ABSTRACT
We have developed a simple method to purify sequence-specific DNA-binding proteins directly from crude cell extracts by using DNA
affinity latex beads. The method enabled us to purify not only DNA-binding proteins, but also their associated proteins. Using beads bearing
the ATF/E4TF3 site from the adenovirus E4 gene promoter, a protein kinase
activity was copurified with the ATF/E4TF3 family. We found that the kinase
interacted with ATF1
in vitro
efficiently. The kinase did not bind directly to DNA. The kinase mainly
phosphorylated ATF1 on serine 36, which was one of target amino acids for
casein kinase (CK) II. Biological features of the kinase were the same as those
of CKII and an anti-CKII serum reacted with the kinase, indicating that the kinase was CKII.
Moreover, it was clearly shown that one of CKII subunits, the CKII
[alpha]
protein bound to glutathione-
S
-transferase (GST) fusion ATF1 but not GST
in vitro
. It has been reported that a specific CKII inhibitor, 5,6-dichloro-1-
[beta]
-D-ribofuranosylbenzimidazole (DRB) inhibits transcription by RNA
polymerase II [Zandomeni
et al
., (1986)
J. Biol. Chem
. 261, 3414-3419]. Taken together, these results suggest that ATF/E4TF3 may recruit
the CKII activity to a transcription initiation machinery and stimulate
transcription.
Generally, transcriptional activators are known to regulate the rate of
initiation by interacting activators with general transcription factors (
1
-
3
). One of the strong transcription activators, Gal4-VP16, which is often used to investigate the transcriptional activation
mechanisms, stimulates transcription dependent on the Gal4 binding site
upstream of the transcription initiation site in the promoter region
in vivo
and
in vitro
(
4
-
6
). In this case, the transcription activation domain of VP16 binds to various
factors in the preinitiation complex, transcription factor (TF)IID (
7
,
8
), TFIIB (
9
,
10
), TFIIH (
11
) and a coactivator, PC4 (
3
), and these interactions are known to be required for increasing the rate of
initiation. However, Yankulov
et al
. have recently reported that transcription activators stimulate transcription
by increasing not only the rate of initiation but also the processivity of RNA
polymerase II (
12
). They have found that Gal4-VP16 and Gal4-E1a can stimulate transcription by enhancing processivity
in vivo
. Although the mechanisms remain unclear, the transcription activators can
recruit elongation factors to the initiation complex. In these contexts, there
is one reasonable interpretation that protein kinase activities are involved in
the regulation of the processivity, since stimulation of processivity is
specifically inhibited by one of the kinase inhibitors, the purine nucleoside analogue 5,6-dichloro-1-[beta]-D-ribofuranosylbenzimidazole (DRB) (
13
,
14
). For the reasons mentioned above, we proposed that some transcription
activators could stimulate transcription by recruiting the kinase activities to
enhance the processivity. It was reported that the transcription factor E2F,
responsible for adenovirus E2A gene expression, associates with the cell cycle
regulating factors, the retinoblastoma protein (pRb) and its related p107
protein, and pRb or p107 containing E2F complexes also contain cyclin E-cdk2 or cyclin A-cdk2 kinases (
15
and references therein). These E2F complexes have sequence-specific DNA-binding activity, implying that E2F can recruit the protein kinase
activities to the position in the vicinity of the preinitiation complex (
16
). In addition, as mentioned above, VP16 can also interact specifically with
TFIIH which has kinase activities to phosphorylate the CTD (
11
,
17
). These data support our speculation that some activators have an ability to
recruit the protein kinase activities to the initiation complex.
We have recently purified and characterized E4TF3, responsible for adenovirus E4
gene transcription, using DNA affinity latex beads (
18
). We showed that E4TF3 purified by the beads stimulated transcription
in vitro
, dependent on binding to E4TF3-binding sites. At least eight polypeptides with different molecular weights 116, 80, 65, 60, 55, 47, 45 and 43 kDa were copurified (
18
). The complex of polypeptides bound specifically to the E4TF3 sequence,
implying that E4TF3 consists of a family of related factors. We now know that
E4TF3 comprises the ATF/CREB family of proteins. Immunoblotting analysis showed that the 65, 47 and 43 kDa species are
identical to ATF2/CRE-BP1, CREB and ATF1 respectively (unpublished data). It has been shown that
the members of the ATF/CREB family are phosphorylated by several protein
kinases and phosphorylation modulates the transcriptional activity. Casein kinase (CK) II (
19
,
20
), calmodulin/Ca
2+
-dependent protein kinase (
21
,
22
), cAMP-dependent protein kinase (PKA) (
23
-
26
), protein kinase C (PKC) (
26
) and DNA-dependent protein kinase (DNA-PK) (
20
) have been reported to phosphorylate either ATF2/CRE-BP1, CREB or ATF1 at several sites
in vitro
.
