T H E J O U R N A L O F C E L L B I O L O G Y
The Journal of Cell Biology, Vol. 168, No. 6, March 14, 2005 921–928
The Rockefeller University Press $8.00
Akt2 phosphorylates Synip to regulate docking and
fusion of GLUT4-containing vesicles
and Masatomo Mori
Department of Medicine and Molecular Science, Gunma University Graduate School of Medicine, Maebashi, Gunma, 371-8511, Japan
Health and Science Center, Gunma University, Maebashi, Gunma, 371-8510, Japan
Core Research for Evolutional Science and Technology, Japan Science and Technology Corporation, Kawaguchi, Saitama 332-0012, Japan
e have identified an unusual potential dual
Akt/protein kinase B consensus phosphoryla-
tion motif in the protein Synip (RxKxRS
Surprisingly, serine 97 is not appreciably phosphory-
lated, whereas serine 99 is only a specific substrate for
Akt2 but not Akt1 or Akt3. Although wild-type Synip
(WT-Synip) undergoes an insulin-stimulated dissociation
from Syntaxin4, the Synip serine 99 to phenylalanine
mutant (S99F-Synip) is resistant to Akt2 phosphorylation
and fails to display insulin-stimulated Syntaxin4 dissocia-
tion. Furthermore, overexpression of WT-Synip in 3T3L1
adipocytes had no effect on insulin-stimulated recruitment
of glucose transporter 4 (GLUT4) to the plasma mem-
brane, whereas overexpression of S99F-Synip functioned
in a dominant-interfering manner by preventing insulin-
stimulated GLUT4 recruitment and plasma membrane
fusion. These data demonstrate that insulin activation
of Akt2 specifically regulates the docking/fusion step
of GLUT4-containing vesicles at the plasma membrane
through the regulation of Synip phosphorylation and
Akt, also known as protein kinase B (PKB), is an important
regulator of several cellular processes, including proliferation,
metabolism, and programmed cell death (Lawlor and Alessi,
2001; Whiteman et al., 2002). Akt has three isoforms, Akt1
), Akt2 (PKB
), and Akt3 (PKB
ping but distinct cellular function. In particular, recent studies
have demonstrated that Akt1 plays an important role on
growth and antiapoptosis, whereas Akt2 functions primarily as
a regulator of glucose metabolism (Cho et al., 2001b; Bae et
al., 2003). For example, Akt1
but relatively normal glucose homeostasis, whereas Akt2
mice displayed insulin-resistant glucose metabolism in liver
and muscle (Cho et al., 2001a). However, the insulin resis-
tance is relatively mild and becomes significantly more pro-
nounced in conjunction with the loss of Akt1 (Jiang et al.,
2003). Consistent with a primary role for Akt2, a family with
severe insulin resistance and overt diabetes was mapped to a
single point mutation in Akt2 (George et al., 2004). Thus,
there seems to exist intracellular signal specificity and some
), each with overlap-
mice have reduced body size
compensation mechanism for the regulation of glucose metab-
olism between Akt1 and Akt2.
It is well documented that Akt is involved in the regula-
tion of glucose metabolism by inhibiting glycogen synthesis
through the inhibition of glycogen synthesis kinase 3 (GSK3)
activity (Cross et al., 1995; Coghlan et al., 2000; Doble and
Woodgett, 2003). However, how Akt regulates glucose uptake
(muscle and adipose tissue), its association with peripheral
insulin resistance, and the molecular basis for the apparent Akt2
specificity is still unknown.
Peptide substrate mapping studies have identified the
preferred Akt1 phosphorylation consensus site as RxRxxS/T
(Alessi et al., 1996). Currently, over 20 substrates for Akt have
been identified; however, none of these substrates has been
reported to exhibit Akt isoform selectivity. Thus, at present, the
molecular basis for the physiologic specificities of Akt isoform
function remains a fundamental issue that has yet to be resolved.
In this regard, we previously identified Synip as a Syntaxin4
interacting protein (Min et al., 1999). Under the basal conditions,
Synip was constitutively bound to Syntaxin4 and prevented the
interaction of Syntaxin4 with both SNAP23 (synaptosome-
associated proteins of 23 kD) and VAMP2 (vesicle-associated
membrane protein 2; Min et al., 1999). Insulin treatment resulted
in a dissociation of the Synip–Syntaxin4 complex allowing for
Correspondence to Shuichi Okada: firstname.lastname@example.org
Abbreviations used in this paper: GLUT4, glucose transporter 4; GSK3, glycogen
synthesis kinase 3; PI3, phosphatidylinositol 3; siRNA, small interfering RNA;
JCB • VOLUME 168 • NUMBER 6 • 2005922
the assembly of a fusogenic Syntaxin4–SNAP23–VAMP2
complex necessary for glucose transporter 4 (GLUT4) translo-
cation (Min et al., 1999).
In this paper, we now demonstrate that Synip is a pre-
ferred Akt2-specific substrate with an unusual dual consensus
phosphorylation site. The specific Akt2-dependent phosphory-
lation of serine 99 is essential for the insulin-stimulated disso-
ciation of Synip from Syntaxin4, translocation, and plasma
membrane fusion of GLUT4-containing vesicles.
Analysis of insulin signal-regulating
Insulin stimulates the translocation of GLUT4 proteins from
intracellular storage sites to the plasma membrane. To date,
two major insulin-mediated signal transduction pathways have
been implicated in the regulation of this process (Saltiel and
Pessin, 2003). The insulin activation and/or targeting of the
type 1A phosphatidylinositol 3 (PI3) kinase generate PI3, 4,
5P3 in the plasma membrane (Okada et al., 1994). PI3, 4, 5P3
recruits and/or activates phosphoinositide-dependent kinase 1
that serves as an immediate upstream kinase for Akt and the
atypical protein kinase C isoforms
1991; Belham et al., 1999). The plasma membrane transloca-
tion of GLUT4 requires the specific interaction of the plasma
membrane t-SNARE proteins Syntaxin4 and SNAP23 with the
v-SNARE protein VAMP2 in GLUT4-containing cargo vesi-
cles (Pessin et al., 1999).
