(Received for publication, February 12, 1997, and in revised form, April 22, 1997)
From the Nuclear Signaling Laboratory, Division for Biochemistry and Molecular Biology, John Curtin School of Medical Research, Canberra, ACT 2601, Australia
The mechanism by which phosphorylation regulates nuclear localization sequence (NLS)-dependent nuclear protein import is largely unclear. Whereas nuclear accumulation of SV40 large tumor antigen (T-ag) fusion proteins is completely dependent on the T-ag NLS (amino acids 126-132), the rate of nuclear import is increased 50-fold by amino acid residues 111-125 and in particular a site for the protein kinase CK2 (CK2) at serine 111/112. Because the first step of nuclear protein import involves the binding of the NLS by an NLS-receptor complex such as the importin 58/97 heterodimer, we established a novel enzyme-linked immunosorbent assay to test whether NLS recognition is influenced by amino acids amino-terminal to the NLS and the CK2 site. We found that recognition of the T-ag NLS by importin 58/97 was enhanced 10-fold in the presence of amino acid residues 111-125 and strongly dependent on importin 97. A T-ag fusion protein in which the spacer between the CK2 site and the NLS was decreased showed 30% reduced binding by importin 58/97. Maximal nuclear accumulation of this protein was reduced by more than 50%, indicating the physiological importance of the correctly positioned CK2 site. Phosphorylation by CK2 increased the T-ag NLS binding affinity for importin 58/97 by a further 40%. We conclude that flanking sequences and in particular phosphorylation at the CK2 site are mechanistically important in NLS recognition and represent the basis of their enhancement of T-ag nuclear import. This study thus represents the first elucidation of the mechanistic basis of the regulation of nuclear protein import through phosphorylation within a phosphorylation-regulated NLS.
Nuclear protein transport is dependent on specific targeting signals called nuclear localization sequences (NLSs),1 defined as the sequences sufficient and necessary for nuclear targeting (1, 2). They are typically short sequences of a single series of basic residues (resembling the NLS of the SV40 large tumor antigen (T-ag)) or of two clusters of basic residues interrupted by a 10-12-amino acid spacer (bipartite NLSs), and they appear to be functional in various protein contexts, suggesting that NLS function is largely independent of secondary/tertiary structure (see Refs. 3-5).
NLS-dependent nuclear protein import can be divided into
two steps, the first of which is energy-independent and involves recognition and targeting of the NLS-bearing protein to the nuclear pore complex (NPC) by a heterodimeric protein complex. The NLS is
specifically recognized by the smaller subunit of the complex or NLS
receptor (6), known variously as importin 58 (7), importin (8),
hSRP1/NP-1 (9), or karyopherin
(10). The larger protein subunit,
importin 97 (11), importin
(12), karyopherin
(13), p97 (14), or
Kap95p (15), binds importin 58 specifically but cannot bind the NLS.
Its role is to target the importin-NLS carrying protein complex to the
NPC through its affinity for NPC components such as nucleoporins (13,
16-18). The second, energy-dependent step of translocation
of the import substrate through the NPC into the nucleus (19, 20)
requires the GTPase Ran/TC4 (21, 22) and the Ran-interacting factor p10/NTF2 (23, 24).
In addition to the NLS, phosphorylation in the vicinity of NLSs has been shown to play a role in regulating nuclear protein import through modulation of NLS function in either a positive or negative fashion (4, 5). The regulatory modules able to confer regulated nuclear protein import on heterologous proteins, called phosphorylation-regulated NLSs (prNLSs) (4, 5), have been identified for a number of nuclear proteins (25-29). Despite clear evidence that prNLSs regulate nuclear protein import (4, 5), the precise mechanism of the action of phosphorylation in terms of regulating NLS function is largely unclear, especially with respect to prNLSs where phosphorylation enhances transport.
