net/publication/324042520
CITATIONS READS
2 126
10 authors, including:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Yusuf Akhter on 22 September 2018.
Gayatri Jamwala,1, Gurjinder Singha,1, Mohd Saleem Dara,1, Paramjeet Singha, Nasima Banoa ,
Sajad Hussain Syeda, Padmani Sandhub, Yusuf Akhter b, Satdarshan P. Monga c and Mohd Jamal
Dara,*
a
Academy of Scientific and Innovative Research (AcSIR), Anusandhan Bhawan, New Delhi,
T
IP
India; Cancer Pharmacology Division, CSIR-Indian Institute of Integrative Medicine,
CR
Jammu-180001, Jammu & Kashmir, India.
b
Department of Biotechnology, Babasaheb Bhimrao Ambedkar University, Vidya Vihar,
US
Raebareli Road, Lucknow, Uttar Pradesh 226025, India.
c
Department of Pathology, University of Pittsburgh, School of Medicine, Pittsburgh, USA
AN
1
These authors contributed equally to this work.
M
Abstract: IGF1R is a ubiquitous receptor tyrosine kinase that plays critical roles in cell
proliferation, growth and survival. Clinical studies have demonstrated upregulation of IGF1R
AC
mediated signaling in a number of malignancies including colon, breast, and lung cancers.
Overexpression of the IGF1R in these malignancies is associated with a poor prognosis and overall
survival. IGF1R specific kinase inhibitors have failed in multiple clinical trials partly because of
the complex nature of IGF1R signaling. Thus indentifying new binding partners and allosteric
sites on IGF1R are emerging areas of research. More recently, IGF1R has been shown to
translocate into the nucleus and perform many functions. In this study, we generated a library of
1
ACCEPTED MANUSCRIPT
IGF1R deletion and point mutants to examine IGF1R subcellular localization and activation of
downstream signaling pathways. We show that the nuclear localization of IGF1R is primarily
defined by its cytoplasmic domain. We identified a cross-talk between IGF1R and Wnt/β-catenin
signalling pathways and showed, for the first time, that IGF1R is associated with upregulation of
T
and IGF1R specific inhibitor(s), we show that cytoplasmic and nuclear activities are two
IP
independent functions of IGF1R. Furthermore, we identified a unique loss-of-function mutation in
CR
IGF1R. This unique loss-of-function mutant retains only nuclear functions and sits in a pocket,
outside ATP and substrate binding region, that is suited for designing allosteric inhibitors of
IGF1R.
US
AN
Keywords: PI3K/Akt, Wnt/β-catenin signalling, IGF1R, Receptor Tyrosine Kinases
M
1. Introduction
ED
The type 1 insulin-like growth factor receptor (IGF1R) is a cell surface transmembrane receptor
activated by IGF-I and II [1]. Ligand binding to the extra-cellular α-subunit causes conformational
PT
changes which results in the autophosphorylation of specific tyrosine residues in the cytoplasmic
CE
domain of β-subunit leading to the phosphorylation of other sites on the receptor and substrate
proteins [2]. IGFIR and IR (insulin receptor), two closely related members of tyrosine kinase
AC
receptor superfamily [3], signal through multiple pathways that include PI3K and MAPK
pathways. IR is required for glucose homeostasis [4] whereas IGF1R regulates cell growth and
development [2]. IGFIR also plays crucial roles in the normal development of many tissues and its
deregulation has been reported in many cancers [5]. Furthermore, signaling mediated via the
IGF1R is believed to be important for the maintenance of tissue resident adult stem cells [6] and
for embryonic stem cell self-renewal, therefore, suggesting its importance in stem cell biology [7,
2
ACCEPTED MANUSCRIPT
8, 9]. IGF1R has been shown to translocate into the nucleus, despite lacking a putative nuclear
localization signal [10, 11, 12]. Reports show that nuclear IGF1R binds to its own promoter region
thereby regulating its own expression [13]. Nuclear localization of IGF1R is seen in many tumors
[10, 14, 15] and its nuclear levels are associated with Geftinib resistance in HCC cells [16].
Although IGF1R is shown to co-localize with RNA polymerase-II, precipitate with histones [10],
T
and associate with LEF1 (lymphoid enhancer factor-1) transcription factor [17], the exact nuclear
IP
roles of IGF1R are still not fully understood.
CR
In this study, we have generated a library of IGF1R mutants to investigate their sub-cellular
localization and their impact on various activities of IGF1R. We identified a cross talk between
US
IGF1R and Wnt/β-catenin signaling pathways and showed that IGF1R is involved in the
AN
upregulation of TCF (T-cell factor) mediated β-catenin signaling in HepG2 cells. Furthermore, we
showed that the cytoplasmic and nuclear activities are two independent functions of IGF1R.
M
cytoplasmic domain) that may provide an opportunity for the designing allosteric inhibitors of
IGF1R.
PT
CE
2.1. Plasmids
Human genes for expression of full length IGF1R and its domains were cloned by PCR
amplification of full length IGF1R human cDNA kindly provided by Dusty Miller [18].
IGF1R-WT and its domains (α, β and CTD) were amplified using primers described in
Supplementary Table 1. The amplified PCR products were purified and digested with NheI (NEB)
3
ACCEPTED MANUSCRIPT
and BamHI (NEB). The digested amplicons were cloned into NheI/ BamHI digested pEGFPN3
vector (Clonetech) using proof-reading enzyme Pfu. (NEB). The vector and the inserts were
ligated using T4 ligase (NEB) at 16 °C overnight. Sequences were verified to encode proteins
position 1055, 1130, 1150 and tyrosine to phenylalanine at position 1161, 1165, and 1166 were
T
generated in full length IGF1R protein, β-subunit and cytoplasmic domain by QuickChange
IP
Site-Directed Mutagenesis kit (Stratagene). For detail of primers used for mutagenesis see
CR
Supplementary Table 1.
US
2.2 Cell lines, reagents and antibodies AN
The human HepG2 liver cancer cells and normal HEK-293 cells were purchased from Sigma
Aldrich and National Centre For Cell Science (NCCS), Pune, India respectively. Both cell lines
M
were maintained in DMEM (Sigma D5523) media supplemented with 10% fetal bovine serum.
