(Received for publication, October 11, 1996)
From the Departamento de Bioquímica y Biología Molecular, Facultad de Biología, Universidad de Santiago, Santiago de Compostela, Spain
Prothymosin (ProT
) is an acidic protein
involved in cell proliferation. Its phosphorylation status is
correlated with proliferative activity. Here we report the isolation
and characterization of a ProT
-phosphorylating kinase (ProT
K)
from mouse splenocytes that seems to be responsible for the in
vivo phosphorylation of ProT
and that differs from other
protein kinases reported to date. This enzyme, mainly located in the
cytosol, has an molecular mass of 180 kDa and appears to be made up of
two proteins of 64 and 60 kDa. Its activity was markedly enhanced by
mitogenic activation of cells. The ProT
residues phosphorylated by
the enzyme in vitro are a Thr at position 7 and another Thr
at positions 12 or 13, both located within casein kinase 2 (CK-2)
consensus motifs; these are the same residues as are phosphorylated
in vivo. The new enzyme shows a number of clear structural
and catalytic differences from CK-2. It phosphorylates histones H2B and
H3, although with weaker activity than ProT
. An enzyme with the same
characteristics was also found in other murine tissues and cell
lines.
Prothymosin (ProT
)1 is a
polypeptide comprising 109-111 amino acid residues, according to
protein (1), cDNA (2, 3), and genome (4) analysis. Its sequence is
highly conserved (5) and includes an extensive central acidic region
(residues 41-85) comprised of Glu and Asp residues. Under
physiological conditions, ProT
behaves as a monomeric protein with a
random coil conformation (6, 7). It is widely distributed in mammalian
tissues and particularly abundant in lymphocytes (8-10).
The changes over time in the levels of ProT (11) and ProT
mRNA (12-14) observed in dividing cells suggest that its function is related to cell proliferation. This is in agreement with the efficiency of antisense oligonucleotides complementary to ProT
mRNA for paralyzing cell growth (15). Furthermore, the proposed implication of the c-Myc protein in the transcriptional regulation of
the ProT
gene (16-18), although controversial (19), would support
this hypothesis.
Although the precise function of ProT remains unclear (for review,
see Ref. 20), there is strong evidence to suggest a nuclear role: its C
terminus region bears a karyophilic signal (21) and ProT
migrates to
the nucleus in proliferating cells (22, 23). Recent work in our
laboratory has shown that ProT
binds histones and cooperates in
nucleosome assembly in vitro (24), suggesting that the
putative nuclear function may be related to chromatin remodeling; this
possibility is consistent with the structural similarities between
ProT
and various nuclear proteins known to be involved in chromatin
activity (25, 26).
Phosphorylation of ProT has recently been reported by us (27) and
corroborated by other authors (28). This post-translational modification constitutes an additional clue to its biological role. The
available data indicate that the extent of phosphorylation of ProT
is positively correlated with cell proliferation activity (27, 28). The
phosphorylation sites are included in the first 14 amino acids of the
sequence: however, there is controversy as to whether the residues
which undergo phosphorylation are Thr residues (our findings; Ref. 27),
or Ser residues (Sburlati et al. (28)).
Regardless of whether the residues phosphorylated are Thr or Ser, the
N-terminal sequence of ProT
(AcSer-Asp-Ala-Ala-Val-Asp-Thr-Ser-Ser-Glu-Ile-Thr-Thr-Lys) suggests
that the in vivo phosphorylation sites are casein kinase 2 (CK-2) consensus sites (29). This is in accordance with our previous
results which show that CK-2 is able to phosphorylate the 14-residue
N-terminal fragment of ProT
in vitro (30). However, phosphoamino acid analysis showed that in vitro
phosphorylation with CK-2 leads to phosphorylation of both Thr and Ser
residues, in approximately equal proportion.
In the work reported here, we aimed to characterize the enzyme
responsible for the phosphorylation of ProT in proliferating cells.
At the same time, we set out to resolve the controversy as to which
amino acid residues undergo phosphorylation. Fractionation of cell
extracts by affinity chromatography on a ProT
-Sepharose column and
then by ion-exchange HPLC yielded a protein kinase, different from
CK-2, which phosphorylated the N-terminal 14-mer of ProT
. Analysis
of the sites phosphorylated both in vivo and by the purified
enzyme in vitro confirmed that the residues phosphorylated in the N-terminal peptide of ProT
are Thr residues, and indicate that the purified enzyme is probably responsible for the
phosphorylation of ProT
observed in vivo.