To determine the specific insulin signaling pathway re-
sponsible for the dissociation of Synip from Syntaxin4, we
(Bellacosa et al.,
treated cells with various pharmacological agents and small in-
terfering RNA (siRNA; Fig. 1). As previously observed, after
insulin stimulation there was an
of Synip protein coimmunoprecipitated with Syntaxin4 (Fig. 1
A). Several studies have suggested that in addition to the PI3
kinase pathway a second PI3 kinase–independent insulin signal
pathway is necessary for the efficient translocation of GLUT4
in adipocytes (Watson et al., 2004). This latter pathway in-
volves the function of the Cbl adaptor protein CAP (Cbl-asso-
ciated protein) as expression of a dominant-interfering CAP
blocks insulin-stimulated GLUT4 translocation without signifi-
cant effect on insulin activation of PI3 kinase signaling and Akt
activation (Baumann et al., 2000). However, expression of a
dominant-interfering CAP mutant (CAP sorbin) had no signifi-
cant effect on the insulin-stimulated dissociation of Synip from
Syntaxin4 with also an
90% reduction in coimmunoprecipi-
tation (Fig. 1 B). Similarly, treatment with a high concentration
M) of PKC inhibitor GF109203X, which inhibits atypical
PKCs (Li et al., 1999), resulted in an
protein coimmunoprecipitated with Syntaxin4 after insulin
stimulation (Fig. 1 C).
In contrast, inhibition of PI3 kinase with Wortmannin or
by expression of a dominant-interfering p85 regulatory subunit
mutant prevented the insulin-stimulated Synip dissociation
(Fig. 1, D and E). Furthermore, reduction of Akt2 protein lev-
els with a specific Akt2 siRNA blocked the insulin-stimulated
dissociation of Synip from Syntaxin4, whereas reduction of
Akt1 protein levels by a specific Akt1 siRNA resulted in a 94%
insulin-stimulated decrease of Synip binding (Fig. 1, F and G).
Under these siRNA conditions, the protein levels of Akt1 and
Akt2 were reduced by more than 90% compared with untreated
90% decrease in the amount
85% decrease of Synip
Synip–Syntaxin4 interaction. FLAG-WT-Synip
is expressed in CHOIR (CHO cell with over-
expressed human insulin receptor) by elec-
troporation. After 48 h of recovery, 6 h of
serum starvation followed. 100 nM insulin
was added for 15 min to see Synip–Syntaxin4
interaction. After 4 mg of whole cell lysates
were made with NP-40 lysis buffer, 2 ?g of
Syntaxin4 antibody was added to precipitate
Syntaxin4. Coimmunoprecipitated FLAG-WT-
Synip was detected by FLAG immunoblotting.
In each experiment it was confirmed that same
amount of Syntaxin4 was immunoprecipi-
tated by Syntaxin4 immunoblotting, and
FLAG-WT-Synip expression was identical by
FLAG immunoblotting in each sample (not
depicted). Data show representative experi-
ments independently performed and each
experiment was repeated three to four times.
(A) A typical result of Synip–Syntaxin4 interac-
tion before and after insulin stimulation. Insulin
causes Synip dissociation from Syntaxin4 after
insulin stimulation, which is recognized as
weaker band intensity compared with basal
condition. It was statistically significant (*, P ? 0.01). (B) The effect of the disruption of CAP/TC10 signal (PI3 kinase–independent pathway) by the design
of sorbin overexpression on Synip–Syntaxin4 interaction. After insulin stimulation, Synip dissociation was still observed. The difference in band intensity
was statistically significant (*, P ? 0.01). (C) The effect of a pharmacological PKC inhibitor such as GF 109203X. Synip dissociation could be still
observed. The difference in band intensity was statistically significant (*, P ? 0.01). (D) The effect of pharmacological PI3 kinase inhibitor, such as Wortmannin,
pretreatment on Synip–Syntaxin4 interaction. Synip dissociation disappeared. (E) The effect of PI3 kinase negative-dominant overexpression. Synip dissociation
also disappeared. (F) The effect of siRNA of Akt1. Synip dissociation could be observed. The difference in band intensity was statistically significant
(*, P ? 0.01). (G) The effect of siRNA of Akt2. Synip dissociation could not be observed.
Analysis of insulin signal-regulating
IDENTIFICATION OF AKT2-SPECIFIC SUBSTRATE • YAMADA ET AL.923
cells or incubated with a random siRNA (unpublished data).
Together, these data strongly suggested that the PI3 kinase–
dependent pathway leading specifically to Akt2 activation is
responsible for the insulin-stimulated dissociation of Synip–
Syntaxin4 complex. Furthermore, these results indicate that
neither the CAP nor atypical protein kinase C pathways are in-
volved in this process.
Akt2 but not Akt1 or Akt3
Next, we determined whether or not Synip phosphorylation by
Akt2 was responsible for the regulation of the Synip–Syntaxin4
interaction. Inspection of the Synip amino acid sequence re-
vealed the presence of an overlapping dual Akt consensus phos-
phorylation site at serine 97 and 99 (RAKLRSESP) (Fig. 2 A).
Because arginine at position
5 is preferred for Akt1 (Alessi et
al., 1996; Obata et al., 2000), this suggests that serine 97 would
be a preferred Akt1 phosphorylation site. However, to date no
Akt2-specific consensus motif has been reported.