ED
The cell lines were transfected with various GFP-tagged constructs of IGF1R using lipofectamine
2000 (Invitrogen) as the transfection reagent. The pIGF1R (sc-135767), GFP (sc-9996), cyclin D1
PT
(sc-718) and GAPDH (sc-166574) antibodies were purchased from Santa Cruz Biotechnology Inc.
CE
The p-ERK (D13.14.4E), total ERK (137F5), total Akt (C67E7), p-Akt (S473) (D9E), p-Akt
(T308) (D25E6), anti-Lamin (4C11) (4777P) and anti-α-tubulin (2144) antibodies were purchased
AC
from Cell Signaling Technology (Danvers, MA). The ATP-competitive IGF1R inhibitor
NVP-ADW742 (S1088) was obtained from Selleck Chemicals. All other reagents and chemicals
4
ACCEPTED MANUSCRIPT
Cytosolic and nuclear fractions were prepared as described previously with minor modifications
[19, 20]. Transfected cells were resuspended in fractionation buffer (250mM Sucrose, 20mM
HEPES (7.4), 10mM KCl, 1.5mM MgCl2, 1mM EDTA, 1mM. EGTA, 1mM DTT, PI
Cocktail).Thereupon, the cells were passed through a 26G needle 10 times and incubated on ice
for 20 min. The homogenate was centrifuged at 3,000 rpm for 5 min. The pellet contained the
T
unbroken nuclei which were then resuspended in 1X RIPA buffer (Sigma) and processed as
IP
described under section 2.4. The supernatant was centrifuged again at 14,000 rpm for 30 min to
CR
remove unbroken cells, and the final supernatant obtained was used as the cytosolic fraction.
6-well-plates. After 24 hr in culture, cells were transiently transfected with 2 µg each of GFP
tagged IGF1R constructs using 4 µl of lipofectamine. The transfected cells were subsequently
ED
grown with or without serum for 24 hours. After that whole cell lysates were prepared in 1X RIPA
PT
buffer (Sigma) with added sodium orthovandate (100mM), protease cocktail (Roche), NaF
(100mM), PMSF (100mM) and EDTA (100mM). Protein estimation was done using Bradford
CE
reagent (Biorad). The samples were boiled with sample buffer containing 1% β-mercaptoethanol,
AC
6%glycerol, 2% SDS, 22mM Tris-HCl pH6.8 and bromophenol blue. Whole cell lysates
corresponding to 70-100ug of protein were loaded. The samples were analysed by SDS-PAGE
with a 12% separating gel. A molecular weight marker (Thermo scientific 26619) was run
simultaneously. After running the gel, the proteins were transferred to PVDF membrane
(Millipore) and then blocked for 1 hour in solution of 5% BSA (Sigma), 0.1% Tween 20
(HiMedia), 150mM NaCl and 20mM Tris-HCl pH7.4. The membranes were then probed with
5
ACCEPTED MANUSCRIPT
specific antibodies for pIGF1R (1:200), p-Akt (S473) (1:1000), p-Akt(T308) (1:1000), total Akt
(1:1000), total Erk (1:1000), p-Erk (1:1000) , GAPDH (1:500), GFP(1:500) and visualized using
ECL (Millipore).
T
HEK-293 cells or HepG2 cells (30,000 cells/ well) were seeded in 4-well chambered coverglass
IP
slides (Thermo Scientific, Lab-Tek 155383) (Rochester, NY). At 24 hours cells were transiently
CR
transfected with 500ng of DNA using 1ul lipofectamine-2000 (Invitrogen). The transfected cells
were supplemented with complete medium for 24 hours. After 24 hours, cells were washed with
US
1X PBS and fixed with 4% paraformaldehyde (Sigma). Subsequently they were permeabilized
AN
with 0.5% Triton X (Sigma) and stained with DAPI (Sigma). Cells were mounted using
Glycerol:PBS solution (9:1) and visualized using a confocal microscope (Olympus Fluoview
M
FV-1000). The magnifications used were 20X (with 1.5X zoom) and 40X.
ED
Reporter assays were performed using HepG2 cells, seeded to attain70-80% confluency on the
CE
next day. Cells were transfected with 1µg each of GFP-IGF1R constructs along with 800 ng of
TopFlash (TOP) or Fopflash (FOP) and 200 ng of Renilla DNA. Inhibitor treatment was given at
AC
the time of media supplementation. Cells supplemented with complete media or serum starved for
24 hours were then washed once with PBS, and lysed in 250µl of Passive Lysis Buffer according
to Dual Luciferase Reporter protocol (Promega), and activities were then read on a luminometer
(TECAN microplate reader Infinite M200 PRO). Luciferase activity was quenched, and Renilla
expression was detected by the addition of 70 μl of STOP-GLO reagent. TopFlash values were
6
ACCEPTED MANUSCRIPT
calculated as ratios of Luciferase signal to Renilla signal. Each condition was assayed in triplicate,
The PDB file of crystal structure of unactivated apo form of IGF1R (PDB ID: 1M7N) was
T
retrieved from the Protein Data Bank (Berman et al., 2000). The mutation to replace Lys1052 with
IP
Arginine was performed using Mutagenesis Wizard of PyMol(DeLano.2002). To compare the
CR
changes in bonds between the amino acid residues the Discovery Studio visualiser 4.1 (DS) was
used. Simulation sessions of 20 ns each were performed using Gromacs (Berendsen et al., 1995) at
US
300 K temperature and 1 bar of atm pressure. Different Gromacs utilities were used to analyze the
AN
structural changes in the IGF1R tyrosine kinase structure before and after mutation.
M
Adobe Photoshop 7.0 was used for processingwestern blot and confocal images. All datawas
analysed using Prism (GraphPad Prism Software, La Jolla, CA, USA). Differences between means
PT
were evaluated by Student’s t-test and analysis of variance. Data are expressed as mean ± SD.
