us to compare m/z intensities for both the
unphosphorylated and phosphorylated peptides and to calculate an estimate of the percentage of
TTP phosphorylated at a particular residue of interest (Table I). The ability to enrich for
phosphopeptides on an Fe3+ POROS 20 IMAC column (Fig. 3) significantly increased our ability
to detect phosphopeptides, allowing us to identify four additional in vitro phosphorylation sites.
The calculated percent phosphorylation from the reverse phase analysis corresponds well with
peptide intensity signals from the IMAC analysis. Because these calculations of stoichiometry
are semi-quantitative, we will describe the TTP residues we have defined as sites of MK2∆3B
phosphorylation as “major” and “minor” phosphosites.
In Table I, the sites identified in muTTP (accession P22893) include serines 52, 58, 105,
176, 178, and 316. Of note, phospho-S52 and phospho-S178 peptides were more than an order
of magnitude more abundant than the other phosphopeptides identified, and are therefore
referred to throughout as major sites. Two additional singly phosphorylated peptides were also
identified as minor sites (T249 or T250 and S264 or S266, respectively), based on the available b
and y ions. T250 or S264 if assigned are in S/T-P motifs that could alternatively be
phosphorylated by small amounts of p38α present in the preparations of GST-MK2 (Fig. 1A).
The other, more prevalent, sites are phosphorylated by MK2∆3B.
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
14
Minor sites S105 and S266 are in motifs that exactly match the current consensus
sequence for MK2 phosphorylation (ΦxRxxS) where Φ is a bulky hydrophobic residue (29).
The major MK2 sites on TTP have similarity to the current consensus but are not exact matches.
S178 has L at p-5; S52 has R at the p-3 position. Their preferential use may indicate
unrecognized determinants, possibly the p+1 F for S178, or secondary and tertiary factors. Other
MK2 substrates, including p16-Arc and vimentin, have non-consensus sites (30,31).
MK2 phosphorylation of TTP creates a 14-3-3 binding site
If S178 is a phosphorylation site for MK2, MK2 phosphorylation should cause a gain of function
for 14-3-3 binding. To test this, GST-TTP was phosphorylated in vitro by either MK2∆3B or
p38α and then tested for altered 14-3-3 binding compared to unmodified GST-TTP. As a source
we used BHK cell lysates that either were transfected with 14-3-3β (Fig. 4A) or were not
transfected (Fig. 4B). After incubation of phosphorylated TTP with lysate, GSH-Sepharose
beads were added to the mixture. After further incubation the beads were recovered and washed
extensively before elution of TTP for analysis by Western blotting with anti-TTP and pan anti-
14-3-3 antibodies. MK2 phosphorylation caused TTP to gel shift. GST-TTP phosphorylated by
MK2 bound 14-3-3 proteins from the lysates, including endogenous 14-3-3, while in contrast the
unphosphorylated and p38α phosphorylated proteins did so only slightly, and only when 14-3-3β
was over-expressed (Fig. 4A).
TTP phosphorylated by MK2 in vitro binds RNA
TTP phosphorylated by MK2∆3B bound a TNF ARE-probe (22) to a similar extent as untreated
TTP (data not shown). This suggests that phosphorylation of the MK2 sites does not inhibit
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
15
RNA binding. Similarly, phosphorylation of TTP by ERK, p38 or JNK did not alter binding of
TTP to a different ARE probe (32).
TTP is predominantly phosphorylated in growing BHK cells
A significant portion of the expressed myc-TTP in BHK cells is phosphorylated at the 24h time
point (Fig. 5; data not shown). Expressed myc-TTP was detected in Western blots using two
different antibodies, a polyclonal antisera to recombinant human TTP (left panel) (13) and
monoclonal antibody to the myc tag (right panel). Endogenous TTP in untransfected BHK cells
was not detectable. In lysates from these cells and others, we usually observe three TTP bands in
different proportions by Western blotting. In BHK cells, the majority of TTP was detected in the
band having the slowest mobility (Fig. 5).
We treated equivalent plates of BHK cells for 1h with or without 10µM SB203580, a
p38α,β inhibitor. SB203580 caused a small but significant downshift, evident by an increase in
the fraction of TTP migrating as the middle band (Fig. 5 and data not shown). This result
suggests that activation of pathways other than p38α,β are sufficient to induce shifted
phosphorylated states of TTP, an expected result since mitogens induce TTP and TTP
phosphorylation (14). Mitogen and cytokine stimuli may induce phosphorylation of the same or
overlapping physiologic sites. Since TTP was predominantly gel shifted and phosphorylated in
BHK cells, we expressed myc-TTP to generate mass spectrometry samples for identification of
in vivo phosphosites on TTP.