Therefore, we performed the following series of experi-
ments. First, we assessed the Akt-dependent phosphorylation
of Synip in cells using an Akt phosphospecific substrate anti-
body. Wild-type Synip (WT-Synip) or Synip serine 99 to phe-
nylalanine mutant (S99F-Synip) was expressed in CHOIR cells
and stimulated with insulin for 15 min. Synip was then immu-
noprecipitated, and phosphorylated signal was detected by
Western blotting with Akt phosphospecific substrate antibody.
As shown in Fig. 2 B, WT-Synip was clearly phosphorylated
after insulin stimulation. In contrast, S99F-Synip was unable to
undergo insulin-stimulated phosphorylation (Fig. 2 B). To as-
sess whether or not serine 99 is a specific Akt2 phosphoryla-
tion site, we examined the in vitro phosphorylation of synthetic
peptides corresponding to this region of Synip (Fig. 2 A). We
were unable to detect any significant Akt1-dependent phos-
phorylation of the WT-Synip peptide, the S99F mutant, the
S97F/S99F double mutant, or a scramble peptide. In contrast,
the WT-Synip peptide was an effective substrate for recombi-
nant Akt2 (Fig. 2 C). As controls, both a scramble peptide with
the same composition as well as mutation of the equivalent po-
sition of serine 99 to phenylalanine markedly reduced Akt2-
dependent phosphorylation (Fig. 2 C). In addition, Akt2 did not
phosphorylate the peptide, which has mutation of the equiva-
(A) Candidate Akt phosphorylation consensus
motif in Synip is presented. (B) Either FLAG-
WT-Synip or FLAG-S99F-Synip was expressed
in CHOIR cells by electroporation. After 15
min of insulin stimulation, Synip was immuno-
precipitated by FLAG antibody and samples
were separated on SDS-PAGE. Phosphorylated
Synip signal was analyzed by Akt phospho-
specific substrate antibody immunoblotting.
These are representative experiments indepen-
dently performed three times. (C) Synthesized
oligo peptides were phosphorylated as de-
scribed in Materials and methods. These were
repeated four to five times. Recombinant Akt2
was capable of phosphorylating the Synip
peptide compared with Akt1 (*, P ? 0.01). A
scramble peptide with the same composition
as well as mutation of the equivalent position
of serine 99 to phenylalanine markedly reduced
Akt2-dependent phosphorylation (*, P ?
0.01). (D) GST-Synip fusion proteins were
phosphorylated as described in Materials
and methods. These were repeated five times.
Recombinant Akt2 was capable of phosphory-
lating GST-WT-Synip (*, P ? 0.05). However,
Akt2 did not phosphorylate GST-S99F-Synip.
GST alone showed negligible count. (E) Akt
and Synip interaction was estimated by the
design of coimmunoprecipitation experiments.
FLAG-Synip was expressed in CHO/IR cells
by electroporation and either Akt1 or Akt2
was immunoprecipitated by specific antibody.
Synip amount associated with Akt was estimated
by FLAG immunoblot. These were repeated
five times. Synip significantly associates with
not Akt1 but Akt2 after insulin stimulation
(*, P ? 0.01).
Synip phosphorylation by Akt2.
JCB • VOLUME 168 • NUMBER 6 • 2005 924
lent position of serine 97 and 99 to phenylalanine. Similarly,
the activated form of purified recombinant Akt3 did not phos-
phorylate the Synip peptide (unpublished data). Confirming the
functional activity of the recombinant Akt1, Akt2, and Akt3,
all three isoforms were fully capable of phosphorylating a
GSK3 control peptide (Fig. 2 C and not depicted).
Because these data demonstrate that the RAKLRSESP
Synip peptide sequence displays Akt2-specific substrate speci-
phosphatase first as described previously (Zhao et al., 1998) to dephosphorylate all the phosphorylated molecules. Thereafter, an ordinal coimmunoprecipitation
experiment was conducted to see Synip–Syntaxin4 interaction. These are representative experiments independently performed two times. (B) Either
GST-WT-Synip or GST-S99F-Synip was phosphorylated by Akt2 first and then incubated with MBP-Syntaxin4. Samples were separated on SDS-PAGE and
GST-Synip associated with MBP-Syntaxin4 was estimated by GST immunoblot (*, P ? 0.01). These were repeated five times. In addition, we stripped
the GST antibody from filters and reprobed them with phosphospecific Akt substrate antibody. We confirmed that there was no signal (not depicted) and
that nonphosphorylated Synip binds to Syntaxin4. (C–E) FLAG-WT-Synip or FLAG-S99F-Synip or FLAG-S97F-Synip was electroporated to CHOIR cells and
3T3L1 adipocytes and the cells stimulated with or without insulin for 15 min. In these experiments, the amount of expressed Synip protein was adjusted to
match the expression levels in 3T3L1 adipocytes that have a lower level of transfection efficiency and protein expression than CHOIR cells (Min et al.,
1999). In comparison to Fig. 1, 8 mg of detergent whole cell lysates were immunoprecipitated with 4 ?g of Syntaxin4 antibody. Coimmunoprecipitated
Synip amount was estimated by FLAG immunoblotting. FLAG-WT-Synip and FLAG-S97F-Synip dissociation could be observed in CHOIR cells and 3T3L1
adipocytes. The difference in band intensity was statistically significant (*, P ? 0.05). These are representative experiments independently performed four
times. (F and G) FLAG-WT-Synip or FLAG-S99F-Synip was electroporated to CHOIR. After Syntaxin4 was immunoprecipitated, coimmunoprecipitated
Synip amount was estimated by FLAG immunoblotting. FLAG-WT-Synip dissociation appeared 5 min after insulin stimulation and continued up to 30 min.
