Cellular Signalling
Cellular Signalling
Cellular Signalling
Cellular Signalling
journal homepage: www.elsevier.com/locate/cellsig
The LMP2A protein of EpsteinBarr virus regulates phosphorylation of ITSN1 and Shb
adaptors by tyrosine kinases
Oleksandr Dergai a, Mykola Dergai a, 1, Inessa Skrypkina a, 1, Liudmila Matskova b, Liudmyla Tsyba a,
Daria Gudkova a, Alla Rynditch a,
a
b
State Key Laboratory of Molecular and Cellular Biology, Institute of Molecular Biology and Genetics, 150 Zabolotnogo Street, Kyiv 03680, Ukraine
Microbiology and Tumor Biology Center (MTC), Karolinska Institute, 171 77 Stockholm, Sweden
a r t i c l e
i n f o
Article history:
Received 24 July 2012
Accepted 4 September 2012
Available online 10 September 2012
Keywords:
EpsteinBarr virus
LMP2A
Intersectin
Shb
Phosphorylation
Syk
a b s t r a c t
Latent Membrane Protein 2A (LMP2A) is an EpsteinBarr virus-encoded protein that is important for the
maintenance of latent infection. Its activity affects cellular differentiation, migration, proliferation and B
cell survival. LMP2A resembles a constitutively activated B cell antigen receptor and exploits host kinases
to activate a set of downstream signaling pathways. In the current study we demonstrate the interaction of
LMP2A with intersectin 1 (ITSN1), a key endocytic adaptor protein. This interaction occurs via both the Nand C-tails of LMP2A and is mediated by the SH3 domains of ITSN1. Additionally, we identied the Shb adaptor and the Syk kinase as novel binding ligands of ITSN1. The Shb adaptor interacts simultaneously with the
phosphorylated tyrosines of LMP2A and the SH3 domains of ITSN1 and mediates indirect interaction of ITSN1
to LMP2A. Syk kinase promotes phosphorylation of both ITSN1 and Shb adaptors in LMP2A-expressing cells.
In contrast to ITSN1, Shb phosphorylation depends additionally on Lyn kinase activity.
Considering that Shb and ITSN1 are implicated in various receptor tyrosine kinase signaling, our results indicate that LMP2A can affect a number of signaling pathways by regulating the phosphorylation of the ITSN1
and Shb adaptors.
2012 Elsevier Inc. All rights reserved.
1. Introduction
The EpsteinBarr virus (EBV) is a ubiquitous virus that infects only
humans. It is a -herpesvirus that primarily infects cells of the oropharyngeal epithelium and B cells, but may, in rare cases, also infect
natural killer cells, smooth muscle cells and peripheral blood T cells
[1]. Due to its association with post-transplant lymphomas, Burkitt's
lymphoma, Hodgkin's lymphoma and nasopharyngeal carcinoma,
EBV is classied as an oncogenic virus [1,2]. While in the majority of
the human population EBV infection occurs symptomless before the
age of 5 years, primary EBV infection late in adolescence frequently
causes a benign self-limiting lymphoproliferative disease, known as
infectious mononucleosis (IM) [3]. IM confers an elevated risk for
Hodgkin's lymphoma later in life [4]. During its life cycle EBV can establish four latency programs with varying sets of expressed genes
[1]. Latent Membrane Protein 2A (LMP2A) is a continuously detected
viral message both in healthy and diseased individuals, which necessitates the investigation of its cellular interactome [5].