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
16
‘On beads’ digestion of TTP bound to anti-tag Ab enables direct in vivo MS/MS analyses
Our success with ‘on beads’ digestion in the in vitro studies led us to devise an ‘on beads’
approach of general utility for studying post translational modifications of proteins in cells.
Myc-antibody was covalently bound to magnetic Dynabeads, chosen because the
physicochemical properties of the beads decrease non-specific interactions. The antibody-linked
beads were used to immunoprecipitate myc-tagged TTP from transfected BHK cells. Then, after
extensive washing with PBS and 100 mM NH4HCO3, the antibody-bound protein was subjected
to trypsin digestion without reduction and alkylation in order to minimize digestion of the myc-
antibody.
Identification of TTP phosphosites in BHK cells
To identify the sites phosphorylated, we analyzed peptides obtained by ‘on beads’ digests of
myc-TTP from proliferating BHK cells by MS/MS. The resultant peptides enabled us to clearly
detect the major MK2 phosphorylation sites (S52 and S178), two minor sites (S316 and
T249/250) that could resolve to an S/T-P site (T250), and an additional S/T-P site, S82 (likely
S80/82/85), which was not phosphorylated in our in vitro experiments (data not shown). S220
was in an unfavorably sized tryptic peptide and was not detectable in vivo in either phospho- or
dephosphoform.
To determine whether the p38 pathway was contributing to phosphorylation of the major
MK2 phosphosites in BHK cells, we treated transfected cells in parallel with the p38α,β
inhibitor. SB203580 caused a small decrease in phosphorylation of S178 in myc-TTP (Fig. 5)
(the S52 site was not well resolved in these analyses, data not shown) consistent with the small
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
17
decrease in the fraction gel shifted. These results necessitated finding a suitable cell line in
which expressed TTP was less extensively phosphorylated so that phosphorylation of TTP might
be induced by experimental manipulation of the p38 pathways.
TTP was less extensively phosphorylated in both RAW 264.7 macrophages and NIH 3T3
cells (data not shown). 3T3 cells transfected well enough to provide sufficient myc-TTP and
RAW cells did not. BHK cells were transfected and maintained in complete medium containing
fetal bovine serum (FBS). Twenty-four h post transfection, 3T3 as well as BHK cells were
normal in appearance without evidence of apoptosis. Apoptosis occurs to a variable but
significant extent when some cell lines, such as 3T3 cells (BHK cells were not studied), are
transfected with TTP constructs using different conditions that included removal of serum for 6 h
(33,34). Continued maintenance in media with FBS may be important in avoiding apoptosis, but
was not studied here.
S52, S178, T250, S316 and SP sites within a cluster (S80/82/85) are phosphosites in 3T3
cells
We analyzed ‘on beads’ digests of myc-TTP in duplicate in the data-dependent mode. These
results confirmed that, as in BHK cells, S52 and S178 were the major phosphosites in 3T3 cells
(data not shown). We also pursued peptides of interest more aggressively, by operating the ion
trap in targeted mode (Fig. 6C). In this way, we obtained the most data points per peptide of
interest as well as better quality spectra. This helped to clarify the minor phosphorylation site as
T250 rather than T249 and revealed that the peptide PGPELSPSPTSPTATPTTSSR was actually
a co-eluting mixture of three singly phosphorylated S*P species: (Ser80, 82, and 85).
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
18
Both anisomycin treatment and MK2 expression increases S52 and S178 phosphorylation
As a test of p38 pathway responsiveness, we treated cells with or without anisomycin, which is
known to activate p38α and MK2 (35). In addition, 3T3 cells were co-transfected with TTP and
kinase defective (KD) or wild type (WT) MK2 plasmids. Treatment with anisomycin resulted in
increased phosphorylation at S52 and S178 (Fig. 6A). Co-expression of WT MK2 with TTP
resulted in a clear increase of S178 and S52 phosphorylation, but not S80/82/85 or T250, which
were possible MAP kinase sites. No increase in phosphorylation was observed with transfection
of the KD construct (Fig. 6B). The data obtained by averaging all three analyses as compared to
the targeted analysis alone produced very similar results. The KD construct did not act as a
dominant negative, as observed in other systems (36). We conclude that MK2 does
phosphorylate S52 and S178 in vivo, but our results do not rule out the possibility that these sites
may also be substrates for other kinases.
S52A/S178A TTP exhibits similar electrophoretic mobility to TTP
To determine the requirement for S52 and S178 for generation of phosphorylated forms with
reduced electrophoretic mobility, we made the double mutant. In cycling BHK cells,
S52A/S178A TTP has reduced electrophoretic mobility similar to wild type TTP (Fig. 6E). This
suggests that S52/S178 are not required for the changes in mobility. The reduced mobility of
S52A/S178A TTP is due to phosphorylation on other sites, evident by increased mobility after
treatment with calf intestinal phosphatase.
