Effect of protein kinase A on accumulation of brefeldin A-inhibited guanine nucleotide-exchange protein 1 (BIG1) in HepG2 cell nuclei.

Abstract

Brefeldin A-inhibited guanine nucleotide-exchange proteins, BIG1 and BIG2, are activators of ADP-ribosylation factor GTPases that are essential for regulating vesicular traffic among intracellular organelles. Biochemical analyses and immunofluorescence microscopy demonstrated BIG1 in nuclei as well as membranes and cytosol of serum-starved HepG2 cells. Within 20 min after addition of 8-Br-cAMP, BIG1 accumulated in nuclei, and this effect was blocked by protein kinase A (PKA) inhibitors H-89 and PKI, suggesting a dependence on PKA-catalyzed phosphorylation. BIG2 localization was not altered by cAMP, nor did BIG2 small interfering RNA influence nuclear accumulation of BIG1 induced by cAMP. Mutant BIG1 (S883A) in which Ala replaced Ser-883, a putative PKA phosphorylation site, did not move to the nucleus with cAMP addition, whereas replacement with Asp (S883D) resulted in nuclear accumulation of BIG1 without or with cAMP exposure, consistent with the mechanistic importance of a negative charge at that site. Mutation (712KPK714) of the nuclear localization signal inhibited BIG1 accumulation in nuclei, and PKA-catalyzed phosphorylation of S883, although necessary, was not sufficient for nuclear accumulation, as shown by the double mutation S883D/nuclear localization signal. A role for microtubules in cAMP-induced translocation of BIG1 is inferred from its inhibition by nocodazole. Thus, two more critical elements of BIG1 molecular structure were identified, as well as the potential function of microtubules in a novel PKA effect on BIG1 translocation.

Full text

Page 1

Effect of protein kinase A on accumulation of
brefeldin A-inhibited guanine nucleotide-exchange
protein 1 (BIG1) in HepG2 cell nuclei
Carmen Citterio*, Heather D. Jones, Gustavo Pacheco-Rodriguez, Aminul Islam, Joel Moss, and Martha Vaughan*
Pulmonary–Critical Care Medicine Branch, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD 20892
Contributed by Martha Vaughan, December 12, 2005
Brefeldin A-inhibited guanine nucleotide-exchange proteins, BIG1
and BIG2, are activators of ADP-ribosylation factor GTPases that
are essential for regulating vesicular traffic among intracellular
organelles. Biochemical analyses and immunofluorescence micros-
copy demonstrated BIG1 in nuclei as well as membranes and
cytosol of serum-starved HepG2 cells. Within 20 min after addition
of 8-Br-cAMP, BIG1 accumulated in nuclei, and this effect was
blocked by protein kinase A (PKA) inhibitors H-89 and PKI, sug-
gesting a dependence on PKA-catalyzed phosphorylation. BIG2
localizationwasnotalteredbycAMP,nordidBIG2smallinterfering
RNA influence nuclear accumulation of BIG1 induced by cAMP.
Mutant BIG1 (S883A) in which Ala replaced Ser-883, a putative PKA
phosphorylation site, did not move to the nucleus with cAMP
addition, whereas replacement with Asp (S883D) resulted in nu-
clear accumulation of BIG1 without or with cAMP exposure, con-
sistent with the mechanistic importance of a negative charge at
that site. Mutation (712KPK714) of the nuclear localization signal
inhibited BIG1 accumulation in nuclei, and PKA-catalyzed phos-
phorylation of S883, although necessary, was not sufficient for
nuclear accumulation, as shown by the double mutation S883D?
nuclear localization signal. A role for microtubules in cAMP-in-
duced translocation of BIG1 is inferred from its inhibition by
nocodazole. Thus, two more critical elements of BIG1 molecular
structure were identified, as well as the potential function of
microtubules in a novel PKA effect on BIG1 translocation.