Although nuclear import in the case of T-ag is completely dependent on the T-ag NLS (amino acids 126-132) (30), measurements of nuclear import kinetics at the single cell level have shown that the sequences amino-terminal to the NLS (amino acid residues 111-125) enhance the rate of nuclear transport about 50-fold (31, 32). The sequence primarily responsible for this effect has been demonstrated to be the protein kinase CK2 (previously casein kinase II) site at serine 112 (31, 32), but the mechanism of action of phosphorylation at the site is unknown. The cyclin-dependent kinase site at threonine 124 flanking the NLS also regulates T-ag nuclear import through negative modulation of the maximum level of nuclear accumulation (25), and the prNLS or regulatory module for T-ag nuclear import is thus called the "CcN motif," where "C" denotes the CK2 site, "c" denotes the cyclin-dependent kinase site, and "N" denotes the NLS (25, 32).
Because the CK2 phosphorylation event enhancing T-ag transport clearly occurs in the cytoplasm prior to nuclear import (32), we hypothesized that phosphorylation may modulate the affinity of the interaction of the NLS receptor with the T-ag NLS. In the present study we address this possibility directly by using a novel ELISA-based binding assay to quantitate NLS recognition by importin subunits. We find that NLS binding by importin 58 was strongly dependent on importin 97. Significantly, sequences flanking the T-ag NLS enhance recognition by importin 58/97 by about 10-fold, whereas CK2 phosphorylation increases binding by a further 40%. This modulation of NLS recognition through CK2 phosphorylation presumably constitutes the basis of enhancement of the rate of nuclear import by the CK2 site in vivo.
T-ag fusion proteins
CcN--Gal, CN-
-Gal, N-
-Gal, Cc-
-Gal,
cN-
-Gal1, cN-
-Gal2, and sh-cN-
-Gal
containing T-ag amino acids fused amino-terminal to the
Escherichia coli
-galactosidase enzyme sequence (amino
acids 9-1023) have been described previously (25, 31, 32) (for amino
acid sequences see Table I). T-ag fusion protein sh-CcN-
-Gal (in
which T-ag amino acids 116-121 are deleted) was derived through
oligonucleotide site-directed mutagenesis (CLONTECH
Transformer kit) of plasmid pPF10, which encodes fusion protein
sh-cN-
-Gal (31). The peptides pep101Lys and pep101Thr (see Table I)
were provided by Imre Pavo and Gabor Toth (Endocrine Unit, Szeged
University Medical School, Szeged, Hungary), whereas pep11Lys was from
Bachem. T-ag fusion proteins were expressed, purified, and labeled
fluorescently as described previously (31). Mouse importin
58-glutathione S-transferase (GST) and 97-GST expression plasmids were expressed as described (7, 11).
|
Cells of the HTC rat hepatoma tissue culture cell line (a derivative of Morris hepatoma 7288C) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (25, 31). Analysis of nuclear import kinetics at the single cell level using microinjected HTC cells in conjunction with confocal laser scanning microscopy (CLSM) (Bio-Rad MRC-600) was as described previously (33). Microinjected HTC cells were fused with polyethylene glycol about 1 h prior to microinjection to produce polykaryons (31). Quantitation of fluorescence using CLSM (see Refs. 34 and 35 for other applications), image analysis of CLSM files using the NIH Image public domain software, and curve fitting have been described previously in detail (26, 33).
In Vitro PhosphorylationIn vitro
phosphorylation of T-ag fusion proteins by rat liver CK2 (Promega) was
performed with 1 mg/ml fusion protein as described previously (33, 36)
in the absence or the presence of [-32P]ATP (3000 Ci/mmol). The stoichiometry of phosphorylation was quantitated as
described (25).
Subsequent to either spotting T-ag fusion proteins onto nitrocellulose membrane (dot blot) or electrophoresis of T-ag fusion proteins on a 7.5% SDS-polyacrylamide gel and transfer to a nitrocellulose membrane (ligand-Western blot), the membrane was blocked in intracellular buffer (IB; 110 mM KCl, 5 mM NaHC03, 5 mM MgCl2, 1 mM EGTA, 0.1 mM CaCl2, 20 mM Hepes, 1 mM dithiothreitol, 5 µg/ml leupeptin, pH 7.4) containing 5% BSA for 4 h at 22 °C. Hybridization was performed for 16 h at 4 °C in IB/1% BSA supplemented with importin 58-GST, importin 97-GST, or precomplexed importin 58 (with or without GST)/97-GST. Precomplexation of mouse importin subunits (molar ratio of 1:1) was carried out in IB for 15 min at 22 °C. The membrane was extensively washed, and bound mouse importin subunit(s) were detected and visualized using a GST-specific antibody (Pharmacia Biotech Inc.), alkaline phosphatase-conjugated second antibody (Sigma) and nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl-1-phosphate (Promega).