LMP2A contains a 119 amino acid-long intracellular N-terminal
tail followed by 12 transmembrane segments and a C-terminal tail
of 27 amino acids [6]. The N-terminal tyrosine-based motifs are phosphorylated by the host kinases providing docking sites for SH2- and
PTB-containing proteins such as Shb, Lyn, Syk, etc. [7,8]. Like B cell receptor (BCR), LMP2A recruits and activates the Syk tyrosine kinase at
the Immunoreceptor Tyrosine-based Activation Motif (ITAM) (Y74
and Y85) allowing LMP2A to generate its own mitogenic signaling
[9,10]. LMP2A-dependent activation of Syk results in AKT activation
leading to tonic prosurvival signaling and prevention of apoptosis in
B cells [7,11,12]. In epithelial cells, LMP2A activates an array of signaling pathways such as the ERK/MAPK, PI3K/AKT and -catenin/Wnt
pathways [1316]. Initial studies of LMP2A signaling showed a suppressive effect on B cell receptor signaling [9,17]. Subsequent studies
have revealed that LMP2A, in addition, may replace constitutive BCR
signaling, allowing B cells with crippled BCR to escape cell death.
Thus, LMP2A may both mimic and inhibit active BCR signaling
[1820].
The LMP2A cytosolic tails contain multiple PXXP (Xany amino acid)
motifs (PRMs) providing docking sites for WW (named for a conserved
Trp-Trp motif) and SH3 (Src homology 3) domains [21]. Except for the
association of WW domain-containing ubiquitin ligases of the AIP4/
NEDD4 family [22,23], which bind to PPPPY motifs of LMP2A, no direct
34
Full-length Shb was amplied with the HA-tag or the Myc-tag in the reverse primer and cloned into pcDNA4/HisMax C to obtain Omni-Shb-HA
or Omni-Shb-Myc, respectively. HA-Syk, HA-Syk K396R, HA-Syk46/201
were kindly provided by Dr. Chung-Wai Chow (Toronto, Canada) [30].
Myc-AIP4 was a kind gift of Dr. Tony Pawson's laboratory. The HA-Lyn
construct was kindly provided by Dr. Hugh R. B. Pelham (Cambridge,
UK) [31].
2.3. Plasmids
3. Results
Full-length 4xFLAG-tagged LMP2A, LMP2A Y74-85 F and Y112F
were described previously [7]. LMP2A wild type (aa 1497, Genbank
accession number YP_401631.1), LMP2A NT (aa 105497) and all
derived LMP2A mutants with P-to-A substitutions were generated
by PCR from constructs carrying wild type LMP2A with subsequent
cloning in the pcDNA4/HisMax C vector (Invitrogen) in frame with
the Omni-tag. Site-specic mutations were introduced by primer extension. The P1 + 4 m LMP2A mutant was obtained by digesting
the 4 m LMP2A (a variant of LMP2A with four motifs mutated: P2,
P3, P4, P5) coding sequence with BamHI and subcloning into
pcDNA4/HisMax B. Plasmids encoding the GST-SH3 domains of
ITSN1 and Omni-ITSN1 were described previously [29]. GST-ITSN1SH3(AE) and GST-NT-LMP2A were obtained by cloning cDNA fragments encoding all ve SH3 domains (aa 7391223, accession number
NP_001001132.1) and the 119 amino acids of the N-terminus respectively in frame with GST into the pGEX 4 T-3 vector (Amersham).
3.1. LMP2A interacts with the endocytic scaffold ITSN1 in different cell
lines
To examine whether LMP2A interacts with ITSN1 immunoprecipitation experiments were performed with endogenous ITSN1 and
LMP2A that was transiently expressed in HEK293 or virally encoded
in CBMI-Ral-Sto cells. LMP2A was detected in immunoprecipitates
of endogenous ITSN1 from HEK293 cells (Fig. 1A, left panel); likewise
ITSN1 was detected in anti-LMP2A precipitates from HEK293 cells
(Fig. 1A, right panel), CBMI-Ral-Sto cells (Fig. 1B) and MCF-7 cells
(data not shown). A partial colocalization of Omni-LMP2A and endogenous ITSN1 was observed in HEK293 cells (Fig. 1E). Overlapping signals of these proteins were found in the perinuclear zone and
throughout the cytoplasm. These data evidence for the complex formation of ITSN1 with LMP2A in vivo.