Materials
The triethylammonium salts of adenosine
5-[
-32P]triphosphate ([
-32P]ATP,
3000 Ci/mmol) and guanosine 5
-[
-32P]triphosphate
([
-32P]GTP, 5000 Ci/mmol) were purchased from Amersham
International. [32P]Orthophosphate (1 Ci/mmol) was from
DuPont NEN. TPCK-treated trypsin, alkaline phosphatase from bovine
intestinal mucosa, poly-L-lysine hydrobromide,
dephosphorylated
-casein from bovine milk, protamine (grade IV) from
salmon, and concanavalin A were from Sigma. Thin-layer cellulose plates
(0.1 mm) were from Merck. Interleukin-2, DNase I, RNase I, histones,
the artificial casein kinase II substrate RRRDDDSDDD and protease V8
(sequencing grade) from Staphylococus aureus were obtained
from Boehringer Mannheim. Antibodies to human casein kinase II were
obtained from Upstate Biotechnology Inc. and casein kinase II from
Promega. ProT
was purified from calf thymocytes as described (10).
All other reagents and materials were of analytical grade.
Methods
Cell Culture and Subcellular FractionationSplenocytes,
thymocytes, and hepatocytes were obtained from 45-day-old female BALB/c
mice. Cells were isolated by pressing a suspension of the freshly
excised and minced tissue in RPMI 1640 medium through a 65-µm mesh
stainless steel screen. In all cases the resulting suspension was
treated with erythrocyte lysis buffer (0.75% NH4Cl, 0.21%
Tris-HCl, pH 7.2) and washed twice in RPMI 1640 before use. Splenocytes
and thymocytes (5 × 106 cells/ml) were grown for
various times in RPMI 1640 containing 10% fetal calf serum, 100 units
ml1 of penicillin, and 100 µg ml
1 of
streptomycin, in the presence or absence of concanavalin A (2.5 µg
ml
1) and interleukin-2 (10 units ml
1), in a
humidified 5% CO2 atmosphere at 37 °C. Cell viability was determined by trypan blue exclusion.
HeLa and NC37 cells were grown in minimum Eagle's medium or RPMI 1640, respectively, at 37 °C in a humidified CO2 atmosphere. In both cases the medium contained 10% fetal calf serum, 100 units ml1 of penicillin, and 100 µg ml
1 of
streptomycin.
Subcellular fractionation of the different cell types was as follows.
Cells were collected from culture, washed twice, suspended (about
5 × 107 cells/ml) in ice-cold lysis buffer (50 mM Tris-HCl, pH 7.4, containing 150 mM NaCl,
1.5 mM MgCl2, 0.5% Nonidet P-40, 0.5 mM dithiothreitol, 5 mM EGTA, 1 mM
EDTA, 50 mM sodium fluoride, 1 mM
phenylmethylsulfonyl fluoride, 10 µg/ml pepstatin, and 10 µg/ml
leupeptin), incubated on ice for 10 min, homogenized (10 strokes) in a
Potter Teflon glass blender, then centrifuged (2000 × g for 10 min) to yield pellet I (nuclei). The supernatant
was ultracentrifuged (110,000 × g for 1 h at
4 °C) to yield pellet II (microsomes) and a supernatant which was
dialyzed against buffer A (50 mM Tris-HCl, pH 7.5, containing 5% (v/v) glycerol, 0.5 mM phenylmethylsulfonyl
fluoride, and 0.5 mM dithiothreitol) to yield the cytosol
fraction. Pellet II was suspended in buffer A containing 0.5 M NaCl and 0.5% sodium deoxycholate, incubated at 0 °C
for 10 min, and centrifuged at 20,000 × g for 15 min:
the supernatant was then dialyzed against buffer A to yield the
microsomal fraction. Nuclei were suspended (about 2 × 108 nuclei/ml) in buffer B (10 mM Tris-HCl, pH
7.4, containing 0.1 mM MgCl2, 1 mM
dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride)
supplemented with 80 µg/ml DNase I and 10 µg/ml RNase I, incubated
at 0 °C for 1 h, and then centrifuged at 20,000 × g for 15 min at 4 °C. The supernatant was dialyzed
against buffer A to yield the nucleoplasm fraction. The pellet was
suspended in buffer B containing 0.5 M NaCl, incubated for
5 min at 0 °C, and centrifuged at 20,000 × g for 15 min; the supernatant was then dialyzed against buffer A to yield the
nuclear envelope fraction. All fractions were either used immediately
or stored at 70 °C until use.