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
19
Discussion
TTP is downstream of MK2 Previous work showed that TTP was a substrate for MK2 without defining the sites of
phosphorylation (7). We confirm phosphorylation of TTP by MK2 and identify specific
phosphorylation sites at S52, S105, S58, S176, S178, and S316. Two other sites (T250 and
S264) we identified in vitro may be phosphorylated by p38α, since it is a minor contaminant in
our preparations of active MK2. All of these sites are novel excepting S178 previously identified
through mutational analysis as a necessary residue for 14-3-3 binding (15).
We showed that phosphorylation of TTP by MK2 creates a binding site for 14-3-3
proteins. (In a different pathway, MK2 phosphorylation of the tuberous sclerosis 2 gene product
also creates a 14-3-3 binding site and serves as a precedent for our finding (37)). Recently, a
further indication of a regulatory network involving MK2 and 14-3-3 proteins was obtained in a
functional proteomics screen to identify MK2 substrates (38). MK2 interacts with 14-3-3ζ and
phosphorylates the monomer preferentially on a site in the dimer interface, S58 in subsequence
GARRSS58, causing decreased dimerization. Thus, both TTP and MK2, but not p38α, are 14-3-
3 interactors.
All of the potential MK2 phosphorylation sites identified as in vitro sites are conserved as
equivalent sites in mouse, rat, bovine, and human TTP. Of these, only the S178 and S316 sites
are present in the closely related human proteins Tis11b/ERF-1 or Tis11d/ERF2, although
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
20
sequences around these sites are not completely conserved. TTP, but not Tis11b or Tis11d,
shows synergy with TNF to cause apoptosis (15).
In our in vivo studies, we observed TTP phosphorylation at S52 and S178, two minor
sites S316 and T250, and the cluster of singly phosphorylated S80/82/85 sites we did not observe
in vitro. These sites were found in both BHK and 3T3 cells. The other minor sites were not
observed in vivo. This lack of detection does not indicate that these sites are not present in vivo,
as factors may have prevented their observation.
TTP phosphorylation at the putative MK2 sites was manipulated in vivo both by
anisomycin treatment that activates p38α and MK2 and by over-expression of MK2.
Importantly, increasing cellular WT MK2 was sufficient to induce phosphorylation. KD MK2
failed to induce phosphorylation, demonstrating that the effect of WT MK2 required its kinase
activity. KD MK2 did not lower basal phosphorylation, similar to previous results where co-
transfection of a similar KD MK2 construct had no detrimental effect on LPS-stimulation of
TNF production in WT macrophages (36). Although it is difficult to prove that a
phosphorylation in vivo is direct for any kinase/substrate pair these correlations strongly support
our conclusion that MK2 directly phosphorylates TTP in vivo.
MK2 could phosphorylate TTP in either the nuclear or the cytoplasmic compartment or
both. MAP kinases p38α,β are activated outside and then enter the nucleus. MK2 is activated in
the nucleus, and once activated may phosphorylate nuclear targets (see (1)) before it is exported
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
21
to the cytoplasm. Export of active MK2 in stressed fibroblasts is rapid, and complete by 20 min
(39), prior to maximal TTP expression.
MK2 may not be the only kinase phosphorylating TTP at the putative MK2 sites we
identify. MAPKAP kinase 3 and MAPKAP kinase 5/Prak are closely related to MK2, and thus
candidates. MK2 itself has two forms in mammals, revealed by in gel kinase assays as p55 and
p45 bands (35), which both become absent in MK2 (-,-) macrophages (2). The origin of these
MK2 forms is still unclear because evidence for cDNAs in EST databases that could correspond
to the shorter form identified biochemically is rare (36). In addition, kinases phosphorylating
MK2 sites in vivo could include kinases in the ERK pathways.
Novel procedure for obtaining Mass Spectrometry samples
So far as we are aware, we are the first to utilize direct digestion of protein samples bound to
affinity purification matrices, either GSH-Sepharose beads or myc-antibody linked to magnetic
Dynabead, to generate peptides for LC MS/MS analysis. In both cases, digestion directly from
the beads was an efficient way of producing sample, with a considerable increase in recovery
that provided sufficient peptides for multiple analyses (data-dependent or targeted) and with
different types of chromatography (RP or IMAC) from a small amount of protein. We
phosphorylated 2µg of TTP, and could identify all of the sites from a single ‘on beads’ digest
using only 10% of the material. In addition, the ability to analyze a protein purified from a crude
lysate after ‘on beads’ digestion should find immediate applications. Many signaling proteins of
importance can be expressed in a functional form as GST-fusions, allowing for direct application
of ‘on beads’ digestion after reduction and alkylation of the recovered protein. These methods
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
22
are applicable to other affinity reagents such as Ni2+ NTA beads for His-tagged proteins
(unpublished results), maltose bound beads for MBP-tagged proteins, and others (40).