The difference in band intensity was statistically significant (*, P ? 0.05; **, P ? 0.01). These are representative experiments independently performed
three times. In this design, insulin stimulation was done at 5, 15, and 30 min.
Synip phosphorylation is required for Synip dissociation from Syntaxin4. (A) In this experiment whole cell lysates were pretreated with alkaline
IDENTIFICATION OF AKT2-SPECIFIC SUBSTRATE • YAMADA ET AL.925
ficity, we examined if serine 99 is the critical Akt2 substrate site
in the full-length Synip protein. GSK3 control peptide and the
full-length Synip proteins expressed as GST-fusions were phos-
phorylated in vitro. As expected, GSK3 control peptide was an
effective substrate for both Akt1 and Akt2, whereas the full-
length Synip protein was only a functional substrate for Akt2
(Fig. 2 D). Consistent with a required role for S99, the Synip
S99F mutant was not phosphorylated by either Akt1 or Akt2.
Although the mechanism of Akt substrate recognition is
not known, if it is through the active site then an increased
binding of Synip to Akt2 compared with Akt1 would be ex-
pected. To test this possibility, we examined the coimmunopre-
cipitation of Synip with Akt1 and Akt2 (Fig. 2 E). The immu-
noprecipitation of Akt1 resulted in the coimmunoprecipitation
of Synip but this interaction was unaffected by insulin stimula-
tion. In contrast, insulin stimulation resulted in a marked in-
crease in the amount of Synip that was coimmunoprecipitated
with Akt2. Together, these data provide compelling evidence
that the Synip serine 99 residue is a selective Akt2 phosphory-
lation site and that the RAKLRSESP sequence context might
provide an Akt2-specific consensus motif.
Phosphorylation of serine 99 is required
for reduced Synip–Synatxin4 binding
To determine the functional consequence of this phosphoryla-
tion event, we nonspecifically dephosphorylated the Synip pro-
tein with alkaline phosphatase before coimmunoprecipitation
(Fig. 3 A). Under these conditions, Synip failed to display an in-
sulin-stimulated dissociation from Syntaxin4, suggesting non-
phosphorylated Synip preferentially binds to Syntaxin4. Consis-
tent with a complete Synip dephosphorylation, we did not detect
any significant difference of Synip–Syntaxin4 complex amount
between unstimulated and insulin-stimulated samples. To di-
rectly assess the effect of Synip phosphorylation, we examined
the in vitro binding of GST-Synip with MBP (maltose binding
protein)-Syntaxin4 after Akt2 phosphorylation (Fig. 3 B). In the
absence of Akt2, both WT-Synip and S99F displayed identical
extents of binding. In contrast, Akt2-dependent phosphorylation
of WT-Synip resulted in a 70% reduction in Syntaxin4 binding,
whereas there was no effect on the S99F mutant.
As expected for site-specific phosphorylation, both the
wild type and a serine 97 to phenylalanine (S97F) mutant dis-
played the typical insulin-stimulated dissociation from Syntaxin4
in both CHO/IR and 3T3L1 adipocytes (Fig. 3, C and D). In con-
trast, the serine 99 to phenylalanine (S99F) mutant was refrac-
tory to insulin-stimulated dissociation (Fig. 3 E). In fact, whereas
5 min of insulin stimulation was sufficient to induce WT-Synip
dissociation from Syntaxin4 (Fig. 3 F), S99F-Synip was persis-
tently associated at least over the 30 min time course examined
(Fig. 3 G). Insulin-stimulated WT-Synip dissociation continued
up to 30 min after insulin stimulation (Fig. 3 F). These data are
consistent with phosphorylation of serine 99 but not serine 97 as
a necessary step in the insulin-stimulated dissociation of the
Synip–Syntaxin4 complex. Together these data provide compel-
ling evidence that the RxKxRSxS motif is essential for Synip
phosphorylation and dissociation from Syntaxin4 by Akt2.
uptake and GLUT4 translocation. (A) In 3T3L1 adipo-
cytes, pcDNA3, FLAG-WT-Synip, FLAG-S99F-Synip,
or FLAG-S97F-Synip was introduced by electroporation
(Min et al., 1999). After 48 h of recovery, serum was
removed for 6 h. Cells were stimulated by insulin for
30 min. 3H-labeled 2-deoxyglucose uptake was mea-
sured as described in Materials and methods. FLAG-
S99F-Synip significantly inhibited 3H-labeled 2-deoxy-
glucose uptake (*, P ? 0.05). (B) In 3T3L1 adipo-
cytes, eGFP-GLUT4 and either pcDNA3, FLAG-WT-
Synip, FLAG-S99F-Synip, or FLAG-S97F-Synip were
introduced by electroporation (Min et al., 1999). After
48 h of recovery, serum was removed for 6 h. Then,
cells were stimulated by insulin for 30 min. EGFP signal
was detected by confocal microscopy, and rim for-
mation was considered as completely translocated
eGFP-GLUT4 in this assay. FLAG-S99F-Synip signifi-
cantly inhibited GLUT4 translocation (*, P ? 0.05).
Experiments were repeated four times. (C) In 3T3L1
adipocytes, eGFP-myc-GLUT4 and either pcDNA3,
FLAG-WT-Synip, FLAG-S99F-Synip, or FLAG-S97F-
Synip were introduced by electroporation (Min et al.,
1999). Myc was inserted in the second loop of
GLUT4 to expose to the outside of the plasma mem-
brane after insulin stimulation. When myc signal was
detected, docking/fusion step was completed and
GLUT4 translocation was accomplished. Myc signal
was estimated as described in Materials and methods.