The various subcellular fractions were chromatographed on
ProT-Sepharose columns as described previously (24). Briefly, the
fraction under study, in buffer A, was first passed through a bovine
serum albumin-Sepharose column (1 mg of bovine serum albumin/ml of
Sepharose) at a ratio of about 10 mg of protein per ml of bovine serum
albumin-Sepharose matrix. The flow-through fractions were then slowly
loaded onto a 1-ml column of ProT
-Sepharose (0.8 mg of ProT
per
ml of Sepharose). The column was then sequentially eluted with 0.3 M NaCl and with 1 M NaCl in buffer A. The
flow-through fraction and the two eluate fractions were dialyzed
against buffer A and concentrated to final volumes of 0.5 ml. ProT
kinase activities of the different concentrates were then assayed as
described below. All operations were carried out at 4 °C.
To further purify components in the affinity chromatography fractions
with ProT kinase activity, these fractions were loaded onto a
Spherogel-TSK DEAE-SPW column (7.5 × 75 mm) previously equilibrated with buffer A, in a Beckman Gold HPLC system. The column
was developed for 10 min with buffer A and then eluted with 0-0.8
M NaCl (linear gradient) in the same buffer. Flow rate was
0.5 ml/min and fractions were collected every minute. ProT
kinase
activity was assayed in aliquots of each fraction as described below.
Fractions with the highest ProT
kinase activity were collected and
concentrated at final volumes of 0.5 ml in a Centricon 10 and stored at
4 °C or
70 °C. Protein concentrations were determined by the
Bradford assay (31).
32P-Labeled ProT was
obtained from proliferating cells following modification of the
procedure of Haritos et al. (32), as we have described
previously (27). Briefly, splenic lymphocytes activated with
concanavalin A and interleukin-2 were incubated with
[32P]orthophosphate (50 µCi/ml), then frozen in liquid
nitrogen, powdered, and suspended in boiling 0.15 M NaCl.
After cooling, the suspension was centrifuged and the supernatant
brought to pH 2.5, clarified by centrifugation, made up to 50 mM Tris-HCl (pH 8) and 70 mM NaCl, and applied
to a DEAE-cellulose column, which was eluted with 3 volumes of loading
buffer (50 mM Tris-HCl, pH 8) containing 0.2 M
NaCl, followed by 6 volumes containing 0.5 M NaCl. The
components eluted with the latter buffer were purified by reverse-phase
HPLC. Phosphorylated ProT
, which co-purified with calf ProT
, was
characterized by SDS-PAGE and isoelectric focusing, and on the basis of
immunoreactivity with an antibody raised against the first 28 amino
acids of ProT
(27). This procedure has previously been shown by us
(10) and by others (5, 9) to yield 43-130 µg of ProT
/g of cell,
depending on tissue. In the present study, the yields of radioactive
ProT
after HPLC purification were about 4 × 105
cpm/108 mitogen-activated splenic lymphocytes.
Peptide
mapping on the basis of tryptic digestion of 32P-labeled
phosphorylated ProT, and phosphoamino acid analysis of the ProT
fragments were as described (27). Briefly, 32P-ProT
,
either purified from metabolically labeled cells or from kinase-assay
reaction mixtures, was digested with TPCK-trypsin for 10 h at
37 °C. Tryptic peptides were then separated by reverse-phase HPLC.
Aliquots of the radioactive peak coeluting with the 14-residue N-terminal fragment of calf ProT
were then subjected to phosphoamino acid analysis, by hydrolysis with 6 N HCl, mixing with
nonradioactive phosphoamino acid standards, thin-layer electrophoresis,
and autoradiography. Alternatively, aliquots of the radioactive peak
(of the HPLC-separated tryptic digests, whether from metabolically
labeled cells or from kinase-assay reaction mixtures) were further
digested with endopeptidase V8 (Boehringer Manheim) (protease:peptide
ratio 1:10) in 25 mM ammonium carbonate (pH 7.8). The
resulting peptides were separated by reverse-phase HPLC in an
Ultrasphere-ODS column (5 µm, 4.6 × 250 mm), and phosphoamino
acid analysis was performed as above. The amino acid compositions of
the peptides derived from V8 digestion of the N-terminal fragment were
determined after acid hydrolysis in a Biochrom 20 analyzer
(Pharmacia).