CE
3. Results
AC
3.1 Expression, activity and sub-cellular localization of GFP tagged IGF1R and its mutants
in HEK293 cells
IR and IGF1R have recently been shown to translocate to the nucleus in many cell lines [21]. Upon
[10, 14, 15, 22], however its specific nuclear role has not been established yet. Here, we analyzed
7
ACCEPTED MANUSCRIPT
the subcellular localization of IGF1R by tagging GFP at its C-terminal end. Confocal microscopy
analysis showed that IGF1R-GFP(WT) was present at the membrane, in the cytoplasm as well as
in the nucleus. By contrast, GFP alone was diffused in the nucleus and cytoplasm as expected.
EGFR-GFP and S45Y-GFP (β-catenin mutant) were used as controls for membranous and nuclear
localization respectively [23, 24] (Fig. 1A and Supplementary Fig. 1A). Expression checked by
T
immunoblot analysis showed that the phosphorylated IGF1R-GFP(WT) is expressed as a protein
IP
of nearly 200 kDa (pro-IGF1R) and 130 kDa (β-domain-GFP) (Fig. 1B lane 2 upper and lower
CR
panel). Subcellular fractionation analysis and quantification of relative amount
of IGF1R-GFP(WT) in the nucleus versus cytoplasm was carried out to confirm the nuclear
US
translocation of IGF1R in HEK293 cells (Figure 1D, Supplementary Fig 1A&B). Since the
AN
predicted size of IGF1R-GFP exceeds the passive diffusion limit of nuclear pore (which is less
than 35–40 kDa), Beta Sehat et al identified three lysine residues at position 1025, 1100, and 1120
M
(represented as K1055, K1130, and K1150 in our study) as trafficking determinants in IGF1R [11].
ED
Thus, we generated IGF1R constructs with mutations for these residues and compared them to
those of wild type IGF1R to ascertain their expression and sub-cellular localization. Confocal
PT
microscopy analysis of GFP tagged single lysine mutants (K1055R and K1130R), double lysine
CE
mutants-DKM (K1055R, K1130R) and triple lysine mutants-TKM (K1055R, K1130R, K1150R)
in the context of IGF1R-GFP(WT) showed no change in their sub-cellular localization (Fig. 1C),
AC
suggesting that these residues have no significant role in nuclear localization of IGF1R. As an
initial check of activity, we determined the autophosphorylation activity of these proteins and
observed that the wild type and K1130R mutant has intact autophosphorylation activity whereas
K1055R in the form of single (K1055R), double lysine mutant (DKM) or triple lysine mutant
8
ACCEPTED MANUSCRIPT
3.2 Role of IGF1R domains and its critical residues in its sub-cellular localization and
pathway activation
IGF1R consists of two covalently linked polypeptide chains each with an extracellular α-subunit
that binds ligand and a transmembrane β-subunit that contains tyrosine kinase activity. Binding of
T
IP
ligand results in trans-autophosphorylation of IGF1R β-subunit which is required for activation of
CR
the receptor and downstream signaling pathways. To investigate whether the intact receptor or its
individual domains localize to the nucleus, we generated deletion mutants of IGF1R as shown in
US
Fig.2A. We transfected HEK293 cells with these plasmids and expression of correct size proteins
along with their autophosphorylation status was analyzed by immunoblotting using protein
AN
specific antibody (Fig. 2C upper panel). Except for α-subunit, IGF1R-WT, β-subunit and CTD
M
were getting autophosphorylated as they harbor an intact tyrosine kinase domain (Fig.2C lower
panel). Upon analyzing sub-cellular localization, GFP tagged α-subunit was predominantly seen at
ED
the membrane whereas β-subunit-GFP was seen at the membrane as well as in the nucleus and
PT
cytoplasm (Fig. 2B and and Supplementary Fig.2). Since, β-subunit possesses a transmembrane
region; we assumed it may not translocate to nucleus, however, it showed prominent nuclear and
CE
cytoplasmic localization. Unlike β-subunit, CTD domain, which lacks transmembrane region,
AC
showed predominant nuclear localization. The nuclear localization of β-subunit and CTD was not
abrogated by the incorporation of the lysine mutations mentioned above (Supplementary Fig.3A
and 3B), reiterating that these residues play no role in nuclear localization of wild type receptor
and its deletion mutants. These results demonstrate that the nuclear localization is rather defined
3.3 Erk and Akt activation in relation to overexpression of IGF1R-GFP(WT) and its
9
ACCEPTED MANUSCRIPT
mutants
The signaling pathways utilized by IGF1R to mediate its diverse effects in normal and cancerous
cells are very well established [25]. Thus, after knowing the sub-cellular localization of
IGF1R-GFP(WT) and its variants, we tried to understand the signaling contribution of these
proteins. We transfected HEK293 cells with plasmids encoding these proteins under basal
T
conditions. We analyzed pathway activation under serum free conditions also to ascertain that the
IP
differences seen in the phosphorylation levels of p-Akt and p-Erk were exclusively due to the
CR
ectopic expression of IGF1R and its mutants alone. Consistent with earlier reports IGF1R-WT
activated PI3K and MAPK pathways [25] while the deletion mutants showed either reduced or no
US
activity (Fig.3A and 3B). The β-subunit showed considerable activity although lower than
AN
IGF1R-GFP(WT) and activated both pathways while others report β-subunit to activate only
PI3K/Akt pathway [26, 27]. The activity of β-subunit can be explained by its diffused sub-cellular
M
localization and by the presence of intact tyrosine kinase domain, as reported earlier [27]. The
ED
CTD, despite showing autophosphorylation (Fig.2C lane 4), did not significantly enhance the
p-Akt and p-Erk levels, suggesting that this domain alone is not sufficient for activating
PT
downstream signaling which contradicts earlier reports [9]. Rosengren et al [26] have reported that
CE
β-subunit expressed in R-cell causes their transformation whereas similar protein in our studies
does not override the activity of wild type protein. Thus, understanding the exact roles of β-subunit
AC
and CTD needs further investigation. In addition, we compared the pathway activation of cells
treated with IGF-1 alone with cells overexpressing recombinant IGF1R and observed that
IGF1R-GFP(WT) is still showing relatively higher pathway activation than the untransfected
10
ACCEPTED MANUSCRIPT
IGF1R-GFP(WT), β-subunit and CTD proteins, and observed that their autophosphorylation
activities were completely abrogated (Fig. 3C and 3F). This corresponded to their reduced levels
of p-Akt and p-Erk as well (Fig.3D). Although the mutants showed no autophosphorylation and
have reduced p-Akt and p-Erk levels, there was no change in their subcellular localization as
observed by confocal microscopy (Fig. 3C and 3F) indicating that the kinase activity of IGF1R
T
does not regulate its sub-cellular localization.