In this paper, to analyze kinases associated with ATF/E4TF3, we performed the purification of ATF/E4TF3 by using the DNA affinity latex beads
directly from HeLa cell nuclear extracts. Interestingly, we found the protein kinase activity involved in ATF/E4TF3. The kinase activity
was able to phosphorylate ATF1
in vitro
and retained on glutathione-
S
-transferase fusion (GST) ATF1 but not GST. Moreover, the kinase activity
did not bind to latex beads bearing the ATF/E4TF3 site, implying that the
kinase could not bind to DNA by itself. Surprisingly, the kinase activity had
the same biological features as those of CKII, and the kinase fraction
contained all subunits ([alpha], [alpha]' and [beta]) of CKII (
27
), indicating an identity of the kinase with CKII. We found also that the CKII [alpha] protein actually bound to ATF1
in vitro
. Zandomeni
et al
. have reported that CKII is involved directly or indirectly in the inhibition by DRB of specific RNA polymerase II-mediated transcription (
28
). Furthermore, DRB is a selective inhibitor of CKII (
29
). Therefore, we thought that ATF/E4TF3 can recruit the CKII activity to the
initiation complex and this may regulate the transcriptional activity.
HeLa spinner cells were grown in MEM containing 10% horse serum as described
previously (
30
).
Nuclear extracts were prepared according to the method of Dignam
et al
. (
31
). Whole cell extracts were prepared according to the method of Manley
et al
. (
32
).
Latex beads carrying E4TF3 sites were prepared as described previously (
18
).
The GST-ATF1 expression vector (pGEX-ATF1) was constructed by inserting a polymerase chain reaction (PCR)-generated fragment that contains ATF1 sequences from 190 to
1002 nucleotides into the plasmid pGEX-2T (Pharmacia) between
Sma
I and
Eco
RI sites (
33
).
The expression and purification of GST and GST fusion protein using glutathione-Sepharose were described previously (
34
).
The bacterial expression vectors and the purification of the histidine tagged
fusion protein were described previously (
20
). Serine 36, 38 and 41 are changed to alanine. Each ATF1 mutant has a
mutation(s) at the indicated position of serine (
20
).
Purification procedures were performed as described previously (
18
). Fifteen milliliters of HeLa cell nuclear extracts (8 mg/ml protein) were
mixed with the latex beads bearing E4TF3 sites. After 30 min at 4oC, the latex beads were collected by centrifugation, and were washed with
TGEN buffer [50 mM Tris (pH 8.0), 20% (w/v) glycerol, 1 mM EDTA, 1 mM
dithiohreitol (DTT), 0.1% Nonidet-40 (NP-40)] containing 0.1 M KCl. Bound proteins were eluted with TGEN
buffer containing 1.0 M KCl. After dialysis of the eluate against TGEN buffer
containing 0.1 M KCl and 12.5 mM MgCl
2
, the dialysate was loaded onto a Mono Q HR5/5 column (Pharmacia; 1 ml) at a flow rate of 1.0 ml/min. The column was washed with 20 ml TGEN buffer containing 0.1 M KCl and 12.5 mM MgCl
2
. Proteins were eluted by 10 ml of a linear gradient of 0.1-0.5 M KCl. Fractions were collected every 30 s. Fractions containing
activity from the Mono Q step were pooled and aliquots (150 [mu]l) were mixed with 50 [mu]l TGMN buffer [50 mM Tris (pH 7.9), 1 mM EDTA, 12.5 mM MgCl
2
, 1 mM DTT, 0.1% NP-40]. The diluted sample was loaded on 4.8 ml 15-35% glycerol gradient. After centrifugation using SW 50.1 rotor
(Beckman) at 48 000 r.p.m. for 24 h at 4oC, gradient was divided into 25 fractions. These fractions were analyzed by
the protein phosphorylation assay. Fractions 18-24 were pooled and used as the kinase fraction.
Phosphorylation of protein was performed at 30oC in a reaction mixture (10 [mu]l) containing 50 mM KH
2
PO
4
-K
2
HPO
4
(pH 7.2), 10 mM MgCl
2
, 3-5% glycerol, 100 mM KCl and 100 [mu]l ATP containing 1 [mu]Ci [[gamma]-
32
P]ATP (3000 Ci/mmol). The reaction was terminated by the addition of 5 [mu]l 4* Laemmli sample buffer and boiled for 5 min. Proteins were subjected
to sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and stained by Coomassie
brilliant blue and the gel was dried for autoradiography. During purification,
protein kinase activity was monitored with GST-ATF1 (0.05 mg/ml) as the substrate.
Quantitation of phosphate incorporation was determined with an AMBIS
radioanalytic system.
Protein kinase A catalytic subunit was purchased from Sigma Chemical Co.
CKII purified from pig testes was kindly provided by Dr Nakaya (
35
).
Two kinases, PKA (0.1 U) and the kinase fraction (1 [mu]l) were used for phosphorylation of GST-ATF1 respectively.