ADP-ribosylation factor ? protein trafficking ? A kinase-anchoring protein
M
roles in the formation of membrane trafficking vesicles. These
vesicles bud from a donor membrane and fuse with a target
membrane to deliver cargo molecules (1–3). Conversion of ARF
from a cytosolic GDP-bound inactive state to a GTP-bound
active form that is tightly membrane-associated is accelerated by
guanine nucleotide-exchange factors (GEFs) (4). All known
ARF GEFs contain a Sec7 domain of ?200 aa that catalyzes
ARF activation (5–7). The large-molecular-weight BIG1, BIG2,
and GBF1, which localize in Golgi regions, are, with different
sensitivities, inhibited by the fungal fatty acid metabolite brefel-
din A (8–10). The ?200-kDa BIG1 and ?190-kDa BIG2 were
first purified together from bovine brain cytosol on the basis of
their BFA-inhibited GEF activities (11) and appeared to exist as
parts of the same macromolecular complex(es) in HepG2 cells
(12). When these cells were growing in medium with serum,
BIG1 was primarily cytosolic and Golgi-associated. After over-
night incubation without serum, a large fraction of endogenous
BIG1 was found in the nuclei localized with nucleoporin p62 at
the nuclear envelope (probably in transit between nucleus and
cytoplasm), as well as in the nuclear matrix and nucleoli (13).
The cAMP-dependent protein kinase A (PKA) is a tetrameric
protein comprising two catalytic (C) and two regulatory (R)
subunits (14). Four isoforms of PKA are designated by their
specific R subunits: RI? and RI? in type I and RII? and RII?
in type II. The holoenzyme is usually tethered by means of
ammalian ADP-ribosylation factors (ARFs) comprise six
20-kDa GTP-binding proteins with essential regulatory
interaction of the R dimer with an A kinase-anchoring protein
(AKAP) that serves as a scaffold for assembly of the kinase,
substrates,andotherenzymes,whichactinconcerttocoordinate
and localize effects of cAMP in cells (15–17). AKAP domain
structures have been reported in diverse proteins, including
BIG1 and BIG2 (18). Consistent with an AKAP function for
BIG1, it was coimmunoprecipitated with RI? and RII? as well
as with C subunits, from HepG2 cytosol (18). PKA activity has
been implicated in several transport pathways (19) and was
reported to play a regulatory role in ARF1 recruitment from
cytosol to intracellular membranes, perhaps by phosphorylating
proteins in the Golgi membrane that serve as binding sites for
ARF1 (20). Activation of PKA also altered the subcellular
localization of other proteins, increasing their presence in either
cytoplasm or nucleus (21, 22).
We report here that, in cells incubated with 8-Br-cAMP, BIG1
was redistributed from membrane to nuclear fractions in a
process that was blocked by PKA inhibitors. BIG1 accumulation
in the nucleus was also specifically blocked by mutation of the
putative PKA phosphorylation site or the nuclear localization
signal (NLS) in the Sec7 domain of BIG1, consistent with its
dependence on both PKA-catalyzed phosphorylation and the
predicted NLS.
Results
Effect 8-Br-cAMP on BIG1 and RI? Subcellular Localization. HepG2
cells were incubated overnight (16 h) without serum before
experiments. To quantify the distribution of BIG1 and RI? in
HepG2 cells, we analyzed cell fractions by Western blotting with
densitometry. We had found that, during incubation of cells with
8-Br-cAMP, amounts of BIG1 decreased in the membrane
fraction to ?40% of the zero time level after 20–30 min, at the
same time increasing to approximately twice the initial level in
nuclei (data not shown). Effects of cAMP at concentrations of
1–1,000 ?M were compared in 20-min incubations. Statistically
significant effects of 1 mM cAMP confirmed those seen in
time-course experiments (data not shown). After exposure to
cAMP,amountsofBIG1inthecytosolwereunchanged,whereas
BIG1 was significantly decreased in membrane fractions and
concomitantly increased in nuclear fractions (Fig. 1A). RI? was
present in all subcellular fractions with no statistically significant
change after cAMP treatment (Fig. 1A). BIG2 was not detected
in the nuclear fraction, and cAMP did not alter its distribution
(Fig. 1A).