IP
CR
3.4 Overexpression of IGF1R promotes the β-catenin transcriptional activity in HepG2 cells
The functions of nuclear IGF1R are poorly understood. There are reports that IGF1R binds to
US
putative enhancer regions and promotes transcription [13]. Warsito D et al showed that IGF-1R
AN
associates with TCF (T cell factor) and increases promoter activity of β-catenin physiological
downstream target genes cyclin D1 and axin 2 [17]. Moreover, ligands IGF-I and II are known to
M
promote β-catenin nuclear translocation and subsequent expression of its target genes c-myc and
ED
transcriptional activity in HepG2 cells using Top-Flash reporter assay- the most robust assay for
CE
measuring TCF-mediated β-catenin transcriptional activity. HepG2 cells are known to have high
β-catenin activity because of the presence of non-degradable form of β-catenin in these cells and
AC
IGF has been shown to promote this activity [23, 24, 29, 30]. Thus, we overexpressed
IGF1R-GFP(WT) in HepG2 cells and observed nearly 3.5-fold increase in β-catenin Top-Flash
luciferase activity (Fig. 4A). Considering that IGF1R nuclear signaling may be dependent on the
presence of IGFs in the serum, we also used serum starved HepG2 cells, however, we found no
significant difference in the activity indicating that IGF1R could increase β-catenin mediated
11
ACCEPTED MANUSCRIPT
Top-Flash activity in a ligand independent manner. In order to validate our results we performed
these experiments using cyclin D1 and NF-κB reporter assays as positive and negative controls.
While NF-kB reporter showed no change in activity we observed nearly 2.3 fold increase in cyclin
T
(supplementary Fig. 5B) consistent with results reported by Warsito D et al [17], indicating that
IP
IGF1R specifically upregulates TCF-mediated beta-catenin activity. We then analyzed the activity
CR
of IGF1R deletion mutants in comparison to vector control (Fig. 4B). Reduced Top-Flash activity
US
membrane. The β-subunit showed relatively higher nuclear activity (a 2.28-fold increase) than the
AN
α-subunit and CTD, however, IGF1R-GFP(WT) still retained the highest activity.
the β-catenin-TCF-4 complex formation, we used TCF-4 dominant negative in our experiments.
ED
TCF-4 dominant negative protein does not bind β-catenin and acts as a potent inhibitor for
Top-Flash reporter activity. The Top-Flash activity was significantly reduced in presence of
PT
dominant negative TCF (DN-TCF) (Fig.4C). This confirms that IGF1R promotes TCF mediated
CE
3.5 IGF1R kinase activity is not required for promoting β-catenin mediated transcription
Previous studies have shown that ATP binding by the tyrosine kinase domain of IGF1R and IR is
critical to initiate autophosphorylation, a necessary proximal step for pathway activation. Cluster
of three tyrosine residues (Y1161, Y1165, and Y1166), which are critical for receptor activity, are
first to be phosphorylated [31]. We showed above that upon mutating these residues the
cytoplasmic activities of IGF1R are completely abrogated. We then examined the Top-Flash
12
ACCEPTED MANUSCRIPT
reporter activity upon mutating these tyrosine residues to phenylalanine, however, no change in
activity was seen when compared to IGF1R-GFP(WT) under both basal and serum free conditions
(Fig.5A and 5B). To further dissect the role of IGF1R tyrosine kinase activity on its nuclear
functions, we used NVP-ADW742, a selective IGF1R inhibitor. Treatment of HepG2 cells with
NVP-ADW742 showed a dose dependent decrease in cell viability with IC50 value of 80 nM
T
(Supplementary Fig.6A). Upon analyzing downstream pathway activation, NVP-ADW742
IP
treatment of HepG2 cells showed a significant decrease in p-Akt levels in a dose dependent
CR
manner (Supplementary Fig.6B). We then determined the effect of various concentrations of
US
observed no change in this activity at lower concentrations of NVP-ADW742 (below its ic50)
AN
suggesting that the kinase activity of IGF1R does not impact the nuclear functions of this protein
(Fig. 5C).
M
3.6 Impact of K1055R mutation on nuclear and cytoplasmic activities of IGF1R and its
ED
deletion mutants
PT
Since K1055R mutations failed to impede the nuclear localization of IGF1R, we explored their
effect on cytoplasmic and nuclear activities in HEK293 and HepG2 cells respectively. Unlike
CE
K1130R mutant, K1055R was seen to abolish the kinase activity of IGF1R in the context of single,
AC
double or triple mutations as shown by reduced p-Akt and p-Erk levels (Fig. 6A). We also
incorporated similar mutations in the β-subunit and CTD and observed that their
autophosphorylation activity was completely abolished as well (Fig. 6B). However, there was no
under serum free and basal conditions (Fig. 6C). Thus, we have identified a unique mutation in
cytoplasmic domain of the receptor which is not involved in substrate and ATP binding. This
13
ACCEPTED MANUSCRIPT
mutation abrogates kinase activity of IGF1R while sparing the nuclear functions. We provide
substantial evidence that IGF1R mutants lacking kinase activity translocate normally into the
nucleus and promote Top-Flash activity. Thus we observed that the nuclear and cytoplasmic
T
IP
3.7 Insilico analysis of novel K1055R mutation and its impact on various activities of IGF1R
CR
Inactivated apo (wild type) and K1052R (K1055 is referred as K1052 in PDB: 1M7N) mutant
versions of IGF1R were subjected to molecular dynamics simulation for 20 ns session using
US
Gromacssoftware (37). The simulation trajectories were used for analyzing Root Mean Square
Deviations (RMSDs) to assess the stability of the protein backbone against time. The RMSD
AN
values of Cα atoms of protein backbone showed that both wild and mutant forms of the protein
M
were stable along the trajectory after an initial increase. The wild type form of protein was having
more RMSD values than the K1052R mutant (Supplementary Fig.7A). The Root Mean Square
ED
Fluctuation (RMSF) analysis for each amino acid residue of protein showed that the fluctuations in
PT
the N and C terminal loop residues dropped from 0.9 nm to a value of 0.3 nm and 0.65 nm to 0.4
nm in K1052R mutant as compared to the wild type respectively (Supplementary Fig.7B). The
CE
hydrogen bond analysis also showed that the number of hydrogen bonds was increased in the case
AC
of K1052R mutant (Supplementary Fig.7C) and there is also difference in the interactions between
the K1052 and the rest of the protein (Supplementary Table 2). The distance between the N
terminal end and αc-helix was calculated which showed that the N terminal came in close
proximity to the αc-helix in the case of K1052R mutant when compared to the wild IGF1R
(Fig.7A). Similar results were verified from the alignment of wild type and K1025R mutant
IGF1R conformers obtained after the simulation session (Fig.7B). Mutation at K1052 also resulted
14
ACCEPTED MANUSCRIPT
in structural changes in the activation loop. Fig.7C shows that two small β-strands and a partial
helix were generated in the activation loop when PDBs derived from the end point of simulation
trajectory of both Wild and K1052R mutant tyrosine kinase were compared using structure
alignment (Fig.7C). The side chain conformational changes were also observed in the three
autophosphorylation tyrosine residues Tyr1158, Tyr1162 and Tyr1163 at the activation loop [32]
T
(Fig.7D). Tyr1158 was shown to have maximum shift in the orientation of its side chain (Fig.7E).