35
Fig. 1. ITSN1 interacts with LMP2A in vitro and in vivo. (A) ITSN1 was immunoprecipitated with -EH2 antibodies (left panel) from HEK293 cells transfected with
FLAG-LMP2A. Immunoprecipitated proteins were detected with -ITSN1 and with -FLAG antibodies. ITSN1 was reciprocally immunoprecipitated with -LMP2A antibodies (right panel). (B) Co-immunoprecipitation of ITSN1 with -LMP2A antibodies using extracts from CBMI-Ral-Sto (EBV positive) and RAMOS NUT (EBV negative)
cell lines. (C) GST-fusion protein containing all 5 SH3 domains of ITSN1 (GST-ITSN1-SH3(AE)) or GST alone were immobilized on glutathione Sepharose and incubated
with lysates of HEK293 cells transfected with FLAG-LMP2A. GST-fusion proteins (lower panel) were visualized with Ponceau S staining; LMP2A (upper panel) was
immunodetected with -FLAG antibodies. TCL: total cell lysate, WB: Western blotting. (D) Separate GST-fusion SH3 domains of ITSN1 or GST alone were used as bait to
precipitate Omni-LMP2A from HEK293 cells. Fusion proteins (lower panel) were stained with Coomassie Blue, and LMP2A was detected with -Omni-antibodies
(upper panel). (E) LMP2A and ITSN1 are co-distributed in HEK293 cells. Cells were transfected with Omni-LMP2A and immunostained with -Omni antibodies followed
by visualization with Alexa Fluor 633 conjugated secondary antibodies; endogenous ITSN1 was detected with -EH2 antibodies and visualized with FITC-conjugated secondary antibodies. Regions of colocalization were visualized with ImageJ software. Scale bar represents 10 m.
3.2. ITSN1 associates with both tails of LMP2A via its SH3 domains
Both cytosolic tails of LMP2A contain PRMs representing canonical
binding sites for WW and SH3 domains [21]. As ITSN1 contains ve
SH3 domains (referred here as ITSN1-SH3(AE)) it is able to bind numerous PRMs. We have shown that ITSN1-SH3(AE) can precipitate
Omni-LMP2A from HEK293 cells (Fig. 1C). A strong interaction of
LMP2A was observed with the SH3D domain and a much weaker interaction with the SH3A domain (Fig. 1D).
Next, we examined which PRM in LMP2A is responsible for the interaction with ITSN1. Five putative PRMs with the consensus PXXP
are present in the LMP2A cytosolic domains (Fig. 2A). First, we found
that an LMP2A mutant lacking the N-terminus with the rst four PRMs
was still capable of binding the ITSN1-SH3(AE) protein (Fig. 2B, lane
4). This observation implied a role for the fth, C-terminal, PRM motif,
but substitution of prolines for alanines (P-to-A) in this motif did not
abolish the interaction with ITSN1-SH3(AE) protein (Fig. 2B, lane 9).
Next, a set of LMP2A mutants with substitutions of P-to-A were generated (Fig. 2A) and tested for their ability to bind ITSN1-SH3(AE).
36
Fig. 2. ITSN1 interacts with both the N- and C-terminal tails of LMP2A in vitro. (A) Schematic representation of LMP2A and its mutants used here. Proline-rich motifs are designated
P1 to P5, gray boxes are transmembrane segments, amino acid sequences of PRMs are presented in white rectangles. P-to-A substitutions and Y-to-F are highlighted in bold.
(B) GST-ITSN1-SH3(AE) was used as bait to precipitate full-length Omni-LMP2A and its truncated or point mutants (designated by the change of P to A or Y to F). GST-fusion proteins (two lower panels) were visualized by Ponceau S staining. Omni-LMP2A (upper panel) was immunodetected with -Omni antibodies.
3.3. Shb scaffold mediates the interaction between ITSN1 and the Nterminal tail of LMP2A
The phosphorylation of Y74 and Y85 in the LMP2A ITAM thus
appeared essential for the interaction with ITSN1 (Fig. 2B, lane 23).