On confocal immunofluorescence microscopy, endogenous
BIG1 and RI? were seen throughout the cells with partial
Conflict of interest statement: No conflicts declared.
Abbreviations: AKAP, A kinase-anchoring protein; BIG1, brefeldin A-inhibited guanine
nucleotide-exchange protein 1; siRNA, small interfering RNA; HA, hemagglutinin; NLS,
nuclear localization signal; ARF, ADP-ribosylation factor; PKA, protein kinase A; wt, wild
type.
*To whom correspondence may be addressed at: Building 10, Room 5N-307, National
Institutes of Health, Bethesda, MD 20892-1434. E-mail: citteric@nhlbi.nih.gov or
vaughanm@nih.gov.
www.pnas.org?cgi?doi?10.1073?pnas.0510571103PNAS ?
February 21, 2006 ?
vol. 103 ?
no. 8 ?
2683–2688
CELL BIOLOGY

Page 2

colocalization most evident in nuclei (Fig. 1B). After incubation
of cells with cAMP for 20 min, BIG1 was clearly more concen-
trated in the nuclei, consistent with the increase quantified in
Fig. 1A. RI? distribution seemed somewhat different, although
the two proteins still appeared partially colocalized in nuclei
(Fig. 1B). RI? immunoprecipitated with BIG1 from the nuclear
fraction both before and after cAMP treatment (Fig. 1C). The
amount of anti-BIG1 antibodies that completely precipitated
BIG1 from nuclei of untreated cells failed to immunoprecipitate
100% of the larger amount of BIG1 in cAMP-treated cells and
also failed to immunoprecipitate all of the RI?. Antibodies
against BIG1 immunoprecipitated BIG2 from cytosolic but not
nuclear fractions (Fig. 1C), consistent with failure to detect
BIG2 in nuclei microscopically or on Western blotting.
Effect of PKA Inhibition on 8-Br-cAMP-Induced BIG1 Translocation. To
assess PKA-catalyzed phosphorylation involvement in BIG1
redistribution, effects of PKA inhibitors H-89 and PKI on BIG1
distribution in cells incubated with cAMP were investigated.
BIG1 distribution was not altered by H-89 or PKI alone, but each
inhibitor significantly decreased the effects of cAMP on BIG1
content of membranes and nuclei (Fig. 2). 8-Br-cGMP did not
mimic the effects of cAMP on BIG1 content of membranes and
nuclei, consistent with a role for PKA in the trafficking of this
protein. RI? distribution was not altered by PKA inhibitors or
by 8-Br-cGMP treatment (Fig. 2), which likewise had no effects
on amounts of BIG2 and RI? in cytosol (data not shown).
BIG1 Distribution in Cells Incubated with BIG2 Small Interfering RNA
(siRNA). Because BIG1 and BIG2 had appeared to exist in the
same macromolecular complex(es) in HepG2 cytosol (12), we
investigated the effect on BIG1 distribution of ‘‘knocking down’’
BIG2 with siRNA. Incubation of HepG2 cells for 72 h with BIG2
siRNA markedly decreased BIG2 protein levels (Fig. 3A) and
Fig. 1.
HepG2 cells were incubated (37°C, 20 min) with 1 mM 8-Br-cAMP before
fractionation of homogenates. Samples (5%) of proteins from each fraction
corresponding to ?50 ?g from cytosol, 30 ?g from membranes, and 20 ?g
from nuclei were separated by SDS?PAGE before Western blotting with
antibodies against BIG1, RI?, or BIG2. Data were similar in three experiments.