It is important to note that the approach we used to minimize antibody contamination of
the in vivo sample (eliminating both reduction and alkylation of the sample) can produce altered
cleavage patterns of a protein and is a potential cause for reduced protein coverage. This is a
potential limitation, depending on the targeted protein. Protein interactors if still present could
also reduce the extent of cleavage. This is offset by the improved recovery of peptides by
avoiding gels, and by the practicality for studying in vivo phosphorylations by minor alteration of
generally available immunoprecipitation protocols. Bosworth, H.B. et al. (41) also used the
resistance of native IgG to trypsin to identify antigen epitopes by MALDI/MS.
For myc-TTP, cleavage ‘on beads’ reduced the extent of digestion in comparison to in
vitro digests somewhat, but we still recovered peptides to define in vivo phosphosites and
estimate stoichiometries. We cannot exclude that we have missed sites. For example, S105 was
a minor in vitro phosphosite, and S105 peptides were not abundant enough to detect in the in
vivo sample. S105 phosphopeptide recovery may be affected by proximity to the Cys3His repeat
and its physical properties. The S105 peptide elutes very early in the gradient and, therefore,
may not retain well on C18. This may also partly explain the observation that the S105
phosphopeptide was not recovered after IMAC, which depends on a final separation by C18.
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
23
Points for TTP regulation by phosphorylation
In quiescent serum-starved fibroblasts or resting macrophages there is little TTP mRNA or
protein. Induction of TTP mRNA requires p38α,β (7) or ERK1,2 dependent pathways (42),
depending on the stimulus. The same pathways also induce mRNAs containing AU-rich
elements that are targets for TTP. In particular, LPS causes p38-dependent induction of TNF
and TTP mRNAs (43), but with differential regulation. TTP mRNA levels increase more
rapidly, peaking at ~1h and declining more rapidly thereafter, whereas levels of TNF mRNA
peak at ~2h and decline more gradually over the next 2-10h (13,43,44). Once translated, TTP
protein may localize to the nucleus (15,45,46), the cytoplasm (47,48) or both (48,49).
Leptomycin causes TTP to accumulate in the nucleus, proving that TTP shuttles. Induced TTP
protein levels are down regulated within 8h in human monocytes. The profiles vary with other
cell types and stimuli (20), suggesting that down regulation is regulated separately from
induction.
The established effector function of TTP is to negatively regulate mRNA stability. TTP
null mice have increased TNF mRNA and plasma TNF, and TNF mRNA is more stable in TTP
(-,-) cells than in WT cells. TTP additions to a cell-free system destabilize TNF mRNA by
stimulating deadenylation (50). Interestingly, endogenous TTP sediments with polysomes in
sucrose gradients (47), suggesting an additional unsuspected role for TTP in translation, and
possibly a positive one since polysomes usually contain the pool of actively translating mRNA.
Confocal microscopy data of endogenous TTP stained with a C-terminal antibody (47) suggest to
us a cytoplasmic localization of TTP into macromolecular assemblies. Endogenous TTP is
concentrated within unidentified, punctate substructures with more distinct borders in resting
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
24
cells and within more diffuse and more numerous substructures in stimulated cells. These
substructures are reminiscent of stress granules (51).
TTP phosphorylation could regulate its life cycle as well as its effector functions in
nuclear or cytoplasmic compartments. The sites are localized schematically in Fig. 7. When
TTP is nuclear in fibroblasts, its export to the cytosol can be stimulated by mitogens (15,46),
suggesting involvement of kinase cascades activated by mitogens.
TTP has an active nuclear export sequence (NES) within the first 14 amino acids at the
N-terminus (48,49). A sequence required for nuclear localization was mapped to sequence
surrounding the first Cys3His RNA binding domain, residues 95-130 in muTTP (48,49). Minor
in vitro phosphosite Ser105 lies within the NLS, and could silence NLS function. Major site
Ser52 and minor site Ser58 are located between the NES and NLS in the N-terminus as are
S80/82/85, making them candidates to regulate cellular localization and/or effector functions.
The N-terminus along with tandem Cys3His RNA binding domains, but not the C-
terminus, are required for induction of apoptosis caused by expression of TTP (34). The N-
terminus is not required to bind mRNA or to destabilize mRNA either in cells or in cell free
assays (50). Thus the requirement for the N-terminal domain could be confined to regulation of
import/export, but the N-terminus’s role might also include modulation of effector functions.