However, since the SH3 domains of ITSN1 per se were not shown to
bind phosphotyrosine motifs we hypothesized that an additional
phosphotyrosine-binding adaptor, such as Shb or Syk, might be involved in the recruitment of ITSN1 to the N-terminal domain of
LMP2A. Both bind phosphotyrosine residues in the ITAM motif and
contain PRMs that interact with SH3 domains [7,8]. Shb is a ubiquitously expressed adaptor protein [34]. It contains a number of
N-terminal PRMs that mediate its interaction with the SH3 domains of
Eps8, the p85 subunit of PI3-kinase, Grb2 and Src [35], a PTB domain
that binds phosphorylated p36/38 (LAT) [36] and an SH2 domain,
which interacts with PDGF-receptor, FGFR1 and the T cell receptor associated -chain [34,35,37]. We conrmed the interaction in vivo between endogenous Shb and ITSN1 in HEK293 cells (Fig. 3A) as well as
the in vitro complex formation of Omni-Shb with the SH3A, B, D and E
domains of ITSN1 (Fig. 3B and C). Overexpression of Shb increased the
efciency of ITSN1 and LMP2A association (Fig. 3D), further arguing
for a role of Shb as a scaffold for the interaction between ITSN1 and
LMP2A.
We examined the possible role of the Syk kinase as a putative accessory protein in ITSN1/LMP2A complex as well. Neither overexpression
of HA-Syk of wild type, nor the catalytically defective (K396R) mutant
or the double mutant in both SH2 domains (46/201, decient in binding
3.4. LMP2A induces tyrosine phosphorylation of the Shb and ITSN1 adaptor proteins
Previously Shb was shown to be tyrosine phosphorylated and activated downstream of surface receptor ligand stimulation [38]. We
investigated whether LMP2A expression would stimulate tyrosine
phosphorylation of Shb. As shown in Fig. 4A (lanes 1 and 5) there
was a dramatic increase in the phosphorylation of Shb in LMP2Aexpressing cells compared to control cells. The Y74/85 F LMP2A mutant, that is unable to bind Syk and is signicantly impaired in binding to Shb, had a moderate effect on Shb tyrosine phosphorylation
(Fig. 4A, lane 6). The Y112F LMP2A (Fig. 4A, lane 7), which is defective for association with Lyn kinase, showed decreased tyrosine
phosphorylation of Shb in comparison to wild type LMP2A (Fig. 4A,
lane 5). This suggests that the LMP2A-associated tyrosine kinases
might mediate phosphorylation of Shb. Indeed, coexpression of wild
type LMP2A with Syk (Fig. 4A, lane 8) or Lyn (Fig. 4B, lane 3) had a dramatic cooperative effect on the level of tyrosine phosphorylated Shb.
Coexpression of LMP2A with catalytically defective Syk K396R
(Fig. 4A, lane 9) had no effect while expression of Lyn, accompanied
37
Fig. 3. The adaptor protein Shb mediates the interaction between LMP2A and ITSN1. (A) Endogenous ITSN1 was co-immunoprecipitated with -Shb serum from HEK293 cells lysates. ITSN1 and Shb were detected with -ITSN1 and -Shb polyclonal antibodies, respectively. B, C. GST-ITSN1-SH3(AE) (B) or separate SH3 domains (C) were used as bait to
precipitate full-length Omni-Shb-HA. (D) HEK293 cells were co-transfected with FLAG-LMP2A and Omni-Shb-HA or GFP. -ITSN1 immunoprecipitates were immunoblotted with
-FLAG antibodies. The lter was stripped and reprobed with -Omni antibodies. (E) Lysates of HEK293 cells co-transfected with Omni-ITSN1, FLAG-LMP2A, HA-Syk or the indicated Syk mutants were subjected to immunoprecipitation with -Omni antibodies. The precipitated proteins were analyzed by Western blotting with -EH2, -FLAG or -HA
antibodies.
by treatment with the Src kinase-specic inhibitor PP1 (Fig. 4B, lane 4),
resulted in signicantly decreased Shb phosphorylation.