FLAG-S99F-Synip significantly inhibited GLUT4 trans-
location (*, P ? 0.05). Experiments were repeated
Synip phosphorylation is required for glucose
JCB • VOLUME 168 • NUMBER 6 • 2005 926
Synip regulates the docking/fusion steps
of GLUT4-containing vesicles by Akt2
Having established a biochemical consequence of Synip phos-
phorylation, we next examined the physiological requirement
of this event. Expression of WT-Synip or S97F-Synip in
3T3L1 adipocytes had no significant effect on insulin-stimu-
lated glucose uptake compared with empty vector-transfected
cells (Fig. 4 A). However, expression of S99F-Synip signifi-
cantly suppressed insulin-stimulated glucose uptake. To deter-
mine if the reduction of insulin-stimulated glucose uptake was
due to an inhibition of GLUT4 translocation, we compared the
cell surface localization of GLUT4 by confocal fluorescent mi-
croscopy (Fig. 4 B). Expression of S99F-Synip but not WT-
Synip or S97F-Synip reduced insulin-stimulated GLUT4 trans-
location compared with control cells (Fig. 4 B). Finally, a
quantitative colorimetric assay for the translocation of a myc
epitope-tagged GLUT4 (Konrad et al., 2002) demonstrated that
expression of S99F-Synip reduced insulin-stimulated GLUT4
translocation by 52% compared with control, WT-Synip, and
S97F-Synip–expressing cells (Fig. 4 C). In this particular as-
say, because the myc epitope tag is located on the extracellu-
lar loop between transmembrane domains 1 and 2, only the
GLUT4 proteins that have undergone plasma membrane fusion
It is well recognized that the interaction of Syntaxin4 as
t-SNARE and VAMP2 as v-SNARE is necessary for insulin-
stimulated GLUT4 translocation in adipose tissue and skeletal
muscle (Foster and Klip, 2000). For example, overexpression
of the Syntaxin4 cytoplasmic domain inhibits insulin-stimu-
lated GLUT4 translocation (Olson et al., 1997). This inhibition
was specific for the VAMP2 binding domain within Syntaxin4
because deletion of this region did not result in inhibition of in-
sulin-stimulated GLUT4 translocation (Olson et al., 1997).
Thus, Syntaxin4 and VAMP2 binding is one of the necessary
steps for GLUT4 translocation. However, it is still unknown
about precise mechanism how insulin controls Syntaxin4–
VAMP2 interaction directly. In this regard, Synip is a Syn-
taxin4 specific binding protein that regulates the interaction be-
tween Syntaxin4 and VAMP2 in an insulin-dependent manner
(Min et al., 1999). Insulin causes Synip dissociation from
Syntaxin4, and this dissociation allows VAMP2 to bind with
Syntaxin4 because Synip and VAMP2 use the same binding
site on Syntaxin4 (Min et al., 1999). However, it has not yet
been determined how insulin causes Synip dissociation from
Syntaxin4. Therefore, we attempted to identify the Synip disso-
ciation mechanism to understand how insulin regulates the
Syntaxin4–VAMP2 interaction and docking/fusion step of
PI3 kinase–independent pathway regulated by CAP/Cbl/
TC10 molecules is recently reported to regulate the final steps
of GLUT4 exocytosis by recruiting Exo70 to the plasma mem-
brane (Inoue et al., 2003). However, as shown in Fig. 1 B,
CAP-sorbin (Baumann et al., 2000), which has a dominant-
negative effect for the CAP/Cbl/TC10 pathway, did not affect
Synip dissociation from Syntaxin4. Furthermore, our data
strongly suggested that the PI3 kinase–Akt pathway specifi-
cally regulates Synip–Syntaxin4 interaction and that the PI3 ki-
nase–atypical PKC pathway is not likely involved (Fig. 1,
C–G). Furthermore, our siRNA experiments strongly sug-
gested that Akt2 but not Akt1 regulates Synip–Syntaxin4 inter-
action (Fig. 1, F and G).
The next question is why Akt2 alone is involved in the
regulation of Synip–Syntaxin4 interaction? What kind of
mechanism does determine the Akt specificity in this pathway?
As shown in Fig. 2 A, Synip seemed to have a consensus motif
including the 99
serine residue as Akt phosphorylation site.
Therefore, we tested the possibility of whether or not Akt phos-
phorylates the 99
serine of Synip. We have observed that
Akt2, but not Akt1, specifically phosphorylates Synip at serine
99 residue after insulin stimulation (Fig. 2 B). However, be-
cause it was formally possible that this was an indirect phos-
phorylation event, we synthesized the WT-Synip peptide, the
S99F Synip mutant, the S97F/S99F Synip double mutant, or a
scramble peptide including the serine 99 and performed in vitro
phosphorylation experiments with recombinant activated forms
of Akt1, Akt2, and Akt3. Surprisingly, only Akt2 was capable
of directly phosphorylating the WT-Synip peptide (Fig. 2 C).
Although serine 97 is also in an appropriate context as an Akt1
substrate, neither Akt1, Akt2, nor Akt3 was capable of phos-
phorylating the S99F Synip mutant, the S97F/S99F Synip dou-
ble mutant, or a scramble peptide. Moreover, the similar
substrate selectivity was observed in the case of in vitro phos-
phorylation using full-length WT-Synip and S99F-Synip (Fig.
2 D). Currently, there are more than 20 reported substrates for
Akt, yet none of them have been shown to display Akt isoform
specificity. It is generally assumed that either Akt and/or sub-
strate spatial compartmentalization is the primary determinant
of substrate selectivity. Although our data do not exclude this
mechanism, our data indicate the presence of at least one un-
usual consensus site that displays relative Akt2 specificity in a
physiologically regulated manner.