(B) Cells, incubated as in A without (untreated) or with 8-Br-cAMP were
reacted with antibodies against BIG1 (green) and RI? (red) and inspected by
confocallaser-scanningmicroscopy.(Scalebar:20?m.)(C)Samplesofproteins
from cytosol (Cy, 200 ?g) and nuclei (Nu, 400 ?g) of cells incubated for 20 min
without or with 8-Br-cAMP (Nu*), in 500 ?l of TKMS buffer (13) were incu-
bated with 4 ?g of rabbit IgG (IgG) or anti-BIG1 antibodies overnight at 4°C.
Samples (40 ?l) of supernatant (S) or of immunoprecipitated proteins (P)
eluted from washed beads in 40 ?l of gel-loading buffer and 50 ?g (2.5%) of
total homogenate proteins (Ly) were separated and reacted with antibodies
against RI?, BIG2, or BIG1. Data were similar in three experiments.
Effect of 8-Br-cAMP on intracellular distribution of BIG1 and RI?. (A)
Fig.2.
Cells were incubated (20 min, 37°C) without or with 1 mM 8-Br-cAMP and?or
1 mM 8-Br-cGMP, 100 ?M H-89, or 10 ?M PKI before SDS?PAGE separation of
proteins from membrane (30 ?g) and nuclear (20 ?g) fractions and reaction
withantibodiesagainstBIG1orRI?.Dataaremeans?SDofvaluesfromthree
experiments, blots from one of which are shown.
EffectofPKAinhibitionon8-Br-cAMP-inducedtranslocationofBIG1.
2684 ?
www.pnas.org?cgi?doi?10.1073?pnas.0510571103 Citterio et al.

Page 3

resulted in the virtual disappearance of BIG2 staining in cells
(Fig. 3C). There were, however, no changes in amounts or
distribution of BIG1 and RI? proteins (Fig. 3A). Similarly,
effects of cAMP on BIG1 distribution were not altered in cells
transfected with BIG2 siRNA (Fig. 3B), nor was the appearance
of BIG1 in nuclei of cAMP-treated cells (Fig. 3C).
Effect of Nocodazole on BIG1 Distribution. Transport by means of
microtubules has been implicated in the delivery of proteins
through the cytoplasm to nuclear pores (23). We investigated the
effect of microtubule disruption by nocodazole on the nuclear
accumulation of BIG1. After incubation of cells for 60 min with
nocodazole, BIG1 in cytosol was significantly increased, and that
in the membrane fraction decreased, whereas it had disappeared
totally from nuclei (Fig. 4A). Addition of cAMP during the last
20 min of incubation with nocodazole had no effect on BIG1
distribution, which was thus higher in cytosol than it was in cells
exposed to cAMP in the absence of nocodazole and much lower
(i.e., not detectable) in nuclei (Fig. 4A). Confocal immunoflu-
orescence microscopy revealed BIG1 partially colocalized with
?-tubulin in cytoplasm especially in the perinuclear region (Fig.
4B). Colocalization was less in cells incubated with cAMP, where
BIG1 was accumulated in nuclei with the remaining cytoplasmic
BIG1 still partially associated with ?-tubulin, but in a different
pattern (Fig. 4B). In cells treated with nocodazole for 1 h, the
microtubule network was disrupted, no BIG1 was seen in nuclei,
and effects of cAMP on BIG1 localization were apparently
completely abolished (Fig. 4B).
Nuclear Accumulation of BIG1 by Means of Phosphorylation by PKA.
To assess the role in nuclear accumulation of a potential PKA
phosphorylation site in the BIG1 Sec-7 domain, Ser-883 was
replaced by alanine or aspartate to generate, respectively, hem-
agglutinin (HA)-tagged BIG1(S883A) or BIG1(S883D). Local-
ization of overexpressed HA-tagged wild-type (wt) BIG1 resem-
bled that of endogenous BIG1 (Figs. 1B and 4B), but it was
perhaps more concentrated in the perinuclear region (Fig. 5A).