ELISA-based Binding AssayFusion proteins (0.5 µg/well) or peptides (1 µg/well) in 50 mM NaHCO3, pH 9.8, were coated in triplicate wells of polystyrene microtiterplates (Nunc) for 16 h at 4 °C. After blocking in IB/5% BSA for 90 min at 22 °C, appropriate dilutions of precomplexed importin 58/97-GST in IB/1% BSA were then added to the microtiterplates and incubated for 16 h at 4 °C. After washing the wells with IB/1% BSA, bound mouse importin 58/97-GST complex was detected as described for the ligand-Western blot assay and visualized with p-nitrophenyl phosphate (Sigma). The change of absorbance at 405 nm was followed with time using a plate reader (Molecular Devices), with values corrected by subtracting both the absorbance at 0 min, and the absorbance in wells incubated without importin 58/97-GST complex.
To correct for differences in coating, T-ag fusion proteins were
subjected to a parallel -galactosidase ELISA assay using
-galactosidase-specific antibodies. Measurements for importin binding were ultimately corrected for any differences in coating efficiencies quantified in the
-galactosidase-based ELISA assay to
enable a true estimate of bound importin.
Initial experiments to investigate binding of the
importin 58/97 subunits to the T-ag NLS (Table I for
T-ag sequences of the fusion proteins) using the dot blot technique
indicated poor binding of importin 58 to the T-ag NLS in the absence of
importin 97, which alone did not bind the T-ag NLS at all (Fig.
1A). This was consistent with the results of
others that the importin 58/97 complex constitutes the high affinity
NLS receptor (37, 38). Interestingly, importin 58/97 only showed strong
binding to the T-ag NLS in the presence, compared with in the absence,
of T-ag amino acids 111-125 flanking the NLS (amino acids 126-132)
(Fig. 1A). To evaluate this further, the ligand-Western blot
technique was used to assess binding of the importin 58/97 complex to
various T-ag fusion proteins differing in the T-ag sequences
amino-terminal to the NLS. As in the dot blot assay, the T-ag fusion
proteins Cc--Gal and N-
-Gal bound importin 58/97 poorly compared
with CcN-
-Gal (Fig. 1B). The sh-cN-
-Gal protein bound
importin 58/97 to intermediate levels, implying that the reduced
binding relative to CcN-
-Gal may have been due to the absence of the
CK2 site. The T-ag fusion protein derivative sh-CcN-
-Gal, containing
a deletion of amino acids 116-121 ("short" version of
CcN-
-Gal), thereby reducing the distance between the CK2 site and
the NLS from 10 to 4 residues (Table I), did not bind importin 58/97 complex as well as CcN-
-Gal (see also below), implying that
positioning of the CK2 site relative to the NLS was crucial to
efficient binding by the importin 58/97 complex.
The functional significance of the altered position of the CK2 site in
sh-CcN--Gal was assessed in nuclear import studies in microinjected
cells of the HTC rat hepatoma line using CLSM. The sh-CcN-
-Gal
fusion protein exhibited maximal nuclear accumulation less than half
that of CcN-
-Gal (Fig. 2) but higher than that of a
protein lacking a functional CK2 site (31, 32), indicating the
functional importance of the correct position of the CK2 site relative
to the NLS for nuclear import. It also exhibited a reduced initial rate
of nuclear import relative to the wild type fusion protein, the reduced
nuclear accumulation of sh-CcN-
-Gal relative to the wild type fusion
protein thus correlating with the decreased binding by importin 58/97
(Fig. 1B and see below). That reduced transport was not
attributable to reduced efficiency of phosphorylation at the CK2 site
was demonstrated by phosphorylation experiments with purified CK2,
where sh-CcN-
-Gal was phosphorylated to an identical extent to
CcN-
-Gal (stoichiometry of phosphorylation of 2.04 and 2.02 mol
Pi/mol tetramer for sh-CcN-
-Gal and CcN-
-Gal, respectively).