As Synip is a Syntaxin4 binding protein, its functional ac-
tivity is assumed to exist at the plasma membrane. Although in-
sulin-stimulated Akt2 translocates to the plasma membrane
(Hanada et al., 2004), Akt1 was reported to undergo nuclear lo-
calization after growth factor stimulation (Pekarsky et al.
As shown in Fig. 2 E, insulin stimulation resulted in a specific
increase in Akt2 but not Akt1 association with Synip. Whether
or not the Synip–Akt2 interaction occurs through direct or indi-
rect interaction, these results are consistent with the specificity
and subcellular compartmentalization of Akt1 and Akt2.
Our data also provide a physiological consequence for
Akt2-specific Synip phosphorylation. As shown in Fig. 3 A,
the dephosphorylated Synip binds Syntaxin4, whereas Akt2
phosphorylated Synip displayed significantly less binding com-
pared with nonphosphorylated Synip (Fig. 3 B). Furthermore,
S99F-Synip binding to Syntaxin4 was refractory to insulin
stimulation and was unaffected by the presence or absence of
active Akt2 (Fig. 3 B). Consistent with these results, only
Synip, which has an Akt2 phosphorylation site, was permissive
IDENTIFICATION OF AKT2-SPECIFIC SUBSTRATE • YAMADA ET AL. 927
for insulin-stimulated glucose uptake and GLUT4 transloca-
tion, whereas the nonphosphorylatable mutant S99F was in-
hibitory. Thus, our data demonstrate that insulin-stimulated
Akt2-dependent phosphorylation of Synip on serine residue 99
results in reduced binding interactions between Synip and
Syntaxin4. These data are consistent with the requirement of
Akt2 for insulin-stimulated glucose uptake and insulin sensitiv-
ity in culture cells, mice, and humans (Cho et al., 2001a; Bae et
al., 2003; George et al., 2004).
In summary, we have identified Synip as the first specific
Akt2 substrate. Akt2-dependent phosphorylation at serine 99
directly modulates the interaction of Synip with Syntaxin4,
suggesting a direct mechanism linking Akt2 function with the
t-SNARE–mediated docking/fusion of GLUT4 cargo vesicles.
Materials and methods
The FLAG M2 mAb was obtained from Sigma-Aldrich, and Syntaxin4
sheep polyclonal and Synip rabbit polyclonal antibodies were prepared
and affinity purified as described previously (Olson et al., 1997; Min et
al., 1999). Phosphospecific Akt substrate antibody was purchased from
Cell Signaling Technology, and Akt1- and Akt2-specific antibodies and
siRNA for Akt1 and Akt2 were obtained from Upstate Cell Signaling Solu-
tions. Although the Akt substrate phosphospecific antibody can recognize
the RxRxxS/T motif, it apparently has a sufficient cross reactivity to the
KxRxxS sequence present in Synip. GST protein expression system includ-
ing GST antibody was obtained from Amersham Biosciences. MBP expres-
sion system was obtained from New England Biolabs, Inc. ECL
Western Blotting Detection System and
from Amersham Biosciences. The anti–sheep and anti–rabbit IgG-HRP
were obtained from Pierce Chemical Co. The VAMP2 antibody was pur-
chased from Calbiochem, and cell culture media and reagents were pur-
chased from Life Technologies. Synthesis of oligo-peptides was done by
Shimazu Corporation. All of other chemicals used in this study were pur-
chased from Sigma-Aldrich.
H-2-deoxyglucose were obtained
CHO cells expressing the human insulin receptor (CHO/IR) were obtained
as described previously (Waters et al., 1995). These cells were maintained
in minimal Eagle’s medium containing 10% FBS at 37
3T3L1 preadipocytes were cultured in DME containing 25 mM glucose,
10% calf serum at 37
C with 8% CO
differentiate into adipocytes as described previously (Min et al., 1999).
C with 5% CO
. Confluent cultures were induced to
Scraped frozen cells were rocked for 10 min at 4
buffer (25 mM Hepes, pH 7.4, 10% glycerol, 50 mM sodium fluoride, 10
mM sodium phosphate, 137 mM sodium chloride, 1 mM sodium ortho-
vanadate, 1 mM PMSF, 10
g/ml aprotinin, 1
g/ml leupeptin). Insoluble material was separated from the soluble ex-
tract by centrifugation for 10 min at 4
the supernatant was determined by BCA method. After the addition of 4.5
g of antibody to the whole cell lysates, samples (typically 2–3 mg of ly-
sates) were incubated for 2 h at 4
Cruz Biotechnology, Inc.) was added and samples were consequently
rocked during the next 1 h at 4
C. After the incubation, samples were ex-
tensively washed three times with the NP-40 lysis buffer. The washed sam-
ples were resuspended in SDS sample buffer (125 mM Tris-HCl, pH 6.8,
20% [vol/vol] glycerol, 4% [wt/vol] SDS, 100 mM DTT, and 0.1% [wt/
vol] Bromophenol blue) and heated at 100
C with NP-40 lysis
g/ml pepstatin, and 5
C, and the total protein amount in
l of protein A–agarose (Santa
C for 5 min.