Punctate collections of HA–BIG1(S883A) were more widely
scattered in the cells than wt, whereas HA–BIG1(S883D), which
Fig. 3.
incubated with BIG2 or nonspecific target (C1) or lamin (C2) siRNA or with
vehicle (M) for 72 h before samples (50 ?g) of total cell proteins were
separated by SDS?PAGE and reacted with antibodies against BIG2, BIG1, or
RI?. Below are means ? SD of BIG2 values quantified by densitometry from
three replicate experiments. (B) HepG2 cells transfected or not with BIG2
siRNA were incubated without or with 1 mM 8-Br-cAMP (20 min, 37°C) before
separation of proteins from cytosol (Cy, 50 ?g), membrane (Me, 30 ?g), and
nuclear (Nu, 20 ?g) fractions and reaction with antibodies against BIG1. (C)
Untreated or Mock (vehicle) or BIG2 siRNA-transfected HepG2 cells were
incubated without or with 8-Br-cAMP (*) as in B before reaction with BIG2 or
BIG1 antibodies for confocal immunofluorescence microscopy. (Scale bar: 8
?m.) Findings were similar in three experiments.
Effect of BIG2 siRNA on BIG1 distribution. (A) HepG2 cells were
Fig. 4.
(37°C,60min)withoutorwith10?g?mlnocodazole(Noc)andwithoutorwith
1 mM 8-Br-cAMP during the last 20 min before separation of proteins from
cytosol(50?g),membrane(30?g),andnuclear(20?g)fractionsbySDS?PAGE
andreactionwithantibodiesagainstBIG1.Dataaremeans?SDofvaluesfrom
three experiments, blots from one of which are above. (B) Cells treated as in
A were reacted with antibodies against ?-tubulin (red) and BIG1 (green) and
inspected by confocal laser-scanning microscopy. (Scale bar: 8 ?m.)
Effect of nocodazole on BIG1 localization. (A) Cells were incubated
Citterio et al.PNAS ?
February 21, 2006 ?
vol. 103 ?
no. 8 ?
2685
CELL BIOLOGY

Page 4

may resemble wt BIG1 phosphorylated by PKA, was concen-
trated in nuclei (Fig. 5A). Treatment of cells with cAMP induced
nuclear accumulation of wt HA–BIG1 but did not alter the
distribution of HA–BIG1(S883A), which remained in the cy-
tosol, or the distribution of HA–BIG1(S883D), which had been
concentrated in nuclei in the absence of cAMP (Fig. 5A). No
HA–BIG1(S883D)wasseeninnucleiofcellsafterincubationfor
1 h with nocodazole (data not shown), consistent with the
requirement for microtubule function in the nuclear accumula-
tion of BIG1. BIG1(NLS), in which the NLS was mutated, was
not seen in nuclei of cells exposed to cAMP, nor was the double
mutant BIG1(S883D?NLS), as expected if the NLS is required
for nuclear localization (Fig. 5A).
Distribution of HA–BIG1 wt and mutants was also assessed by
Western blotting with anti-HA antibody. wt BIG1 and all of the
mutants were present in the cytosol with no apparent change in
cells exposed to cAMP (Fig. 5B). After incubation of cells with
cAMP, the amount of wt HA–BIG1 in nuclei was clearly
increased and that in membranes was perhaps decreased, but
cAMP had no effects on amounts of any of the mutant proteins
recovered in membrane or nuclear fractions (Fig. 5B). Amounts
of BIG1(S883D) in nuclei were similar before and after cAMP
treatment (Fig. 5B).
Western blotting with BIG1 antibodies of proteins in lysates of
HepG2 cells transfected with empty vector or the HA–BIG
constructs revealed the overexpressed proteins at similar levels,
all greater than the amount of endogenous BIG1 in cells
transfectedwithemptyvector(Fig.5C).Amountsofendogenous
BIG1 in cells overexpressing the several constructs were ?30%
of those in the control (empty vector) cells and ?15% of those
of the overexpressed proteins (Fig. 5C).