Assays of the ProT
phosphorylating activity of cell extracts or rat liver CK-2 were
performed in 25-µl reaction mixtures containing 50 mM
Tris-HCl (pH 7.4), 150 mM KCl, 26 mM
MgCl2, 1.6 mM EGTA, 1 mM EDTA, 3.3 mM dithiothreitol, 80 ng/ml protamine, 83 mM
-glycerol phosphate, and 100 µM
[32P]ATP. Dephosphorylated calf thymus ProT
(5 µg),
or that indicated under "Results," was used as substrate. After 30 min at 37 °C, the reaction was stopped with 5 µl of 100 mM ATP. Activity was quantified (a) by
separating the components of the reaction mixture by SDS-PAGE (33),
with autoradiography to estimate 32P-ProT
concentration
and kinase activity subsequently expressed in arbitrary units, or
(b) by separating the components by reverse-phase HPLC (as
described above), then quantifying the radioactivity co-eluting with
ProT
, with activity expressed as amount of 32P
incorporated into the substrate. Assays of immunoreactivity with
antibodies raised against the first 28 amino acids of ProT
(see
above) were also used to characterize 32P-ProT
in the
various reaction mixtures.
Immunodepletion assays were carried out in 50-µl aliquots of reaction
mixture buffer, without ATP, containing 30 µl of the purified kinase
and different concentrations of antibody to the subunit of CK-2.
The mixture was incubated overnight at 4 °C with gentle shaking.
Forty µl of Sepharose-bound protein A in reaction mixture buffer was
then added, and the mixture gently shaken for 2 h at 4 °C.
Immunocomplexes were collected by centrifugation and kinase activity
was assayed in the supernatant after adding 100 µM
[32P]ATP.
Phosphorylation of ProT is particularly marked in
mitogen-activated splenic lymphocytes, as compared with non-activated
lymphocytes and other proliferating cell types (27, 28). We thus
searched for the ProT
kinase (ProT
K) in extracts of these cells.
The first step was to isolate cellular components which show affinity for ProT
by using a ProT
-Sepharose column that we have previously shown to be effective for this purpose (24). ProT
phosphorylation activity in the various affinity chromatography fractions was assayed
under conditions equivalent to those used for assaying CK-2 activity,
since the in vivo phosphorylation sites of ProT
are
similar to those of this enzyme (27, 28). Subcellular fractions
(cytosol, nucleoplasm, microsome, and nuclear envelope) from
mitogen-activated splenic lymphocytes were chromatographed on
ProT
-Sepharose columns and ProT
phosphorylation activity was
assayed in the different eluates. As shown in Fig. 1,
components with high affinity for ProT
-Sepharose (lanes
4) in the cytosol (Fig. 1A), nuclear envelope (Fig.
1B), and nucleoplasm (Fig. 1C) fractions all
showed ProT
phosphorylation activity, as judged by the radioactivity
migrating with ProT
in the SDS-PAGE analysis, whereas no such
activity was observed when ProT
was omitted from the reaction
mixtures (Fig. 1, lanes 3). Components obtained from the
microsome fraction showed no detectable ProT
phosphorylation activity (data not shown). Affinity chromatography appears to be
efficient for isolating the ProT
phosphorylating activity, since no
phosphorylated product migrating with ProT
was detected in the
flow-through fractions (Fig. 1, lane 1), and since the phosphorylated product was absent (Fig. 1, B and C,
lanes 2) or present in only low concentration (Fig. 1A, lane
2) in the moderate-affinity fractions. ProT
phosphorylation
activity was imperceptible in the various crude subcellular fractions
(not shown).
Similar results were obtained when an HPLC-based assay (see
"Methods"; results not shown) was used for determination of ProT phosphorylation activity in the different fractions. That the radioactive peaks co-eluting with calf ProT
in HPLC were indeed radiolabeled ProT
was confirmed by immunoprecipitation with the anti-28-mer antibody (see "Methods"; results not shown).
Densitometric scanning of the gels shown in Fig. 1 indicates that
ProT phosphorylation activity is highest in the high-affinity fraction (HAF, lane 4) of the cytosol extract. Activities in
the HAFs of the nuclear envelope and nucleoplasm extracts were about 25 and 5.9%, respectively, of that in the HAF of the cytosol extract. Activity in the moderate-affinity fraction of the cytosol extract was
about 4.8% of that in the HAF of the cytosol extract. Total protein
concentrations were 15, 9, and 4 µg/108 cells in the HAFs
of cytosol, nucleoplasm, and nuclear envelope extracts,
respectively.
ProT phosphorylation activity in the HAF of the cytosol extract from
non-activated splenic lymphocytes was about 14% of that in the HAF of
the cytosol extract from mitogen-activated splenic lymphocytes (Fig.