Quantitation of T-ag Fusion Protein Binding to the Importin 58/97 Complex
To define the role of the T-ag amino-terminal sequences
in NLS binding by importin 58/97 in quantitative terms, a novel
ELISA-based binding assay was established (see "Experimental
Procedures" for details) whereby T-ag fusion proteins coated onto
microtiterplates were incubated with increasing amounts of importin 58 and 97 or importin 58/97 complex. Negligible association of importin 97 with CcN--Gal was observed, while only low efficiency binding was
evident when importin 58 alone was used compared with that when the
importin 58/97 complex was used (not shown), consistent with dot blot
analysis (Fig. 1A). The apparent dissociation constant (KD) of CcN-
-Gal for importin 58 alone was 10 times that for the importin 58/97 complex.
The role of the different regions of the CcN motif in recognition of
the T-ag NLS by importin 58/97 was assessed using the CN--Gal,
sh-CcN-
-Gal, cN-
-Gal1, cN-
-Gal2,
sh-cN-
-Gal, and N-
-Gal T-ag fusion protein derivatives (Fig.
3A and Table I). The amino-terminal sequences
flanking the NLS enhanced binding by importin 58/97 whereby binding to
wild type protein (CcN-
-Gal) was about 10-fold higher than that to
the fusion protein containing the T-ag NLS alone (N-
-Gal) (Table I).
Cc-
-Gal bound negligible amounts of importin 58/97, indicating the
specificity of the assay (Table I). The sh-cN-Gal protein, which lacks
the CK2 site, bound importin 58/97 to levels about 50% that of
CcN-
-Gal, implying that the CK2 site was necessary for efficient
importin binding. The cN-
-Gal1 protein, containing
nonphosphorylatable residues in place of the CK2 site serines 111 and
112 (32), and cN-
-Gal2, carrying a mutated CK2
recognition site (substitution of Asp-Asp-Glu115 by
Asn-Asn-Gln) but retaining serines 111 and 112, also bound importin to
a reduced level, supporting the idea of the CK2 site being directly
involved in importin recognition. This was further supported by the
fact that the sh-CcN-
-Gal protein, the "short" version of the
CcN motif containing a functional CK2 site, bound higher levels of
importin 58/97 than sh-cN-Gal (deleted for amino acids 111-119,
including the CK2 site) (Fig. 3A); binding, however, was
reduced relative to wild type, implying that although the CK2 site is
essential for high affinity binding by importin, its correct
positioning relative to the NLS is a critical factor. The lack of a
correctly positioned CK2 site in sh-CcN-
-Gal increased the
KD by almost 2-fold relative to CcN-
-Gal (Fig.
3A and Table I), while the complete lack of a functional CK2
site (sh-cN-Gal, cN-
-Gal1, and cN-
-Gal2)
increased the KD even further. Substitution of
serines 120 and 123 together with threonine 124 (the site of
phosphorylation by the cyclin-dependent kinase
cdc2, which reduces the maximal level of T-ag fusion protein nuclear accumulation) (25) by alanine residues (the CN-
-Gal protein)
had no significant effect on importin recognition relative to
CcN-
-Gal (Table I), further underlining the central importance of
the CK2 phosphorylation site in importin binding. Taking the results
from Table I (the rightmost data columns), the KD relative to wild type for the respective fusion proteins correlated very well (regression coefficient of 0.99) with their initial rates of
nuclear import in vivo determined either previously (31, 32)
or in this study, implying that the apparent dissociation constant is
critical in determining the initial rate of nuclear import.