We suggested that Syk and Lyn kinases might mediate the phosphorylation of Shb in the absence of LMP2A. Fig. 4A (lanes 2 and 4)
shows that overexpression of Syk (lane 2) as well as Lyn (lane 4)
led to a signicant increase in the amount of tyrosine phosphorylated
Shb. Indeed, coexpression of Shb with Syk K396R (Fig. 4A, lane 3) had
no impact on the Shb phosphorylation level (Fig. 4A, lane 1). Inhibition of Src kinases with PP1 while coexpressing Shb with Lyn
(Fig. 4B, lane 2) led to a reduction of Shb tyrosine phosphorylation
compared to the DMSO control (Fig. 4B, lane 3).
Further, we studied whether LMP2A affects phosphorylation of
ITSN1. We detected tyrosine phosphorylation of ITSN1 in immunoprecipitates of cells overexpressing LMP2A and Syk, but not Lyn and
LMP2A or Syk alone (Fig. 5). Thus, our results indicate that LMP2A
promotes tyrosine phosphorylation of Shb and ITSN1 adaptors via
its association with the tyrosine kinases Syk and Lyn.
4. Discussion
Here we report that LMP2A interacts with ITSN1, a key component
of the cellular endocytic machinery. ITSN1 is a multidomain adaptor
protein implicated in the regulation of clathrin-mediated endocytosis
as well as in cellular signaling [2628,39]. We conrmed the presence
of the ITSN1/LMP2A complex in vivo, by co-immunoprecipitation and
38
Fig. 4. LMP2A induces Syk- and Lyn-dependent phosphorylation of Shb in HEK293 cells. (A) Cells were co-transfected with Shb and LMP2A, Syk, its mutants and Lyn as indicated in
the table at the top of the gure. 24 h post-transfection, the cells were lysed and Shb was immunoprecipitated with -Omni antibodies. The samples were probed with -pTyr
antibodies (upper panel), stripped and reprobed with -Omni antibodies. p-Shb intensity was normalized to total Shb and plotted on the graph as the average mean SD of
three independent experiments. (B) Lyn-dependent phosphorylation of Shb is suppressed by PP1 Src kinase inhibitor. HEK293 cells were contransfected with Omni-Shb-Myc,
FLAG-LMP2A and HA-Lyn as indicated in the table at the top. 24 h post-transfection the cells were treated with 25 M PP1 for 6 h, and the p-Shb level was estimated as indicated
above.
Related to Shb, adaptor protein Shc is phosphorylated by Lynactivated Syk kinase during BCR signaling [52]. We observed that treatment with the Src kinase inhibitor PP1 led to a signicant reduction of
Shb phosphorylation in cells coexpressing LMP2A and Lyn. This demonstrates that Shb phosphorylation depends on the activity of Lyn that is
consistent with previous observations [47].
ITSN1 is tyrosine phosphorylated in HEK293 cells overexpressing
both Syk and LMP2A. Syk alone had moderate if any effect on ITSN1
phosphorylation. We suggest that Syk must be activated by LMP2A
to gain the potential for ITSN1 phosphorylation. Kinase-specic prediction of tyrosine phosphorylation sites with GPS 2.1 (http://gps.
biocuckoo.org/) has identied several potential sites for Syk/Zap70
in the C-terminal part of ITSN1. No Lyn-specic phosphorylation
was predicted, consistent with our observation that Lyn and LMP2A
coexpression has no effect on ITSN1 phosphorylation. Currently,
only reports from large-scale mass spectrometry experiments provide evidence that ITSN1 undergoes tyrosine phosphorylation. No
data about sites of this modication or its functional signicance are
available at present.