2). As is also shown in Fig. 2, activity in the HAF of
the nuclear envelope extract from non-activated cells was about 35% of
that in this fraction from activated cells, whereas activity in the HAF
of the nucleoplasm extract from non-activated cells was scarcely
detectable.
Taken together, these results indicate that the ProTK activity in
splenic lymphocytes is influenced by proliferation status and is
largely contained in the cytosol. Some activity is contained in nuclear
envelopes, while activity in the nucleoplasm is negligible. To
characterize this activity, we further purified components of the
cytosol, nuclear envelope, and nucleoplasm extracts of activated
splenic lymphocytes with affinity for ProT
-Sepharose by ion-exchange
HPLC. ProT
phosphorylation activity in the HAFs of the cytosol and
nuclear envelope fractions showed different elution patterns, but was
in both cases concentrated in single clearly defined peaks (Fig.
3A). Kinase activity eluting at the same
position as that in the HAFs from cytosol was scarcely detectable either in the eluate fractions obtained by ion-exchange HPLC of the
HAFs of the nucleoplasm fraction or in eluate fractions obtained by
ion-exchange HPLC of moderate-affinity fractions (see Fig. 1A,
lane 2) from cytosol (result not shown). The specific activity of
the HPLC-purified ProT
K from cytosol extract was about 300-fold higher than in the extract prior to HPLC. No such assessment of the
effectiveness of the second purification step is possible for the
nuclear envelope extract, since protein content in the peak-activity
HPLC fraction was below detection limits. Gel filtration (Fig.
3B) indicated that the HPLC-purified kinase from the cytosol extract has a molecular mass of about 180 kDa, while the molecular mass
of the kinase from the nuclear envelope extract was about 130 kDa.
Identification of Sites of Phosphorylation
To investigate
whether the purified kinases are responsible for the phosphorylation of
ProT observed in vivo, we performed experiments aimed at
more detailed characterization. In various animal cells ProT
has
been shown to be phosphorylated in its 14-residue N-terminal region.
However, controversy remains as to whether the residues phosphorylated
are Thr, as we have found (27), or Ser, as proposed by Sburlati
et al. (28). We thus set out to identify sites
phosphorylated in vivo and sites phosphorylated in
vitro by the kinases purified from cytosol and nuclear envelopes. To this end, 32P-ProT
purified by reverse-phase HPLC
(see "Methods") from metabolically labeled mouse splenocytes, or
from kinase activity assay reaction mixtures, was trypsin-digested, and
the resulting peptides were separated by reverse-phase HPLC. The
elution pattern of the peptides derived from the tryptic digestion of
32P-ProT
phosphorylated in vitro by the
cytosol-fraction ProT
K is shown in Fig.
4A. The observed pattern is identical to
those obtained with ProT
phosphorylated in vitro by the
nuclear envelope fraction ProT
K or with ProT
phosphorylated
in vivo (results not shown). As can be seen from Fig.
4A, all the radioactivity co-eluted with the 14-residue
N-terminal fragment of ProT
. Phosphoamino acid analysis of the
radioactive tryptic peptides derived from ProT
phosphorylated
in vivo or in vitro by the purified enzymes are
shown in Fig. 4B. This result indicates that only Thr
residues were phosphorylated by the cytosolic enzyme and in the ProT
phosphorylated in vivo, whereas both Ser and Thr residues,
in similar proportion, were phosphorylated by the kinase from nuclear
envelopes. This result confirms our previous findings as regards the
phosphorylation of ProT
in vivo (27), and at the same
time indicates that the cytosolic ProT
K, which phosphorylates only
Thr residues, is probably that responsible for the phosphorylation
observed in vivo. To investigate this hypothesis, we next
performed a structural study of the radioactive tryptic fragments of
ProT
phosphorylated in vivo and in vitro which
co-purified with the N-terminal fragment.
The sequence of the first 14 amino acid residues of ProT (see Fig.