IP
These changes may lead to loss of autophosphorylation and ultimately yield disrupted kinase
CR
activity. Further, we have observed one potential binding pocket in IGF1R receptor involving
K1052 residue and the N-terminal loop. This pocket is constituted by the side chains of residues
US
W989, K1052, E1054, F1055, N1056, C1057, R1061 and L1062 (Fig.7F). There is also similar
AN
pocket reported in human ephrin receptor tyrosine kinase (PDB ID: 2QON) which is responsible
4. Discussion
PT
IGF1R signaling regulates cell growth, survival and plays significant roles in stem cell
proliferation and differentiation. IGF1R primarily mediates its downstream activities via PI3K and
CE
MAPK signaling pathways. IGF-I mediated phosphorylation of Akt and GSK3β is known to
promote β-catenin stability and nuclear localization suggesting that β-catenin is an important
AC
downstream molecule of IGF1R signaling. However, the precise downstream targets and
cross-talk between these two signaling pathways remain yet to be defined. RTKs are found in the
nucleus in two forms, either as intact molecules (e.g. EGFR and FGFR-I) [34, 35] or as
cytoplasmic fragments produced by the sequential action of two distinct membrane-localized
proteases. Here, we sought to understand how IGF1R is translocated into the nucleus and the
impact of its subcellular localization on downstream signaling. We evaluated the subcellular
localization by expressing GFP fusion IGF1R proteins in HEK293 and HepG2 cells. Upon
transfection of IGF1R-GFP(WT) in HEK293 cells, we were able to express a fully functional and
15
ACCEPTED MANUSCRIPT
correct sized protein. IGF1R (wildtype) showed predominant membranous localization, however,
significant cytoplasmic and nuclear levels were also observed. Although lacking a canonical
nuclear localization sequence (NLS), IGF1R is reported to have the necessary structural
requirement to move into the nucleus. Beta Sehat et al [11] identified three lysine residues as
important trafficking determinants for IGF1R nuclear localization. We investigated whether these
lysine residues were sufficient for IGF1R nuclear localization in HEK293 cells, and thus we
generated single (K1055R and K1130R), double (K1055R, K1130R) and triple lysine mutants
T
(K1055R, K1130R, K1150R). Upon expressing correct size proteins, K1055R mutant was seen to
IP
be defective for autophosphorylation and subsequent downstream signaling activity in comparison
CR
to IGF1R-GFP(WT), K1130R and K1150R mutants. However, the nuclear localization of these
three lysine mutants remained unchanged. To further elucidate the effect of K1055R mutation, we
US
performed in silico analysis. Mutating K1055 (referred as K1052 in PDB:1M7N) to Arginine lead
to changes in the interaction between the N-terminal loop and the αc-helix, bringing them close to
AN
each other. This interaction caused conformational changes in the orientation of the side chain of
the Y1158 responsible for the blockage in the transfer of autophosphorylation activity, a
M
prerequisite for the kinase activity of IGF1R [36]. Thus, we report for the first time, a unique
loss-of-function mutation in IGF1R located outside substrate and ATP binding region. In
ED
Wnt/β-catenin signaling pathways. Upon expression in HEK293 cells, α-subunit was exclusively
present at the membrane while the cytoplasmic domain (CTD) was present equally in the nucleus
CE
and cytoplasm but none at the membrane. The β-subunit, possessing membrane spanning domain
and extracellular region, showed a more diffused subcellular localization and was present in the
AC
cytoplasm, nucleus and at the membrane too. Again, upon incorporating point mutations for the
trafficking determinants of three lysine residues into the CTD and β-subunit, the localization status
of proteins remained unchanged reiterating that these three lysine residues do not regulate the
nuclear localizing of IGF1R and its individual domains, nuclear localization of IGF1R is rather
determined by its CTD domain alone, how that exactly happens needs further investigation. We
then investigated the IGF1R mediated PI3K pathway activation by assessing the p-Akt status and
MAPK pathway activation by assessing the p-Erk levels in cells transiently expressing
IGF1R-GFP (WT) and its deletion mutants. We observed significant differences in the activities of
16
ACCEPTED MANUSCRIPT
full-length and truncated proteins. The IGF1R-GFP (WT) induced maximal phosphorylation of
p-Akt, whereas the β-subunit showed moderate p-Akt activation, α-subunit and CTD failed to
show any significant kinase activity. Similar results were obtained for the p-Erk status of these
proteins. A very important aspect of this study is that that CTD which retains the
autophosphorylation potential is not fully active for substrate phosphorylation (p-Akt and p-Erk);
thus whether CTD which is pursued for the development of inhibitors [37] suffices the structural
requirement for screening these inhibitors needs to be revisited. As mentioned above, IGF1R is
T
reported to be localized in nucleus and the ligand binding and kinase activity are shown to be
IP
critical for its nuclear localization [10,11,14]. IGF-1 and insulin have been shown to promote
CR
β-catenin transcriptional activity through PI3K signalling [30]. Therefore, we sought to analyze
the nuclear activity of IGF1R and its mutants and have identified a specific role for IGF1R that
US