Discussion
The intracellular trafficking of BIG1 and BIG2 remains incom-
pletely understood, and whether a multiprotein complex is
involved in the intracellular actions of these proteins is not clear,
although the proteins were initially purified together from
bovine brain cytosol in an ?670-kDa complex (11). Subse-
quently, yeast two-hybrid screens and coimmunoprecipitation
revealed the association of BIG1 and BIG2 with other proteins,
such as PKA regulatory and catalytic subunits (18). As reported
here, BIG1 was present in the cytosol, membrane, and nuclear
fractions of HepG2 cells and rapidly accumulated in the nuclei
when cells were incubated with cAMP; i.e., the intracellular
redistribution of BIG1 appeared to reflect changes in cell cAMP
concentration and subsequent PKA activation. In fact, when
PKA activity was inhibited with relatively specific kinase inhib-
itors H-89 and PKI, nuclear accumulation of BIG1 was pre-
vented, consistent with a requirement for PKA-catalyzed phos-
phorylation. BIG2, however, was not detected in nuclei, whether
or not BIG1 was present.
When cellular cAMP levels are low, PKA exists as holoen-
zyme, which is too large to enter the nucleus by diffusion (15).
Upon binding cAMP, the R dimer dissociates from two active C
subunits that then diffuse into the nucleus, where they can
phosphorylate cAMP response element-binding proteins, lead-
ing to cAMP effects on transcription (15). Structures and
functions of the R and C subunits are rather well understood
(14),andanadditionallevelofregulationisestablishedbymeans
of interaction of the R dimer with an A kinase-anchoring protein
(AKAP) that tethers PKA and other proteins with which it
functions as an integrated molecular machine (16). A variety of
such multimolecular complexes comprise different assortments
of protein components assembled on specialized AKAP scaf-
folds at specific subcellular localizations (17).
Most AKAPs had initially been thought to interact exclusively
with RII subunits, but PKA can also be tethered via RI? and
RI? (17). Three short sequences in the N-terminal region of the
BIG2 molecule were identified as AKAP domains that bind
different PKA R subunits (18). In a yeast two-hybrid screen and
then with coimmunoprecipitation of in vitro translated, epitope-
tagged proteins, Li et al. (18) demonstrated the interaction of
RI? with BIG2. They also reported that antibodies against RI?
precipitated BIG1 and BIG2 from HepG2 cytosol; RI? was
precipitated by antibodies against BIG2 or BIG1 (18). Here, we
have shown coimmunoprecipitation of BIG1 and RI? from
Fig. 5.
S883D?NLS double mutant were incubated without or with 8-Br-cAMP for 20 min before processing for immunofluorescence with anti-HA antibodies (green).
(Scale bar: 8 ?m.) (B) Cells were fractionated, and samples of proteins (5% of each fraction) were analyzed by Western blotting with HA antibodies. (C) Proteins
(50?g)fromtotallysatesofHepG2cellstransfectedwithemptyvector(EV)orHABIG1wtandmutantswereanalyzedbyWesternblottingwithBIG1antibodies
to detect both endogenous and overexpressed BIG1.
Overexpression of BIG1 and mutants in HepG2 cells. (A) HepG2 cells, 24 h after transfection with HA-tagged BIG1 wt, S883A, S883D, NLS mutant, or
2686 ?
www.pnas.org?cgi?doi?10.1073?pnas.0510571103Citterio et al.

Page 5

HepG2 cell nuclei. This interaction was not dependent on BIG2,
which was not detected in the nuclei. The BIG1–RI? interaction
is consistent with the reported presence in BIG1 of an AKAP
sequence identical to one in BIG2 that interacted with both RI
and RII subunits in yeast two-hybrid experiments (18).