4) forms part of the most conserved region of the protein and is
identical in all mammals studied to date (5). Protease V8 (in ammonium
carbonate buffer, pH 7.8) specifically cleaves this fragment between
residues Glu-10 and Ile-11, giving rise to a decapeptide (AcSDAAVDTSSE)
and a tetrapeptide (ITTK). Since only one of these peptides contains
Ser residues, V8 digestion is a useful tool for identification of the
sites phosphorylated by the purified enzymes. V8 treatment of the
N-terminal peptide from ProT
phosphorylated by the cytosolic kinase
(Fig. 4C), the nuclear envelope kinase (Fig. 4D),
or from ProT
phosphorylated in vivo (Fig. 4E), in all
cases efficiently cleaved the N-terminal fragment to yield a
decapeptide and a tetrapeptide, the sequences of which were confirmed
by amino acid analysis (results not shown). The coincidence of the
radioactive peaks in Fig. 4, C-E, with those
derived from V8 cleavage of the N-terminal fragment of ProT
(in the
same figures) provides further confirmation that both the in
vitro (Fig. 4, C and D) and in
vivo (Fig. 4E) phosphorylation sites are contained
within the N-terminal fragment. Moreover, these elution patterns again
indicate that the cytosolic enzyme phosphorylates the same sites as
in vivo. Specifically, the Thr at position 7 in the fragment
AcSDAAVDTSSE and, to judge from the amount of radioactivity, the Thr in
the fragment ITTK at positions 12 or 13 (or both partially) are
phosphorylated both by the cytosolic enzyme (Fig. 4C) and
in vivo (Fig. 4E). By contrast, the nuclear envelope kinase (Fig. 4D) phosphorylates Thr residues at
positions 12 or 13 (or both) in the tetrapeptide, as well as the Thr at position 7, and either one or two Ser residues in the decapeptide; the
latter conclusion is inferred from the fact that the results of
phosphoamino acid analysis of the decapeptide phosphorylated by the
nuclear envelope kinase (not shown) are similar to those of analysis of
the whole 14-mer phosphorylated by this enzyme (Fig. 4B).
These data support our contention that the 180-kDa cytosolic ProT
K
is that responsible for the phosphorylation of ProT
in
vivo.
We next carried out
experiments aimed at characterizing the cytosolic ProTK and also the
kinase from the nuclear envelope fraction which, although apparently
not responsible for the phosphorylation of ProT
in vivo,
shows similar specificity to that of the cytosolic enzyme. Since the
specificities of both enzymes implied a consensus phosphorylation site
identical to that of CK-2 (29), the first step was to investigate the
behavior of the two enzymes with CK-2 substrates and effectors. The
results of kinase activity assays performed with dephosphorylated
casein and with the synthetic peptide RRRDDDSDDD, an artificial CK-2
substrate, are shown in Fig. 5A. The two
kinases clearly differ in substrate specificity, since the nuclear
envelope kinase phosphorylated both substrates while the cytosolic
ProT
K did not phosphorylate either. Experiments involving
immunodepletion with antibodies to the
subunit of CK-2 (Fig.
5B) confirmed the difference between the two kinases, since
only the activity of the nuclear envelope kinase was blocked. Moreover,
the behavior of the nuclear envelope enzyme in these experiments,
together with the molecular size data, suggest that this enzyme is in
fact CK-2. Phosphopeptide mapping and phosphoamino acid analysis of
ProT
phosphorylated by CK-2 from rat liver (data not shown) gave
identical results to those obtained with the nuclear envelope kinase
(see Fig. 4), which corroborates this conclusion. The antibody to the
subunit of CK-2 was likewise ineffective for depleting the minor
ProT
phosphorylation activity present in the HAF of nucleoplasm and
in the fraction of the cytosolic extract with moderate affinity for
ProT
-Sepharose (see Fig. 1, nucleoplasm panel, lane 4,
and cytosol panel, lane 2); as indicated above, these
activities were undetectable after ion-exchange HPLC. This suggests
that this minor kinase activity may correspond to the main 180-kDa
ProT
kinase.
The response of the cytosolic ProTK to CK-2 effectors was, however,
quite similar to that of CK-2; it was markedly inhibited by heparin (at
0.2-0.4 µg/ml) and activated up to 4-fold by protamine or polylysine
at 0.04-0.08 µg/ml. Kinetic study indicated that the cytosolic
ProT
K has a Km for ProT
of 50 µM
and a Vmax of 235 pmol/min/mg. The enzyme uses
ATP (with a Km of 55 µM) as phosphate
donor; GTP is an inefficient donor.
Determination of cytosolic ProTK activity in extracts obtained at
different periods (0-24 h) after the start of mitogenic activation of
non-synchronized splenic lymphocytes indicated a linear increase of up
to 10-fold over the first 16 h of activation, activity
subsequently remaining roughly constant over the next 8 h (results
not shown). SDS-PAGE analysis of the HPLC-purified enzyme (results not
shown) indicate that the kinase activity seems to reside in two
proteins of 64 and 60 kDa; taken together with the gel filtration data
(see Fig. 3B), this suggests that the enzyme has an
oligomeric structure in vivo.