Samples were separated by SDS-PAGE and electrophoretically transferred
to polyvinylidene difluoride membranes. The samples were immunoblotted
with monoclonal or polyclonal specific antibody as indicated in the figures
To assess Synip–Syntaxin4 interaction, 90% of samples were ap-
plied on SDS-PAGE and subjected to FLAG immunoblotting to detect
FLAG-Synip signal. For the estimation of immunoprecipitated Syntaxin4
amount, the rest of 10% immunoprecipitated samples were applied on a
different SDS-PAGE and subjected to Syntaxin4 immunoblotting. Although
figures were not presented, FLAG-Synip and Syntaxin4 signal from whole
cell lysate samples were checked and it was confirmed that electropora-
tion efficiency was the same among each sample and Syntaxin4 expres-
sion was not affected by sample treatment in each experiment. The pri-
mary monoclonal and polyclonal antibodies for immunoblottings were
detected with HRP-conjugated anti–mouse or anti–rabbit IgG antibodies
(Pierce Chemical Co.). Specific signals were visualized by ECL
Western Blotting Detection System (Amersham Biosciences). Each band
was developed to X-ray film from Amersham Biosciences and scanned by
UMAX MagicScan 4.3 system.
Transfection of CHO/IR and 3T3L1 adipocytes
CHO/IR cells were quantitatively transfected by electroporation as previ-
ously described (Waters et al., 1995). 3T3L1 adipocytes were put into
suspension by mild trypsinization and electroporated with a total of 600
g of plasmid under low-voltage condition (0.16 kV, 950
were allowed to adhere to collagen-coated tissue culture dishes for 30–48 h,
and the adipocytes were then serum starved for 2 h before incubation in
the absence or presence of 100 nM insulin for 15 min at 37
case of cotransfection of siRNA and plasmids, 2 nmol siRNA was used.
F). The cells
C. In the
In vitro Synip phosphorylation
Akt substrate phosphorylation assay kit was purchased from Upstate Cell
Signal Solutions and we followed the manufacture’s procedure. 200 mU
of recombinant Akts were incubated with 2.5 mM of synthesized oligo-
Synip peptides and 4.3
M of GST-Synip fusion proteins with supplied ki-
nase reaction buffer. The kinase reaction was stopped by the addition of
ice cold TCA and applied on phosphocellulose paper. After extensive
washing of phosphocellulose paper, scintillation cocktail was added to the
phosphocellulose paper and the radiation activity was measured.
EGFP-GLUT4 translocation assay
g of eGFP-GLUT4 plasmid with 550
electroporated to 3T3L1 adipocytes. The cells were allowed to adhere to
collagen-coated tissue culture dishes for 30–48 h, and the adipocytes
were serum starved for 2 h before incubation in the absence or presence
of 100 nM insulin for 15 min at 37
washed in PBS and fixed for 10 min in PBS containing 4% PFA and 0.2%
Triton X-100. The samples were mounted on glass slides with Fluorescent
Mounting Medium (DakoCytomation). Cells were imaged using a confo-
cal fluorescence microscope (model MRC-1024; Bio-Rad Laboratories).
g of interesting plasmids was
C. Transfected adipocytes were
Colorimetric assay and 2-deoxyglucose uptake
In the case of colorimetric assay, 50
g of interesting plasmids was electroporated to 3T3L1 adipocytes.
2-Deoxyglucose uptake and colorimetric assay were performed as de-
scribed previously (Olson et al., 1997; Konrad et al., 2002; Brozinick et
g of eGFP-cMyc-GLUT4 with 550
Each band of Western blots was scanned and analyzed by Molecular Im-
ager FX (Bio-Rad Laboratories). We set and used the same size rectangle
box to surround each band and analyzed the intensity of each band alone
by the program in Molecular Imager FX. Before we calculated the band in-
tensity, we subtracted the background intensity, which is estimated from
the band-free area by using the same rectangle box.
All values are expressed as mean
tistical significance by analysis of variance and
SEM. Data were evaluated for sta-
t test with InStat 2 program.
We would like to thank Dr. Jeffrey E. Pessin (State University of New York at
Stony Brook, Stony Brook, NY) for critical suggestions on our manuscript.
Submitted: 31 August 2004
Accepted: 25 January 2005
Alessi, D.R., F.B. Caudwell, M. Andjelkovic, B.A. Hemmings, and P. Cohen.
1996. Molecular basis for the substrate specificity of protein kinase B;
comparison with MAPKAP kinase-1 and p70 S6 kinase.
Bae, S.S., H. Cho, J. Mu, and M.J. Birnbaum. 2003. Isoform-specific regulation
of insulin-dependent glucose uptake by Akt/protein kinase B.
JCB • VOLUME 168 • NUMBER 6 • 2005928
Baumann, C.A., V. Ribon, M. Kanzaki, D.C. Thurmond, S. Mora, S. Shige-
matsu, P.E. Bickel, J.E. Pessin, and A.R. Saltiel. 2000. CAP defines a
second signalling pathway required for insulin-stimulated glucose
Belham, C., S. Wu, and J. Avruch. 1999. Intracellular signalling: PDK1—a kinase
at the hub of things.
Bellacosa, A., J.R. Testa, S.P. Staal, and P.N. Tsichlis. 1991. A retroviral onco-
gene, akt, encoding a serine-threonine kinase containing an SH2-like
Brozinick, J.T., Jr., B.R. Roberts, and G.L. Dohm. 2003. Defective signaling
through Akt-2 and -3 but not Akt-1 in insulin-resistant human skeletal
muscle: potential role in insulin resistance.
Cho, H., J. Mu, J.K. Kim, J.L. Thorvaldsen, Q. Chu, E.B. Crenshaw III, K.H.
Kaestner, M.S. Bartolomei, G.I. Shulman, and M.J. Birnbaum. 2001a.
Insulin resistance and a diabetes mellitus-like syndrome in mice lacking
the protein kinase Akt2 (PKB beta).
Cho, H., J.L. Thorvaldsen, Q. Chu, F. Feng, and M.J. Birnbaum. 2001b. Akt1/
PKBalpha is required for normal growth but dispensable for mainte-
nance of glucose homeostasis in mice.