These may be regulated by N-terminal phosphorylations we describe. The N-terminus (aa 1-100
in muTTP) fused to GAL4 DNA has apparent transactivating activity in PC12 cells, which is
inhibitable by phorbol ester (42). Observed transactivating functions of TTP in macrophages are
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
25
also altered directly or indirectly by p38 kinase cascades. Expression of TTP robustly down
regulates basal luciferase activity from reporters of cytokine genes TNF and IL8, and it decreases
relative induction of the same reporters that, in the absence of TTP expression, are markedly
activated by LPS (13). Stimulation of p38α,β with LPS blocks suppression, and suppression is
blocked by p38α,β inhibition.
Major phosphosite S178 is in the C-terminal domain. Phosphorylation of S178 by MK2
may increase TTP levels in the cytoplasmic compartment following activation of the p38
pathway. S178 is not required for mitogen-stimulated export of TTP from the nucleus but is
required for 14-3-3 binding. Co-expression of TTP with 14-3-3β increased the level of
cytoplasmic TTP (15). Ser178 is not required for TTP effector function, assayed by ability of
S178A TTP to induce apoptosis.
Currently we do not know whether phosphorylation impacts the effector activity of TTP
as opposed to its life cycle, but phosphorylation could change protein/protein interactions that
facilitate deadenylation activity. Multienzyme complexes coordinate what happens at the 5’ and
3’ ends of mRNA (52). MK2 has also been shown to phosphorylate poly(A) binding protein 1
(53) and hnRNP AO (54). If TTP has a different function that favors, or is permissive, for
translation of cytokine, this would have to have a biologic switch. The major block in TNF
production in MK2 (-,-) cells is translation not stability (4). In TTP (-,-) mice, TNF is
overproduced. Thus as a hypothesis, phosphorylation of TTP at an N-terminal site (S52, S58) or
Ser105 in the nucleus could silence the NLS and/or enhance the NES activity, causing export of
TTP. Phosphorylation of S178 could increase its concentration or dwell time or bring another
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
26
protein to it via the other pocket of bound bifunctional 14-3-3, provided its other pocket does not
bind internally to another phosphosite in TTP. Loss of TTP does not block translation. If MK2
regulates translation in part by TTP phosphorylation, TTP should be a repressor of translation
when dephosphorylated and an activator (or neutral to) translation when phosphorylated.
Acknowledgements
We thank Mary Harp for technical assistance and many colleagues (named herein) who have
generously provided reagents. This work was supported by the United States Public Health
Service, National Institutes of Health Grants GM62890 (to T.W.S.), GM37537 (to D.F.H.), and
and DK/GM60720 (to M.T.W.). Support for C.A.C. was provided by Training Grant T32-
DK07320 (to C.A.C.)
References 1. Shi, Y., and Gaestel, M. (2002) Biological Chemistry 383(10), 1519-36 2. Kotlyarov, A., Neininger, A., Schubert, C., Eckert, R., Birchmeier, C., Volk, H. D., and
Gaestel, M. (1999) Nat Cell Biol 1(2), 94-7. 3. Han, J., Brown, T., and Beutler, B. (1990) J Exp Med 171(2), 465-75 4. Neininger, A., Kontoyiannis, D., Kotlyarov, A., Winzen, R., Eckert, R., Volk, H. D.,
Holtmann, H., Kollias, G., and Gaestel, M. (2002) J Biol Chem 277(5), 3065-8. 5. Lasa, M., Mahtani, K. R., Finch, A., Brewer, G., Saklatvala, J., and Clark, A. R. (2000)
Molecular & Cellular Biology 20(12), 4265-74 6. Kotlyarov, A., and Gaestel, M. (2002) Biochem Soc Trans 30(Pt 6), 959-63. 7. Mahtani, K. R., Brook, M., Dean, J. L., Sully, G., Saklatvala, J., and Clark, A. R. (2001)
Mol Cell Biol 21(19), 6461-9. 8. Blackshear, P. J. (2002) Biochemical Society Transactions 30(Pt 6), 945-52 9. Worthington, M. T., Amann, B. T., Nathans, D., and Berg, J. M. (1996) Proc Natl Acad
Sci U S A 93(24), 13754-9. 10. Amann, B. T., Worthington, M. T., and Berg, J. M. (2003) Biochemistry 42(1), 217-21. 11. Lai, W. S., Carballo, E., Thorn, J. M., Kennington, E. A., and Blackshear, P. J. (2000) J
Biol Chem 275(23), 17827-37. 12. Taylor, G. A., Carballo, E., Lee, D. M., Lai, W. S., Thompson, M. J., Patel, D. D.,
Schenkman, D. I., Gilkeson, G. S., Broxmeyer, H. E., Haynes, B. F., and Blackshear, P. J. (1996) Immunity 4(5), 445-54.
13. Zhu, W., Brauchle, M. A., Di Padova, F., Gram, H., New, L., Ono, K., Downey, J. S., and Han, J. (2001) Am J Physiol Lung Cell Mol Physiol 281(2), L499-508.