Proteins larger than 45 kDa require a nuclear localization
sequence (NLS) for entry into the nucleus (15). Nuclear impor-
tation of proteins is a two-step process involving the dimeric
importin-???, in which the ?-subunit directly binds the NLS
motif and serves as an adaptor for importin-?. NLS–importin-?
complexes interact with nuclear pore complexes by means of
importin-? and are translocated into the nucleus in an energy-
dependent process (19, 24). BIG1 was described in HepG2 cells
colocalized, in part, with nucleoporin p62 at the nuclear enve-
lope, perhaps in transit between nuclear and cytoplasmic com-
partments (13). The NLS in a protein destined for nuclear
localization contains a unipartite, or a bipartite, basic amino acid
cluster, such as KKKRK in SV40 large tumor antigen (25) or
RKR-Xn-RKRKR in T cell protein tyrosine phosphatase (26),
which is recognized by an importin-??? heterodimer. Protein
phosphorylation in the vicinity of the NLS is reported to play a
major role in modulating NLS-dependent nuclear import and
can facilitate NLS recognition by the NLS-binding importin-?
subunit (23, 27, 28).
Protein kinases, including PKA, regulate the subcellular lo-
calization of a number of proteins. Phosphorylation of S312 in
the dorsal protein of Drosophila by PKA increased its affinity for
importin-? and was accompanied by enhanced nuclear accumu-
lation (21). The contribution of a negative charge to that effect
was suggested by the observation that replacement of Ser-312
with Glu lowered the affinity slightly less than did the Ala
substitution (21). BIG1 is a protein of 1,849 aa with a predicted
NLS sequence 711KKPKR715, which belongs to the same class
of monopartite NLS modules present in SV40 T antigen. The
BIG1 sequence 880KKIS883 was identified as a potential PKA-
phosphorylation site, C-terminal to the NLS sequence. Mu-
tagenesis of this site in BIG1 showed that the PKA-catalyzed
phosphorylation of Ser-883 was necessary for nuclear accumu-
lation of BIG1 in response to 8-Br-cAMP. The mutant in which
Ala replaced Ser-883, BIG1(S883A), was not present in nuclei
after cAMP stimulation, whereas the S883D mutant, in which
Asp may resemble a phosphorylated Ser, was present in nuclei
whether or not cells were treated with cAMP. The effect of
Ser-883 replacement by Asp is consistent with a mechanistic
importance for the negative charge at that site, as suggested
earlier (21).
Mutation of the NLS in BIG1 resulted in the absence of its
nuclear localization with or without cAMP treatment. The
double-mutant S883D?NLS, not surprisingly, also failed to
accumulate in nuclei after incubation with cAMP. PKA-
catalyzed phosphorylation of Ser-883 was necessary for nuclear
translocation, but not sufficient, because the presence of a
functional NLS was required. The phosphorylated S883 presum-
ably represents a signal in addition to, or recognized in concert
with, the NLS, which is specifically recognized by the nuclear
transport apparatus. We note that BIG2, which has not yet been
found in nuclei, contains a NLS corresponding to that in BIG1.
BIG2 appears to lack a PKA substrate site (10), but perhaps
phosphorylation by another kinase(s) will result in its nuclear
accumulation.
The NLS is, clearly, not a sole determinant of nuclear local-
ization. Phosphorylation sites together with the NLS constitute
regulatory modules for nuclear localization, which have been
called phosphorylation-regulated NLSs (or prNLSs) (28). Al-
though numerous prNLSs have been identified, the mechanism
that underlies prNLS-dependent regulation of nuclear transport
is still unclear. One example of a prNLS is the CcN motif of the
viral T antigen, transport of which is regulated by dual phos-
phorylations catalyzed by protein kinase CK2 and the cyclin-
dependent kinase cdc2. Phosphorylation of the CK2 site acceler-
ated NLS-dependent nuclear import, whereas phosphorylation of
the cdc2 site adjacent to the NLS inhibited transport and markedly
reducedmaximalnuclearaccumulation(28).Ageneralroleforthis
‘‘CcN motif’’ in regulating nuclear protein transport was suggested
by the similar arrangements of CK2 and cdc2 kinase sites with NLS
in several other proteins.