Coghlan, M.P., A.A. Culbert, D.A. Cross, S.L. Corcoran, J.W. Yates, N.J.
Pearce, O.L. Rausch, G.J. Murphy, P.S. Carter, L.R. Cox, et al. 2000. Se-
lective small molecule inhibitors of GSK-3 modulate glycogen metabo-
lism and gene transcription.
Cross, D.A., D.R. Alessi, P. Cohen, M. Andjelkovich, and B.A. Hemmings.
1995. Inhibition of glycogen synthase kinase-3 by insulin mediated by
protein kinase B.
Doble, B.W., and J.R. Woodgett. 2003. GSK-3: tricks of the trade for a multi-
J. Cell Sci.
Foster, L.J., and A. Klip. 2000. Mechanism and regulation of GLUT-4 vesicle fu-
sion in muscle and fat cells. Am. J. Physiol. Cell Physiol. 279:C877–C890.
George, S., J.J. Rochford, C. Wolfrum, S.L. Gray, S. Schinner, J.C. Wilson,
M.A. Soos, P.R. Murgatroyd, R.M. Williams, C.L. Acerini, et al. 2004.
A family with severe insulin resistance and diabetes due to a mutation in
AKT2. Science. 304:1325–1328.
Hanada, M., J. Feng, and B.A. Hemmings. 2004. Structure, regulation and func-
tion of PKB/AKT—a major therapeutic target. Biochim. Biophys. Acta.
Inoue, M., L. Chang, J. Wang, S.H. Chiang, A.R. Saltiel. 2003. The exocyst
complex is required for targeting of Glut4 to the plasma membrane by
insulin. Nature. 422:629–633.
Jiang, Z.Y., Q.L. Zhou, K.A. Coleman, M. Chouinard, Q. Boese, and M.P.
Czech. 2003. Insulin signaling through Akt/protein kinase B analyzed by
small interfering RNA-mediated gene silencing. Proc. Natl. Acad. Sci.
Konrad, D., P.J. Bilan, Z. Nawaz, G. Sweeney, W. Niu, Z. Liu, C.N. Antonescu,
A. Rudich, and A. Klip. 2002. Need for GLUT4 activation to reach max-
imum effect of insulin-mediated glucose uptake in brown adipocytes iso-
lated from GLUT4myc-expressing mice. Diabetes. 51:2719–2726.
Lawlor, M.A., and D.R. Alessi. 2001. PKB/Akt: a key mediator of cell prolifer-
ation, survival and insulin responses? J. Cell Sci. 114:2903–2910.
Li, D., G. Sweeney, Q. Wang, and A. Klip. 1999. Participation of PI3K and
atypical PKC in Na?-K?-pump stimulation by IGF-I in VSMC. Am. J.
Min, J., S. Okada, M. Kanzaki, J.S. Elmendorf, K.J. Coker, B.P. Ceresa, L.J.
Syu, Y. Noda, A.R. Saltiel, and J.E. Pessin. 1999. Synip: a novel insulin-
regulated syntaxin 4-binding protein mediating GLUT4 translocation in
adipocytes. Mol. Cell. 3:751–760.
Obata, T., M.B. Yaffe, G.G. Leparc, E.T. Piro, H. Maegawa, A. Kashiwagi,
R. Kikkawa, and L.C. Cantley. 2000. Peptide and protein library
screening defines optimal substrate motifs for AKT/PKB. J. Biol.
Okada, T., Y. Kawano, T. Sakakibara, O. Hazeki, and M. Ui. 1994. Essential
role of phosphatidylinositol 3-kinase in insulin-induced glucose trans-
port and antilipolysis in rat adipocytes. Studies with a selective inhibitor
wortmannin. J. Biol. Chem. 269:3568–3573.
Olson, A.L., J.B. Knight, and J.E. Pessin. 1997. Syntaxin 4, VAMP2, and/or
VAMP3/cellubrevin are functional target membrane and vesicle SNAP
receptors for insulin-stimulated GLUT4 translocation in adipocytes. Mol.
Cell. Biol. 17:2425–2435.
Pekarsky, Y.Y., A. Koval, C. Hallas, R. Bichi, M. Tresini, S. Malstrom, G.
Russo, P. Tsichlis, and C.M. Croce. 2000. Tcl1 enhances Akt kinase ac-
tivity and mediates its nuclear translocation. Proc. Natl. Acad. Sci. USA.
Pessin, J.E., D.C. Thurmond, J.S. Elmendorf, K.J. Coker, and S. Okada. 1999.
Molecular basis of insulin-stimulated GLUT4 vesicle trafficking. Location!
Location! Location! J. Biol. Chem. 274:2593–2596.
J. Biol. Chem.
Saltiel, A.R., and J.E. Pessin. 2003. Insulin signaling in microdomains of the
plasma membrane. Traffic. 4:711–716.
Waters, S.B., K. Yamauchi, and J.E. Pessin. 1995. Insulin-stimulated disassoci-
ation of the SOS-Grb2 complex. Mol. Cell. Biol. 15:2791–2799.
Watson, R.T., M. Kanzaki, and J.E. Pessin. 2004. Regulated membrane traffick-
ing of the insulin-responsive glucose transporter 4 in adipocytes. Endocr.
Whiteman, E.L., H. Cho, and M.J. Birnbaum. 2002. Role of Akt/protein kinase B
in metabolism. Trends Endocrinol. Metab. 13:444–451.
Zhao, H., S. Okada, J.E. Pessin, G.A. Koretzky. 1998. Insulin receptor-mediated
dissociation of Grb2 from Sos involves phosphorylation of Sos by ki-
nase(s) other than extracellular signal-regulated kinase. J. Biol. Chem.