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
27
14. Taylor, G. A., Thompson, M. J., Lai, W. S., and Blackshear, P. J. (1995) J Biol Chem 270(22), 13341-7.
15. Johnson, B. A., Stehn, J. R., Yaffe, M. B., and Blackwell, T. K. (2002) J Biol Chem 277(20), 18029-36.
16. Stoecklin, G., Ming, X. F., Looser, R., and Moroni, C. (2000) Mol Cell Biol 20(11), 3753-63.
17. Khokhlatchev, A., Xu, S., English, J., Wu, P., Schaefer, E., and Cobb, M. H. (1997) J Biol Chem 272(17), 11057-62.
18. Ben-Levy, R., Hooper, S., Wilson, R., Paterson, H. F., and Marshall, C. J. (1998) Curr Biol 8(19), 1049-57.
19. Engel, K., Schultz, H., Martin, F., Kotlyarov, A., Plath, K., Hahn, M., Heinemann, U., and Gaestel, M. (1995) J Biol Chem 270(45), 27213-21.
20. Fairhurst, A.-M., Connolly, J., Hintz, K., Goulding, N., Rassias, A., Yeager, M., Rigby, W., and Wallace, P. (2003) Arthritis Res Ther 5(4), R214 - R225
21. Chrestensen, C. A., and Sturgill, T. W. (2002) J Biol Chem 277(31), 27733-41. 22. Worthington, M. T., Pelo, J. W., Sachedina, M. A., Applegate, J. L., Arseneau, K. O., and
Pizarro, T. T. (2002) J Biol Chem 277(50), 48558-64. 23. Poteet-Smith, C. E., Smith, J. A., Lannigan, D. A., Freed, T. A., and Sturgill, T. W.
(1999) J Biol Chem 274(32), 22135-8 24. Ficarro, S. B., McCleland, M. L., Stukenberg, P. T., Burke, D. J., Ross, M. M.,
Shabanowitz, J., Hunt, D. F., and White, F. M. (2002) Nature Biotechnology 20(3), 301-5 25. Eng, J., McCormack, A.L., and Yates, J.R. (1994) J Am. Soc. Mass Spectrom. 5, 976-989 26. Ben-Levy, R., Leighton, I. A., Doza, Y. N., Attwood, P., Morrice, N., Marshall, C. J., and
Cohen, P. (1995) Embo J 14(23), 5920-30. 27. Meng, W., Swenson, L. L., Fitzgibbon, M. J., Hayakawa, K., Ter Haar, E., Behrens, A.
E., Fulghum, J. R., and Lippke, J. A. (2002) J Biol Chem 277(40), 37401-5. 28. Shevchenko, A., Wilm, M., Vorm, O., and Mann, M. (1996) Analytical Chemistry 68(5),
850-8 29. Stokoe, D., Caudwell, B., Cohen, P. T., and Cohen, P. (1993) Biochem J 296(Pt 3), 843-
9. 30. Singh, S., Powell, D. W., Rane, M. J., Millard, T. H., Trent, J. O., Pierce, W. M., Klein, J.
B., Machesky, L. M., and McLeish, K. R. (2003) J Biol Chem. 278(38), 36410-7 31. Cheng, T. J., Tseng, Y. F., Chang, W. M., Chang, M. D., and Lai, Y. K. (2003) J Cell
Biochem 89(3), 589-602. 32. Cao, H., Dzineku, F., and Blackshear, P. J. (2003) Arch Biochem Biophys 412(1), 106-
120. 33. Johnson, B. A., Geha, M., and Blackwell, T. K. (2000) Oncogene 19(13), 1657-64. 34. Johnson, B. A., and Blackwell, T. K. (2002) Oncogene 21(27), 4237-46. 35. Cano, E., Doza, Y. N., Ben-Levy, R., Cohen, P., and Mahadevan, L. C. (1996) Oncogene
12(4), 805-12. 36. Kotlyarov, A., Yannoni, Y., Fritz, S., Laass, K., Telliez, J. B., Pitman, D., Lin, L. L., and
Gaestel, M. (2002) Mol Cell Biol 22(13), 4827-35. 37. Li, Y., Inoki, K., Vacratsis, P., and Guan, K. L. (2003) J Biol Chem 278(16), 13663-71. 38. Powell, D. W., Rane, M. J., Joughin, B. A., Kalmukova, R., Hong, J.-H., Tidor, B., Dean,
W. L., Pierce, W. M., Klein, J. B., Yaffe, M. B., and McLeish, K. R. (2003) Mol. Cell. Biol. 23(15), 5376-5387
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
28
39. Engel, K., Kotlyarov, A., and Gaestel, M. (1998) Embo J 17(12), 3363-71. 40. Terp, K. (2003) Applied Microbiology and Biotechnology 60(5), 523-533 41. Bosworth, H. B., Clipp, E. C., McNeilly, M., Christakis, N. A., Parker, J., Tulsky, J. A.,
and Parker, C. E. (2002) Journal of Palliative Medicine 5(6), 829-41 42. Murata, T., Hikita, K., and Kaneda, N. (2000) Biochem Biophys Res Commun 274(2),