involves promotion of β-catenin mediated transcriptional activity in HepG2 cells. We show that
IGF1R-GFP(WT) is necessary for both cytoplasmic tyrosine kinase activity and nuclear functions,
AN
albeit β-subunit also shows moderate levels of nuclear and cytoplasmic activity. As expected,
α-subunit is devoid of any activity whereas the CTD displays modest nuclear activity, although
M
both these domains fail to show any detectable p-Akt and p-Erk activation. We carried out nuclear
activities of IGF1R-GFP(WT) under basal and serum free conditions, however, no change in
ED
activity was observed indicating that the overexpression of IGF1R bypasses the requirement for
stimulation with its cognate ligands (IGF-I and II). To elucidate further that IGF1R specifically
PT
acts through β-catenin pathway we carried out reporter assays using TCF-DN proteins. As
expected, Top-Flash activity was reduced in the presence of TCF-DN in IGF1R-GFP transfected
CE
HepG2 cells. Finally, we show the kinase activity of IGF1R is not required for promoting
β-catenin transcriptional activity by using kinase dead point mutants (K1055R, DKM, TKM, DTM
AC
and TTM) and IGF1R specific small molecule inhibitor, NVP-ADW742. This is for first time such
a detailed study has been carried out to segregate these two roles of IGF1R. Thus, we report a
cross-talk between IGF1R and Wnt/β-catenin signaling and show that the nuclear IGF1R is
involved in upregulation of TCF mediated β-catenin transcriptional activity. Moreover, we show
that disrupting the interaction of K1055 (K1052R) with its cognate binding residue resulted in the
generation of inactive IGF1R that failed to show any kinase activity. The alignment of PDBs of
wild type and K1052R mutant IGF1R tyrosine kinase derived after the simulation session of 20ns
showed that in case of K1052R mutant there are changes in the secondary structure content of the
17
ACCEPTED MANUSCRIPT
activation loop. More importantly, a potential binding pocket constituted of residues W989,
K1052, E1054, F1055, N1056, C1057, R1061 and L1062 was observed in this region which
generates new hope for designing allosteric inhibitors for IGF1R. This binding pocket provides a
unique platform relevant for the development of allosteric inhibitors for IGF1R. The highly
anticipated IGF1R allosteric inhibitors are expected to exclude the unwanted issues related to
kinase inhibitors which target IGF1R and IR in equal measure due to high degree of sequence
homology. In conclusion, we observed kinase dependent and kinase independent roles for IGF1R.
T
IGF1R kinase activity is critical for its conventional cytoplasmic activities whiles as the kinase
IP
independent roles impact its nuclear functions. The cross-talk between IGF1R and Wnt/β-catenin
CR
signaling pathways reported here, is relevant to further understand the role played by IGF1R in cell
proliferation and differentiation.
US
AN
5. Conclusion
In this study, we generated a library of IGF1R constructs in order to study the subcellular
M
localization, and cytoplasmic and nuclear activities of this protein. We compared the nuclear and
ED
cytoplasmic activity of IGF-1R and its mutants. We observed the nuclear localization of IGF1R is
PT
primarily mediated by its C-terminal domain. Upon nuclear localization, IGF1R was seen to
IGF1R promoted its nuclear activity at normal levels, while the same mutants failed to activate
AC
PI3K and MAP kinase signaling in the cytoplasm. This was further proved by the use of ,
NVP-ADW742, a IGF-1R specific inhibitor, which blocked the kinase mediated cytoplasmic
activities of IGF-1R while sparing its nuclear activities. Moreover, we identified a unique
loss-of-function mutation in the C-terminal domain of IGF1R that helped to further elucidate the
kinase dependent and independent function of IGF-1R. In conclusion, we have identified a link
between two signaling pathways-IGF1R and Wnt/β-catenin signaling, and we believe this
18
ACCEPTED MANUSCRIPT
cross-talk is relevant to further address how IGF-1R signaling is involved in cancer and stem cell
biology. Unique loss of function mutation identified in this study would help in the identification
of allosteric small molecule inhibitors to block the activity of IGF1R in cancer cells.
T
IP
CR
US
AN
M
ED
PT
CE
AC
19
ACCEPTED MANUSCRIPT
Acknowledgments
We are also grateful to Dr. P. R Sharma for helping with confocal microscopy. This work was
supported by a grant from the Council of Scientific and Industrial Research (CSIR), Government
of India, New Delhi, under Network Project BSC-0108. This manuscript represents IIIM
T
IP
Commision, New Delhi, for providing the fellowship. Yusuf Akhter’s lab is supported by
CR
extramural funds from Govt. of India agencies (UGC, SERB-DST and ICMR).
US
AN
M
ED
PT
CE
AC
20
ACCEPTED MANUSCRIPT
References:
and cellular aspects of the insulin-like growth factor I receptor. Endocr Rev
T
IP
16, 143–163.
CR
3] Ullrich A, Gray A, Tam AW, Yang-Feng T, Tsubokawa M, Collins C, Henzel W,
US
primary structure: comparison with insulin receptor suggests structural
7] Wang L, Schulz TC, Sherrer ES, Dauphin DS, Shin S, Nelson AM, Ware CB,
Zhan M, Song CZ, Chen X, Brimble SN, McLean A, Galeano MJ, Uhl
EW, D'Amour KA, Chesnut JD, Rao MS, Blau CA & Robins AJ (2007)
21
ACCEPTED MANUSCRIPT
8] Nguyen TT, Sheppard AM, Kaye PL & Noakes PG (2007) IGF-I and insulin
Activate mitogen-activated protein kinase via the type 1 IGF receptor in mouse
T
Chem 270, 29825-29830.