We conclude that Shb and ITSN1 bind each other independently of
post-translational modication, at least what concerns phosphorylation
of tyrosines. Syk and LMP2A coexpression dramatically increased phosphorylation of both Shb and ITSN1 but expression of Syk and its mutants
had no effect on ITSN1 interaction with LMP2A that was shown to be at
39
5. Conclusions
In this study, we have identied multiple proteinprotein interactions that take place in LMP2A-based signalosome. The key endocytic
adaptor protein ITSN1 associates directly and indirectly with LMP2A
and with LMP2A-signalosome components such as Shb and Syk. Tyrosine kinases associated with LMP2A drives phosphorylation of ITSN1
and Shb adaptors.
Supplementary data to this article can be found online at http://
dx.doi.org/10.1016/j.cellsig.2012.09.011.
Conict of interest
The authors declare no competing nancial interests.
Authorship
Contributions: O.D., M.D., I.S. and D.G. performed experiments, analyzed data, designed research, and wrote the manuscript; L.T., L.M.
and A.R. analyzed data and wrote the manuscript.
Acknowledgments
least partially mediated by Shb. This agrees with our observation that
ITSN1 forms a complex with Shb in HEK293 cells independently of mitogen stimulation (unpublished data). We conclude that the interaction
between Shb and ITSN1 primarily depends on the association of the
ITSN1 domains with N-terminal PxxP motifs in Shb.
Previously it was shown that clustering of LMP2 proteins controls
their constitutive phosphorylation and downstream signaling [7]. We
suggest that the ITSN1 interaction with the C-terminus of LMP2A and
Shb-dependent binding to the N-terminus of LMP2A may allow to
form of clusters of LMP2A-dimers into higher-order oligomers where
ITSN1 could function as a multivalent scaffold. A hypothetical model
of ITSN1/LMP2A interaction and LMP2A-induced phosphorylation of
the adaptor proteins is presented in Fig. 6.
The current observations about interaction between ITSN1 and
LMP2A are in line with our previous nding that endocytic adaptor
amphiphysin binds PRM of LMP2A via its SH3-domain and this is important for the secretion of LMP2A in exosomes [53]. Whether ITSN1
affects LMP2A trafcking and what other impacts ITSN1 could have
on the LMP2A signaling activity, remain challenges for future studies.
We thank Dr. G. Winberg, Prof. I. Erenberg and A.-L. Haenni for helpful discussions and comments on the manuscript. We are grateful to
Drs. V. Gorchev and S. Karakhim for providing expertise and help in
obtaining the confocal images. We thank Dr. M. Welsh for providing
-Shb antibodies and for fruitful discussions. This work was supported
by an INTAS grant (no. 05-1000004-7762). Skrypkina I. and Dergai O.
were supported by a Visby scholarship from the Swedish Institute.
References
[1] E. Kieff, A.B. Rickinson, Fields Virology, fth ed. Lippincott Williams & Wilkins,
2007..
[2] G. Niedobitek, N. Meru, H.J. Delecluse, Int J Exp Pathol 83 (2001) 149170.
[3] D.A. Thorley-Lawson, A. Gross, N Engl J Med 350 (2004) 13281337.
[4] K.T. Bieging, M. Swanson-Mungerson, A.C. Amick, R. Longnecker, Cell Cycle 9
(2010) 901908.
[5] D.A. Thorley-Lawson, Nat Rev Immunol 1 (2001) 7582.
[6] R. Longnecker, E. Kieff, J Virol 64 (1990) 23192326.
[7] L.V. Matskova, C. Helmstetter, R.J. Ingham, G. Gish, C.K. Lindholm, I. Ernberg, et al.,
Oncogene 26 (2007) 49084917.
[8] C.L. Miller, A.L. Burkhardt, J.H. Lee, B. Stealey, R. Longnecker, J.B. Bolen, et al., Immunity 2 (1995) 155166.
[9] S. Fruehling, R. Longnecker, Virology 235 (1997) 241251.
[10] M. Merchant, R.G. Caldwell, R. Longnecker, J Virol 74 (2000) 91159124.
[11] M. Fukuda, R. Longnecker, J Virol 78 (2004) 16971705.