We have recently reported that ProT interacts with core histones
in vitro (24). It was thus of interest to investigate whether the cytosolic ProT
K phosphorylates histones. Kinase activity assays using histone H1 and core histones from calf thymus as substrates (Fig. 6) indicated that histones H2B and H3
are phosphorylated by the ProT
K. On the basis of densitometric
scanning of the autoradiograph, H2B and H3 phosphorylation activities
were, respectively, 2 and 4.5 times lower than ProT
phosphorylation
activity. The cytosolic ProT
K is quite stable in the fractions from
affinity chromatography (about 2 months at 4 °C, 7 months at
70 °C), but shorter-lived after purification by ion-exchange HPLC
(about 8 days at 4 °C, 3 weeks at
70 °C).
ProT
The ProT
kinase was isolated from other cell types following the same procedures
as for mouse splenocytes. Subcellular fractionates obtained from
hepatocytes and mitogen-activated thymocytes (both from mouse),
semiconfluent HeLa cell cultures, and NC37 cells were subjected to
ProT
-Sepharose affinity chromatography, and ProT
phosphorylation
activity was assayed in the resulting eluate fractions. Kinase
activity, only detected in components with high affinity for ProT
,
was found to be appreciable in cytosolic fractions from all the cell
types tested. ProT
kinase activity was also detected in nuclear
envelope extracts of NC37 cells (results not shown), although substrate
specificity and immunodepletion with anti-CK-2 indicated that the
enzyme was CK-2, as in nuclear envelope extracts of splenocytes.
ProT
kinase activities in cytosolic fractions from the different
cell types are shown in Fig. 7. As can be seen from this
figure, activities appear to be dependent on the cell's proliferation
activity. Cells with moderate proliferation activity (hepatocytes,
mitogen activated thymocytes, and semiconfluent HeLa cells) showed
activity which was about 26-40% of that observed in mitogen-activated
splenocytes, while activity in extracts from NC37 cells was in the
range of that in mitogen-activated splenocytes. Fractions from
unstimulated thymocytes and confluent HeLa cells showed negligible
ProT
K activity (results not shown). Ion-exchange HPLC separation of
the cytosolic components with high affinity for ProT
from NC37,
semiconfluent HeLa, and mitogen-activated mouse thymocytes gave kinase
activity elution patterns similar to those observed with cytosolic
components from mouse splenocytes. Peptide mapping and phosphoamino
acid analysis of the HPLC-purified enzymes, following the same
procedure as for cytosolic ProT
K from splenocytes, indicated that
the cytosolic kinases from NC37, semiconfluent HeLa, and
mitogen-activated thymocytes show the same specificity for ProT
as
the splenocyte enzyme.
We have isolated and purified a 180-kDa protein kinase whose
specificity for ProT suggests that it is responsible for the phosphorylation of this protein observed in vivo in mouse
splenic lymphocytes and other cells (27, 28). The data presented in this study, particularly the affinity of the enzyme for Sepharose-bound ProT
, and the results of peptide mapping and phosphoamino acid analysis of ProT
phosphorylated in vivo and in
vitro by the purified enzyme, support this hypothesis.
Affinity chromatography on a ProT-Sepharose column has been shown to
be effective for isolating components which can interact with ProT
,
including core histones (24). The results presented here corroborate
the efficiency of this chromatographic system for isolating possible
cellular targets of ProT
. Thus, although ProT
is similarly
phosphorylated by the 180-kDa protein kinase and by CK-2, only the
former (Fig. 3A, upper panel) was separated by
ProT
-Sepharose affinity chromatography from cytosolic extracts of
the various cell types studied, in which both enzymes are present with
similar levels of activity. However, in nuclear envelope extracts,
which lack the 180-kDa kinase activity, the enzyme present (characterized as CK-2) was bound to ProT
-Sepharose (Fig. 3A, lower panel). We checked these points by evaluating the activity in cytosol and nuclear envelope extracts of the respective kinases, after purification by conventional chromatography on DEAE-cellulose and
phosphocellulose (data not shown).
That the protein kinase isolated from the cytosol of the various
mammalian cells is that responsible for the phosphorylation of ProT
in vivo is supported by our comparative analysis of the pattern of incorporation of [32P]orthophosphate into
ProT
in vivo or by the purified enzyme in
vitro (Fig. 4). In both cases (i.e. phosphorylated
in vivo or in vitro), 32P-ProT
was
identified by SDS-PAGE, isoelectric focusing, reverse-phase HPLC, and
immunoreactivity. Furthermore, the phosphoamino acid analysis and
peptide mapping were performed by two different procedures to guarantee
the identification of the phosphopeptides. These comparative analysis
not only indicate that the cytosolic kinase is that responsible for
phosphorylation of ProT
in vivo, but also indicate that
the kinase from nuclear envelopes, apparently identical to CK-2,
phosphorylates ProT
in vitro in such a way as to rule out
any role in the phosphorylation of this protein observed in
vivo.