To begin to understand how BIG1 is delivered to a nuclear
pore for translocation to its intranuclear site(s) and function(s)
that remain to be defined, we assessed the involvement of
microtubules, which function widely in both endocytotic and
exocytotic trafficking pathways. Many animal viruses rely on
microtubule-based transport for delivery to the nuclear enve-
lope, where the viral genome is released through the nuclear
pores (29, 30). Salman et al. (23) demonstrated integration of the
microtubuledeliveryandnucleocytoplasmictransportsystemsin
an animal cell model, showing that the same NLS responsible for
translocation through nuclear pores also invoked active, dynein-
mediated transport along microtubules. In intact cells, the
microtubule network radiates from the centrosome near the
nucleus with plus ends pointing toward the cell periphery. In this
orientation, dynein could move NLS-bearing molecules to the
nuclear envelope for nuclear import. The absence of nuclear
accumulation of BIG1 in cells after depolymerization of micro-
tubules by nocodazole is consistent with their involvement in the
delivery of BIG1 for nuclear import. BIG1 was present, along
with BIG2 and exocyst proteins, in microtubules purified from
HepG2 cells by taxol polymerization (31). It seems most prob-
able that the interactions of BIG1 and BIG2 with microtubules
are independent and quite different, both structurally and
functionally.
Materials and Methods
HepG2 human liver carcinoma cells were purchased from Amer-
ican Type Culture Collection; penicillin, streptomycin, 8-bromo-
cAMP, 8-bromo-cGMP, and nocodazole were purchased from
Sigma; H-89 (catalog no. EI-196) and PKI (catalog no. P-203)
were purchased from Biomol; and goat and horse sera were
purchased from Vector Laboratories.
Antibodies. Preparation and purification of antibodies against
BIG1andBIG2werereportedinrefs.12and31.Mouseantibody
against ?-tubulin was purchased from Sigma, chicken polyclonal
anti-RI? antibody was purchased from Biomol, and polyclonal
antibody against HA was purchased from Santa Cruz Biotech-
nology. Horseradish peroxidase-conjugated anti-rabbit IgG and
anti-mouse IgG were purchased from Amersham Pharmacia,
and anti-chicken IgY was purchased from Promega.
Cell Culture. HepG2 cells were grown and incubated, unless
otherwise indicated, on collagen I-coated 10-cm dishes (Becton
Dickinson) at 37°C, 5% CO2?95% air in DMEM (GIBCO) with
10% FBS (GIBCO), penicillin (100 units?ml), and streptomycin
(100?g?ml).Beforeexperiments,cellswereincubatedovernight
(16 h) in the same medium without FBS.
Cell Fractionation, Western Blotting, and Immunoprecipitation. Pro-
teins from HepG2 cell fractions prepared as described in ref. 13
were separated by SDS?PAGE in 10-well 4–12% gels (Invitro-
gen) and transferred to nitrocellulose membranes, which were
divided for reaction with anti-BIG1 (0.5 ?g?ml), anti-RI? (5
?g?ml),oranti-BIG2(1:200).Secondaryantibodies,horseradish
peroxidase-conjugated goat anti-rabbit and goat anti-mouse
IgG, were detected with SuperSignal chemiluminescent sub-
strate (Pierce). Densitometry was performed by using a Chemi-
Imager 5500 (Alpha Innotech, San Leandro, CA). Immunopre-
cipitation is described in the Fig. 1 legend.
Citterio et al. PNAS ?
February 21, 2006 ?
vol. 103 ?
no. 8 ?
2687
CELL BIOLOGY