[12] O. Hatton, L.K. Phillips, M. Vaysberg, J. Hurwich, S.M. Krams, O.M. Martinez, J Biol
Chem 286 (2011) 3736837378.
[13] S.Y. Chen, J. Lu, Y.C. Shih, C.H. Tsai, J Virol 76 (2002) 95569561.
[14] R. Swart, I.K. Ruf, J. Sample, R. Longnecker, J Virol 74 (2000) 1083810845.
[15] M. Fukuda, R. Longnecker, J Virol 81 (2007) 92999306.
[16] J.A. Morrison, N. Raab-Traub, J Virol 79 (2005) 23752382.
Fig. 6. A proposed model of LMP2A-induced phosphorylation of Shb and ITSN1 adaptors. The protein complex assembled by LMP2A is shown. ITSN1 directly associates with the
C-terminal domain of LMP2A via its SH3 domains; Shb binds phosphotyrosines of the LMP2A N-terminal tail. ITSN1 interacts with Shb through its SH3 domains and PRM of Shb.
In this way ITSN1 is recruited to LMP2A indirectly via Shb adaptor. The described interactions make possible an assembly of LMP2A dimers into high-order oligomers. The
LMP2A-associated kinases Lyn and Syk phosphorylate Shb, while only Syk phosphorylates ITSN1. It is not clear, whether Lyn directly phosphorylates Shb or whether its effect consists only in Syk activation.
40
[37] M. Welsh, Z. Songyang, J.D. Frantz, T. Trb, K.A. Reedquist, T. Karlsson, et al., Oncogene 16 (1998) 891901.
[38] K. Holmqvist, M.J. Cross, C. Rolny, R. Hgerkvist, N. Rahimi, T. Matsumoto, et al., J
Biol Chem 279 (2004) 2226722275.
[39] W.M. Henne, E. Boucrot, M. Meinecke, E. Evergren, Y. Vallis, R. Mittal, et al., Science 328 (2010) 12811284.
[40] D.T. Lynch, J.S. Zimmerman, D.T. Rowe, J Gen Virol 83 (2002) 10251035.
[41] C.W. Dawson, R.J. Port, L.S. Young, Semin Cancer Biol 22 (2012) 144153.
[42] M.F. Pang, K.W. Lin, S.C. Peh, Cell Mol Biol Lett 14 (2009) 222247.
[43] T. Portis, M. Ikeda, R. Longnecker, Trends Immunol 25 (2004) 422426.
[44] R. Hooshmand-Rad, L. Lu, C.H. Heldin, L. Claesson-Welsh, M. Welsh, Exp Cell Res
257 (2000) 245254.
[45] C.K. Lindholm, Biochem Biophys Res Commun 296 (2002) 929936.
[46] M.J. Cross, L. Lu, P. Magnusson, D. Nyqvist, K. Holmqvist, M. Welsh, et al., Mol Biol
Cell 8 (2002) 28812893.
[47] K. Holmqvist, M. Cross, D. Riley, M. Welsh, Cell Signal 15 (2003) 171179.
[48] K. Gustafsson, G. Calounova, F. Hjelm, V. Kriz, B. Heyman, K.O. Grnvik, et al., BMC
Immunol 12 (2011) 110.
[49] M. Rozakis-Adcock, R. Fernley, J. Wade, T. Pawson, D. Bowtell, Nature 363 (1993)
8385.
[50] N.W. Gale, S. Kaplan, E.J. Lowenstein, J. Schlessinger, D. Bar-Sagi, Nature 363
(1993) 8892.
[51] X.K. Tong, N.K. Hussain, E. de Heuvel, A. Kurakin, E. Abi-Jaoude, C.C. Quinn, et al.,
EMBO J 19 (2000) 12631271.
[52] K. Nagai, M. Takata, H. Yamamura, T. Kurosaki, J Biol Chem 270 (1995)
68246829.
[53] O. Dergai, M. Dergai, I. Skrypkina, L. Tsyba, A. Yaruchik, A. Rynditch, Biopolym Cell
28 (2012) 234238.