Phosphorylation sites of the 180-kDa ProTK match consensus motifs
for CK-2 (29). Furthermore, the ProT
kinase is, like CK-2, highly
sensitive to inhibition by heparin and to activation by polyamines.
However, the purified enzyme cannot be included in the CK-2 family in
view of its structure (molecular size and subunit composition), and
given that it does not phosphorylate substrates of CK-2, is not
inhibited by anti-CK-2 antibodies (Fig. 5) and is unable to use GTP as
phosphate donor. There are also differences as regards subcellular
location (34-36). On the other hand, the ability of CK-2 to
phosphorylate acidic nuclear proteins structurally related to ProT
,
such as HMG-14 (37), protein P1 (38), and nucleolin (39), suggests that
this enzyme is related to ProT
K.
Interestingly, the ProTK phosphorylates histone H2B and, to a lesser
extent, histone H3 (Fig. 6). Although no biological significance can at
present be inferred from this result, it should be noted that these
proteins are not substrates of CK-2 (29) and that the pattern of
phosphorylation of ProT
by the ProT
K does not bear any
resemblance to that reported for kinases which phosphorylate H2B (40,
41). In addition, the kinases that phosphorylate this histone are cAMP
or cGMP-dependent (40); by contrast, signal transduction
experiments performed in our laboratory2
indicate that ProT
is not directly phosphorylated by protein kinase
A or C, and its phosphorylation in proliferating splenic lymphocytes
seems to be dependent on protein kinase C activity. The present data
thus suggest that the purified 180-kDa ProT
K does not fall into any
of the protein kinase categories described to date (41). That this is a
new enzyme should be confirmed by further investigation of its
structure and substrate specificity. It should be noted that the
widespread distribution of ProT
in mammalian cells, together with
its high concentration (0.03-0.15 pg/per cell), imply a central role
in the cell proliferation; it is thus quite reasonable to hypothesize
that there is a specific kinase for this protein.
An important question is whether the phosphorylation of ProT is
biologically significant. In view of the present data, we would argue
that it is, at least in proliferating cells, for two main reasons:
first, the existence of the protein kinase reported herein, which
appears to be highly specific for ProT
; second, the high
concentration of phosphorylated ProT
in proliferating cells. The
amount of [32P]orthophosphate incorporated into ProT
after 20 h of metabolic labeling of mitogen-stimulated mouse
splenocytes (in the conditions indicated in "Methods") was about
0.3 nmol/108 cells. As ProT
concentration in splenic
lymphocytes was about 0.8 nmol/108 cells (in accordance with
estimates reported by other authors; Refs. 5 and 9), and since peptide
mapping suggests that only two Thr residues are phosphorylated in
ProT
, the above finding suggests that about 18% of ProT
molecules are phosphorylated in mitogen-stimulated mouse splenocytes
(assuming no phosphorylation before labeling and a 100% yield in the
recovery of 32P-ProT
from labeled cells). Finally, the
direct relationship between the activity of the ProT
K and the
concentration of phospho-ProT
in proliferating cells (present
results; and Ref. 27) argues for a significant biological role.
Phosphorylation makes ProT more similar to other structurally
related acidic nuclear proteins which likewise undergo phosphorylation in vivo, such as nucleoplasmin (42), nucleolin (39), P1
(37), and HMGs (37). However, the putative effects of phosphorylation on the behavior of ProT
are unknown. Interestingly, the recently demonstrated capacities of ProT
to interact with histones and to
enable nucleosome assembly activity in vitro (24) do not appear to be affected by phosphorylation. Although this does not rule
out the possibility that phosphorylation affects in vivo interactions between ProT
and histones or other molecules in the
cell, it seems probable that phosphorylation is involved in some other
aspect of ProT
function, such as degradation to yield thymosins
(10) or nuclear import. In this connection, it should be noted that the
rate of nuclear import of some proteins is regulated by phosphorylation
at putative CK-2 sites (43). In this sense, ProT
could be included
in the group of proteins (nucleoplasmin, lamins, etc.) whose nuclear
activity is regulated by phosphorylation of CK-2 consensus site motifs
(44). Clearly, both structural and biochemical characterization of the
180-kDa ProT
K, together with elucidation of the effects of
phosphorylation on the behavior of ProT
may help to shed light on
the function of this protein.