Peptides comprising T-ag amino acids 112-132 (pep101Lys), T-ag amino acids 112-132 with a nonfunctional NLS (pep101Thr), and T-ag amino acids 126-132 (pep11Lys) (Table I) were also tested. Consistent with the studies using the T-ag fusion proteins, binding of the importin 58/97 complex to pep101Lys (KD of 17 nM) was significantly higher than that to pep11Lys, underlining the importance of the flanking regions to T-ag NLS recognition by importin 58/97. Peptide pep101Thr showed negligible binding (Fig. 3B and Table I), again confirming the specificity of the ELISA-based binding assay.
CK2 Site Phosphorylation Increases the Affinity of Binding of Importin 58/97 to the T-ag NLSOur previous work had demonstrated
that nuclear accumulation of CcN--Gal is much more rapid than that
of cN-
-Gal1, cN-
-Gal2, and sh-cN-
-Gal,
all of which lack a functional CK2 site (Refs. 31 and 32 and Table I).
To test whether phosphorylation at the CK2 site influences binding of
importin 58/97 to the T-ag NLS, CcN-
-Gal was phosphorylated with CK2
to a stoichiometry of 1.0-1.5 mol Pi/mol tetrameric T-ag
fusion protein, and importin 58/97 binding was quantified using the
ELISA-based binding assay. The clear result was that phosphorylation at
the CK2 site increased binding of importin 58/97, with about a 40%
decrease in the value of the KD compared with
unphosphorylated CcN-
-Gal (Table II). Differences
between non- and prephosphorylated T-ag proteins were much more
marked (up to 2.5-fold differences in the amount of importin 58/97
complex bound) at lower importin 58/97 concentrations (data not shown);
it should also be stressed that the quantitative differences here
almost certainly underestimate the effect of CK2 phosphorylation in
terms of increasing the affinity for importin 58/97 because CK2
prephosphorylated CcN-
-Gal tetramers represent a mixed population of
unphosphorylated and phosphorylated CK2 sites as a result of the
stoichiometry of phosphorylation (i.e. 2.5 mol of CK2 sites
per tetramer are not phosphorylated). This almost certainly results in
underestimation of the KD for importin association
of CK2-phosphorylated protein.
|
To confirm that the effects on importin binding did not result from
trace amounts of CK2 in the protein samples in the case of
prephosphorylated protein, control experiments were carried out using
sh-cN--Gal, N-
-Gal, and Cc-
-Gal preincubated with CK2.
Phosphorylation of sh-cN-
-Gal and N-
-Gal was undetectable, as
expected (Table II). There was no increase in the maximal binding of
the CK2-preincubated proteins sh-cN-
-Gal and N-
-Gal compared with
the nonpreincubated T-ag fusion proteins (Table II). We also found no
increased binding of importin 58/97 to CK2 preincubated
-Gal (not
shown). Significantly, a 3-fold increase in binding of importin 58/97
to CK2 prephosphorylated Cc-
-Gal was observed relative to
nonphosphorylated Cc-
-Gal (Table II). This enhancing effect of CK2
site phosphorylation on binding by importin 58/97 even in the absence
of a functional NLS is a further indication that the CK2 site probably
participates directly in binding to importin 58.
In conclusion, this study establishes that the mechanistic basis of CK2 site-mediated enhancement of T-ag nuclear import is through the CK2 site and CK2 site phosphorylation increasing the affinity of interaction of the NLS receptor importin 58/97 with the T-ag NLS. As supported by the correlation between the KD and initial nuclear import rate in vivo (see above), the higher affinity of binding presumably results in more rapid kinetics of association of the NLS receptor with the T-ag transport substrate and faster docking at the NPC ultimately leading to an accelerated rate of transport, which has been observed both in vivo and in vitro. (25, 31-33). This study thus represents the first elucidation of the mechanistic basis of the regulation of nuclear protein import through phosphorylation within a prNLS.
We thank Prof. Yoshihiro Yoneda (Osaka University Medical School, Osaka) for providing the mouse importin 58 and 97 expression plasmids, Drs. Imre Pavo and Gabor Toth (Endocrine Unit, University of Szeged) for peptides pep101Lys and pep101Thr, the Clive and Vera Ramaciotti Foundation for its support of the work, and Patricia Jans and Lyndall J. Briggs for skilled technical assistance.