IP
10] Aleksic T, Chitnis MM, Perestenko OV, Gao S, Thomas PH & Turner GD
CR
(2010) Type 1 Insulin-like Growth Factor Receptor Translocates to the Nucleus
US
11] Sehat B, Tofigh A, Lin Y, Trocmé E, Liljedahl U, Lagergren J & Larsson
AN
O (2010) SUMOylation Mediates the Nuclear Translocation and Signaling of the
12] Packham S, Warsito D, Lin Y, Sadi S, Karlsson R, Sehat B & Larsson O (2015)
ED
22
ACCEPTED MANUSCRIPT
Goros MW, Michalek JE, Wang S, Mansoor A, Druker BJ, Wagers AJ,
Tyner JW & Keller C (2013) Dynamic and nuclear expression of PDGFRα and
T
16] Bodzin AS,Wei Z, Hurtt R, Gu T & Doria C (2012) Gefitinib resistance in HCC
IP
mahlavucells: upregulation of CD133 expression, activation of IGF1R signaling
CR
pathway and enhancement of IGF1R nuclear translocation. J Cell Physiol
227, 2947-2952.
US
17] Warsito D, Sjöström S, Andersson S, Larsson O & Sehat B (2012) Nuclear
AN
IGF1R is a transcriptional co-activator of LEF1/TCF. EMBO Rep 13, 244-250.
18] Kaleko M, Rutter WJ & Miller D (1990) Overexpression of the human Insulin
M
19] Singh P, Dar MS, Singh G, Jamwal G, Sharma PR, Ahmad M & Dar MJ (2016) Dynamics of
PT
GFP-fusion p110α and p110β isoforms of PI3K signaling pathway in normal and cancer cells. J
CE
21] Sarfstein R & Werner H (2013) Minireview: nuclear insulin and insulin-like
23
ACCEPTED MANUSCRIPT
tumor cells requires elevated expression of the receptor and the SUMO-
23] Dar MS, Singh P, Singh G, Jamwal G, Hussain SS, Rana A, Akhter Y, Monga SP
T
& Dar MJ (2016) Terminal regionsof β-catenin are critical for regulating its adhesion and
IP
transcription functions. Biochim Biophys Acta 1863(9), 2345-2357.
CR
24] Dar MS, Singh P, Mir RA & Dar MJ (2017) Beta-catenin N-terminal domain: An
US
enigmatic regionprone to cancer causing mutations. Mut Res 773, 122-133.
AN
25] Baserga R (2000) The contradictions of the insulin-like growth factor 1 receptor.
24
ACCEPTED MANUSCRIPT
G, Perret C & Capeau J (2001) Insulin and IGF-1 stimulate the beta-catenin
pathway through two signalling cascades involving GSK-3beta inhibition and Ras
T
31] Kato H, Faria TN, Stannard B, Roberts Jr CT & LeRoith D (1994) Essential role
IP
of tyrosine residues 1131, 1135, and 1136 of the insulin-like growth factor-I
CR
(IGF-I) receptor in IGF-I action. Mol Endocrinol 8, 40-50.
US
Crystal structure of bisphosphorylated IGF-1 receptor kinase: insight into domain
AN
movements upon kinase activation. Structure 9, 955-965.
34] Brand TM, Iida M, Li C & Wheeler DL (2011) The nuclear epidermal growth
factor receptor signaling network and its role in cancer. Discov Med 12, 419-432.
PT
muscle myosin.Biophys.J.93,2083-2090.
25
ACCEPTED MANUSCRIPT
T
IP
CR
US
AN
M
ED
PT
CE
AC
26
ACCEPTED MANUSCRIPT
Figure Legends
Fig. 1. Expression and localization of IGF1R-GFP (WT) and its mutants in HEK-293 cells.
(A) HEK-293 cells were transiently transfected with IGF1R-GFP (WT) plasmid to analyze its
expression and localization via confocal microscopy at 40X magnification. EGFR-GFP and
T
respectively. GFP plasmid was used as transfection control. Arrow marks the position of nuclear
IP
IGF1R. (B) HEK-293 cells were transiently transfected with IGF1R-GFP (WT) plasmid to
CR
determine its expression and phosphorylation status by immunoblotting using anti-GFP and
US
transfected with plasmids of IGF1R-GFP single, double (DKM) and triple (TKM) lysine mutants
AN
were analyzed for their expression and localization via confocal microscopy at 20X magnification
(with 1.5X zoom). Arrow marks the position of nuclear IGF1R. (D) HEK-293 cells were
M
transiently transfected with IGF1R-GFP (WT) plasmid to determine its expression in cytoplasmic
ED
and nuclear fractions using anti-GFP antibody (representative of three independent experiments).
α-tubulin and Lamin were used as controls for cytoplasmic and nuclear fractions respectively.
PT
(E) Immunoblotting (representative of three independent experiments) shows the expression and
CE
phosphorylation status of IGF1R-GFP (WT) and its lysine mutants in transiently transfected
HEK-293 cells in the absence of serum. *Marks the position of non-specific band
AC
Fig. 2. Role of IGF1R domains and its critical residues in nuclear localization. (A) Schematics
of IGF1R deletion and point mutants. The amino acid position of important residues and domains
is shown. The lysines residues are important for SUMO protein binding and tyrosines residues
indicate the autophosphorylation sites. All the constructs were tagged with GFP at the C-terminal.
27
ACCEPTED MANUSCRIPT
(B) HEK-293 cells were transiently transfected with indicated plasmids to analyze their
expression and localization via confocal microscopy at 40X magnification. GFP was used as
transfection control. (C) HEK-293 cells were transiently transfected with IGF1R-GFP (WT)
plasmid and its deletion mutants for determining their expression level and phosphorylation status
T
independent experiments). * Marks the position of degradation product as seen on immunoblot.