526-32. 43. Carballo, E., Lai, W. S., and Blackshear, P. J. (1998) Science 281(5379), 1001-5. 44. Baseggio, L., Charlot, C., Bienvenu, J., Felman, P., and Salles, G. (2002) Eur Cytokine
Netw 13(1), 92-8. 45. DuBois, R. N., McLane, M. W., Ryder, K., Lau, L. F., and Nathans, D. (1990) J Biol
Chem 265(31), 19185-91. 46. Taylor, G. A., Thompson, M. J., Lai, W. S., and Blackshear, P. J. (1996) Mol Endocrinol
10(2), 140-6. 47. Brooks, S. A., Connolly, J. E., Diegel, R. J., Fava, R. A., and Rigby, W. F. (2002)
Arthritis Rheum 46(5), 1362-70. 48. Phillips, R. S., Ramos, S. B., and Blackshear, P. J. (2002) J Biol Chem 277(13), 11606-
13. 49. Murata, T., Yoshino, Y., Morita, N., and Kaneda, N. (2002) Biochem Biophys Res
Commun 293(4), 1242-7. 50. Lai, W. S., Kennington, E. A., and Blackshear, P. J. (2003) Molecular & Cellular
Biology 23(11), 3798-812 51. Kedersha, N., and Anderson, P. (2002) Biochemical Society Transactions 30(Pt 6), 963-9 52. Wilkie, G. S., Dickson, K. S., and Gray, N. K. (2003) Trends Biochem Sci 28(4), 182-8 53. Bollig, F., Winzen, R., Gaestel, M., Kostka, S., Resch, K., and Holtmann, H. (2003)
Biochem Biophys Res Commun 301(3), 665-70. 54. Rousseau, S., Morrice, N., Peggie, M., Campbell, D. G., Gaestel, M., and Cohen, P.
(2002) Embo J 21(23), 6505-14
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
29
Figure Legends
Fig. 1. Activation of MK2 in bacteria requires co-expression with active p38α. A)
Purification of GST-MK2∆3B after co-expression with activated ERK2 or p38 shows that only
p38 interacts with and activates MK2. Panel 1 - SDS-PAGE analysis of GST-MK2∆3B
expressed alone. Panel 2- SDS-PAGE analysis of GST-MK2∆3B co-expressed with active
ERK2. Panel 3- Western analysis of GST-MK2∆3B co-expressed with active ERK2 probed for
anti-ERK2. Panel 4- Panel 3 probed for anti-active ERK. Panel 5-SDS-PAGE analysis of GST-
MK2∆3B expressed with active p38α. Panel 6- Western analysis of GST-MK2∆3B co-
expressed with active p38 probed for anti-p38. An electrostatic interaction, disrupted by 2 M
LiCl, is observed between MK2 and p38 but not between MK2 and ERK2. B) SDS-PAGE
analysis of GST-MK2∆3B (GST-MK2EE∆3B) purified either alone or with activated ERK2 or
p38, only MK2 expressed with p38 is gel shifted. C) Kinase activity of purified GST-MK2∆3B
proteins using the p81 paper assay with a peptide substrate (KKLNRTLSVA). MK2 activated
by p38 has a much greater specific activity in comparison to the mutational activated protein
MK2EE∆3B. Results are representative of at least 2 independent experiments.
Fig. 2. Activated MK2 phosphorylates TTP in vitro. A) Kinase assays were analyzed by SDS-
PAGE, and incorporation of 32P into the substrates was visualized using a phosphoimager.
Active GST-MK2∆3B phosphorylates TTP and HSP27. The lower panel is the Coomassie stain
of these gels. B) SDS-PAGE analysis (4-20% gradient) of concentrated GST-TTP (6µg)
phosphorylated with activated GST-MK2∆3B (0.05µg), silver stain (left panel) and
PhosphoImage (right panel). Results are indicative of more than 3 independent experiments.
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
30
Fig. 3. MS/MS analyses of MK2 phosphorylated TTP. A) Reverse phase and IMAC MS/MS
analysis of TTP peptides. TTP was treated as described in Fig. 2 and then the protein was re-
purified with GSH-Sepharose and digested with trypsin (1% of the digest analyzed for reverse
phase, 10% for IMAC).