IP
Fig. 3. Pathway analysis of IGF1R-GFP (WT) and its mutants in HEK-293 cells. (A)
CR
Immunoblot analysis of HEK-293 cells transiently transfected with IGF1R-GFP (WT) and its
deletion mutants was performed in the presence and absence of serum using indicated antibodies
US
to determine their effect on downstream signaling. GFP was used as control. The figure is
AN
representative of three independent experiments. (B) Densitometry analysis of western blot
experiments (n=3, mean ± SD) (*** P<0.001) shows full-length IGF-1R has significantly higher
M
signaling capacity than its deletion mutants followed by its β-domain. (C) Immunoblotting shows
ED
the expression and autophosphorylation status of IGF1R-GFP (WT) and its tyrosine mutants using
anti-GFP and anti-phospho-IGF-1R antibodies in the absence of serum. Confocal images show the
PT
expression and localization of GFP-tagged IGF1R double tyrosine mutant (DTM) and triple
CE
tyrosine mutant (TTM) at 20X magnification (with 1.5X zoom). Arrow marks the position of
nuclear IGF1R. (D) Immunoblotting shows the effect of IGF1R-GFP tyrosine mutants on
AC
downstream signaling in the absence of serum. The experiment was done three times and a
representative blot is shown. (E) Densitomerty analysis of western blots shows significant increase
in downstream signaling by IGF-1R-GFP (WT) as compared to its kinase dead mutants (n=3,
mean ± SD) (*** P<0.001). (F) HEK-293 cells were transfected with β-domain-GFP and
CTD-GFP wild type and their tyrosine mutants. Images were taken with confocal microscope at
28
ACCEPTED MANUSCRIPT
20X magnification (with 1.5X zoom) to determine subcellular localization. Arrow marks the
β-domain-GFP and CTD-GFP WT and their respective DTM and TTM mutants in HEK-293 cells
transiently transfected with these constructs. These experiments were carried out in the absence of
T
IP
Fig. 4. TopFlash activity of IGF1R full length and deletion mutants in HepG2 cells. (A)
CR
Reporter activity was determined upon co-transfection of Top-Flash (TOP) or Fop-Flash (FOP)
and Renilla plasmids into HepG2 cells along with GFP and IGF1R-GFP (WT) under basal and
US
serum free conditions (materials and methods). (B) Reporter assays were performed using
IGF1R-GFP (WT) and deletion mutants with GFP as control. IGF1R-GFP (WT) showed highest
AN
activity followed by β-domain (C) Reporter assays were performed using IGF1R-GFP (WT) and
M
dominant negative TCF. GFP (pEGFPN3 vector DNA) and pcDNA (vector DNA) were used as
controls for IGF1R-GFP (WT) and DN-TCF respectively. Results represent the average ± SD of
ED
Fig. 5. TopFlash activity of IGF1R-GFP tyrosine mutants in HepG2 cells. (A) and (B)
Reporter activity was determined upon co-transfection of Top-Flash (TOP) or Fop-Flash (FOP)
CE
and Renilla plasmids into HepG2 cells along with IGF1R-GFP (WT) and tyrosine mutants under
AC
basal and serum free conditions with GFP as control (material and methods). (C) Reporter activity
was determined upon co-transfection of Top-Flash (TOP) or Fop-Flash (FOP) and Renilla
plasmids into HepG2 cells along with IGF1R-GFP (WT) and IGF1R tyrosine kinase inhibitor
NVP-ADW742 at the indicated concentrations with 0.01% DMSO under basal conditions
(materials and methods). The data shown represents the average ± SD of three experiments. ***P
< 0.001.
29
ACCEPTED MANUSCRIPT
Fig. 6. Impact of K1055R mutation on nuclear and cytoplasmic activities of IGF1R and its
deletion mutants in HEK-293 and HepG2 cells. (A) HEK-293 cell lysates were analyzed by
immunoblotting to determine the effect of wild type and lysine mutant proteins on downstream
signaling using the indicated antibodies. The experiment was performed under serum free
T
(n=3, mean ± SD) shows K1055R, DKM and TKM have significantly lower signaling activity as
IP
compared with IGF-1R-GFP (WT) and K1130R mutant. (** P<0.05, *** P<0.001). (B) Serum
CR
starved HEK-293 cell lysates were immunoblotted to determine the expression and
US
antibodies. (C) Reporter activity was determined upon co-transfection of TopFlash (TOP)
AN
or FopFlash (FOP) and Renilla plasmids into HepG2 cells along with IGF1R-GFP (WT) and
lysine mutants under basal and serum free conditions with GFP as control (material and methods).
M
The data shown represents the average ± SD of three independent experiments. (***P < 0.001)
ED
PT
Fig. 7. Distance between the αc -helix and N terminal domain reduced leading to a more compact
CE
structure after mutating the Lys1052 and induction of orderness in the activation loop resulted in the
loss of the autophosphorylation activity. (A) Distance between the αc-helix and N terminal was
AC
calculated from the simulation trajectory of 20 ns. The distance is 2.5 in the case of Wild type whereas
reduced to 1 nm in the case of K1052R mutant tyrosine. (B) Alignment of tertiary structures of wild type
(yellow colour) and K1052R (Cyan colour) mutant IGF1-R tyrosine kinase showed that the N terminal
domain of the K1052 mutant IGF1R tyrosine kinase has moved closer to the αc helix as compared to the
wild type IGF1R tyrosine Kinase and there is downward movement of the αc helix. The protein structures
are shown in the cartoon model and represented in different colours. (E) Alignment of tertiary structures of
30
ACCEPTED MANUSCRIPT
the wild type and K1052R mutant showing the changes in secondary structure contents of the activation
loop. Wild type (yellow) IGF1-R tyrosine kinase has disordered activation loop, while in the K1052R
mutant (blue) IGF1-R tyrosine kinase the activation loop develop two small beta strands and a partial helix.
(F) Tyr1158 residues of the K1052R mutant IGF1-R tyrosine kinase involved in the autophosphorylation
activity showing changes in the orientation of the side chain compared to the wild type. The residues are
T
shown in sticks model in yellow (wild type) and blue (K1052R mutant).
IP
CR
US
AN
M
ED
PT
CE
AC
31
ACCEPTED MANUSCRIPT
Conflict of interest
T
IP
CR
US
AN
M
ED
PT
CE
AC
32
ACCEPTED MANUSCRIPT
Highlights:
4. We show that cytoplasmic and nuclear activities are two independent functions of IGF1R.
T
IP
CR
US
AN
M
ED
PT
CE
AC
33
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
View publication stats Figure 7