Fig. 4. GST-TTP phosphorylated in vitro by MK2 interacts with endogenous 14-3-3. A)
GST-TTP (5µg) was incubated in various kinase reactions with or without activated GST-
MK2∆3B, or p38 (0.05µg) and then the reaction was combined with BHK lysates transfected
with 14-3-3. After incubation, the GST-TTP was bound to GSH-Sepharose beads, which were
washed and then analyzed by SDS-PAGE for interacting proteins. B) GST-TTP was treated as
in A but pre-treated with or without SB203580 (10µM), and the BHK cells used to make the
lysates were not transfected with 14-3-3, so only endogenous protein is detected. Results are
representative of 4 independent experiments
Fig. 5. Myc-TTP transfected into BHK cells is predominantly phosphorylated. BHK cells
transfected with myc-TTP were treated with or without SB203580 (10µM) 60 min. prior to
harvest. Anti-TTP, and Anti-myc antibodies were used on duplicate Western blots to gauge the
effect of p38 pathway inhibition on protein mobility. Results are representative of at least 4
independent experiments.
Fig. 6. In vivo analysis of TTP phosphorylation sites and modulation by MK2. A) Change in
relative TTP phosphorylation by treatment with anisomycin; 3T3 cells were transiently
transfected with expression vectors containing TTP and then treated with 5 µg/ml anisomycin for
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
31
20 minutes prior to harvest. Dynabead-purified proteins were digested into tryptic peptides,
which were analyzed in data-dependant mode (4% of digest). The results are presented as % of
peptide phosphorylated in treated and untreated sample. Results are the average ± range of two
MS/MS analyses. B) Lysates from cells co-transfected with TTP and either WT or KD MK2
were treated as in A, results of three independent runs averaged ± S.D. *- in targeted analysis on
samples from B (see part C), this was determined to be a co-eluting peak of three singly
phosphorylated species (S80/82/85) C) Elution order and abundance of eight TTP peptides
analyzed using targeted analysis of sample from B. D) Western blot analysis of cellular lysates
and an IP from 3T3 cells. E) Western blot analysis (anti-myc) of lysates from WT and
S52A/S178A mutant transfected BHK cells treated with or without calf intestine alkaline
phosphatase. The S80/82/85 peptide is ***PGPELSPSPTSPTATPTTSSR.
Fig. 7. Identified Sites in muTTP. NES (Red) is present in the first 14 residues. A segment
required for NLS activity includes a loop of the first Cys3His domain (Blue). MK2 sites S52 and
S178 were sensitive to anisomycin and MK2 expression. S80/82/85 were anisomycin sensitive
and MK2 expression insensitive, and are putative p38 sites.
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
32
Table I - Summary of in vitro MK2 phosphorylation sites on TTP.
Phosphopeptide
Signal for Nonphos C18a
Signal for Phospho C18a
Percent Phosphorylatedb
Signal IMAC a
175QSIS*FSGLPSGR186
175QS*IS*FSGLPSGR186 1.3x108 1.3x108
1.4x108 53 1.2x108 3.3x106
50STS*LVEGR57 1.1x108 2.2x108 67 3.6x108 103TYS*ESGR109 4.6x107 8.2x106 15 Not
detected 308RLPIFNRIS*VSE319 9.0x106 3.5x106 28 2.9x107
58S*CGWVPPPPGFAPLAPR74 1.8x108 3.6x106 248ST?T?PSTIVGPLGGLAR263 1.5x108 2.3x106
264S?PS?AHSLGSDPDDYASSGSSL.. .GGSDSPVFEAGVFGPPQTPAPPR307
2.8x107 1.4x107
a) The signals for the non- and phosphorylated peptides are the sum of peak intensities for all detected charge states.
b) The percent phosphorylation is an estimate based on the ratio of the phosphorylated signal to the total signal.
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
0%
10%
20%
30%
40%
50%
60%
Ser 52 Ser 178 Ser 80,82,85* Thr 250
Peptide
Perc
ent p
hosp
hory
latio
n
UntreatedAnisomycin
Fig. 6, Chrestensen et al.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Ser 52 Ser 178 Ser 80,82,85* Thr 250
Peptide
Perc
ent p
hosp
hory
latio
n
Wild typeKinase dead
B
A
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
Fig. 7, Chrestensen et al.
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from
W. Pelo, Mark T. Worthington and Thomas W. SturgillCarol A. Chrestensen, Melanie J. Schroeder, Jeffrey Shabanowitz, Donald F. Hunt, Jared
14-3-3 bindingMK2 phosphorylates tristetraprolin on in vivo sites including S178, a site required for
published online December 19, 2003J. Biol. Chem.
10.1074/jbc.M310486200Access the most updated version of this article at doi:
Alerts:
When a correction for this article is posted•
When this article is cited•
to choose from all of JBC's e-mail alertsClick here
by guest on February 18, 2018http://w
ww
.jbc.org/D
ownloaded from