32
P-labeled GST-ATF1 (200 ng) was fractionated on a 10% SDS polyacrylamide gel, and the gel containing the labeled GST-ATF1 was excised and washed vigorously by shaking in 500 [mu]l of 25% 2-propanol for 6 h. Then the gel was washed in 500 [mu]l 10% methanol for 2 h on a rotator. After drying the gel, 50 [mu]l HCl was added and incubated at 110oC in an oil bath for 2 h (
36
). The solution was lyophilyzed and resuspended in 5 [mu]l H
2
O. Aliquots (2 [mu]l) of sample were mixed with a phosphoamino acid standard, then subjected to thin layer electrophoresis using Pharmacia Multiphor II Electrophoresis
system. After drying, standards were visualized by spraying with 5% ninhydrine
in methanol and incubated at 65oC for 5 min. Radiolabeled amino acids were detected by autoradiography.
Glutathione-Sepharose beads (25 [mu]l) that had been incubated with bacterial supernatant fractions
containing GST (25 [mu]l) and GST-ATF1 (125 [mu]l) were washed three times with NETN buffer [20 mM Tris (pH 8.0),
100 mM NaCl, 1 mM EDTA, 0.5% NP-40] and once with TGEN buffer containing ~1 mg/ml immobilized GST and GST-ATF1. Twenty five microliters of the kinase fraction was
diluted with an equal volume of TEMN buffer. The diluted samples were mixed
with 25 [mu]l of GST-coupled beads to remove non-specific binding proteins, and the flow-through fractions were collected. The beads bearing
immobilized GST and GST-ATF1 were incubated with aliquots (15 [mu]l) of the flow-through fraction at 4oC for 60 min and were washed five times with TGEN buffer containing 0.1 M KCl, before bound proteins were eluted
with 25 [mu]l TGEN buffer containing 1.0 M KCl and 12.5 mM MgCl
2
. The ATF1 kinase activity of each sample was analyzed.
35
S-labeled CKII [alpha] protein was synthesized
in vitro
by transcription of CKII [alpha] cDNA (pET-CKII [alpha]) by using T7 RNA polymerase (Takara) and subsequent translation in rabbit
reticulocyte lysate basically according to the manufacturer's instructions (Promega). The
CKII [alpha] expression vector, pET-CKII [alpha], was constructed by inserting a PCR-generated fragment that contains human CKII [alpha] sequences from 157 to 1332 nucleotides (
37
) into the plasmid pET14b between
Nde
I and
Bam
HI sites.
35
S-labeled CKII [alpha] protein was incubated with beads (25 [mu]l) bearing ~1 mg/ml immobilized GST and GST-ATF1 for 60 min at 4oC. After washing beads, bound proteins were
eluted by boiling in 1* SDS sample dye and subjected to SDS-PAGE and autoradiographed.
CKII was purified from whole cell extracts prepared as described by Manley
et al
. (
32
). Forty milliliters of the extracts (10 mg/ml proteins) were loaded onto a 40
ml heparin-agarose column equilibrated with HGEN buffer (20 mM Hepes, 20% glycerol,
1 mM EDTA, 1 mM DTT, 0.1% NP-40) containing 0.1 M KCl and 12.5 mM MgCl
2
. Bound proteins were eluted stepwise as described previously (
30
). The E4TF3 DNA-binding activity was completely eluted at 0.4 M KCl (
30
). The CKII activity was eluted at 1.0 M KCl. The active fractions were pooled
and proteins were precipitated by ammonium sulfate (final 40%). Then the
precipitate was resuspended with 2 ml HG(10)EN (HGEN except 10% glycerol)
buffer containing 0.1 M KCl and 10 mM MgCl
2
and loaded onto a Superdex 200 HR (10/30) column (Pharmacia) at a flow rate of 0.5 ml/min. The CKII activity was eluted from 7 to 13 ml after injection. The active fractions were pooled and loaded on a Mono
SPC16/5 column (Pharmacia; 0.1 ml) equilibrated with HG(10)EN buffer containing
0.1 M KCl and 10 mM MgCl
2
at a flow rate of 0.1 ml/min by using the SMART System (Pharmacia). Proteins
were eluted by 2.2 ml of a linear gradient of 0.1-0.5 M KCl. The CKII activity was recovered separately in flow-through fractions and gradient fractions eluted between 0.22 and
0.28 M KCl. The gradient fractions were pooled and dialyzed against TGEN buffer
containing 0.1 M KCl and 10 mM MgCl
2
. The flow-through fractions and the dialysate were loaded independently onto a Mono
QPC16/5 column (Pharmacia, 0.1 ml) at a flow rate of 0.1 ml/min and proteins
were eluted by 2.2 ml of a linear gradient of 0.1-0.5 M KCl. In both cases, the CKII activity was eluted at the same KCl
(0.29 M) and used for further analysis as HeLa cell CKII.
Immunoblotting was carried out as described previously (
30
). Proteins were subjected to SDS-PAGE and the fractionated proteins were transferred to Immobilon
(Millipore) by wet electrophoretic transfer. After blocking the membrane with
blots, the membrane was soaked with 500-fold diluted anti-CKII rabbit serum kindly provided by Dr Kikkawa (
38
). Then the membrane was developed according to the manufacturer's instructions
(Amersham).
We purified the ATF/E4TF3 family of proteins directly from HeLa cell nuclear
extracts by using DNA affinity latex beads having ATF/E4TF3 binding sites (
18
). To examine whether the purified fraction contained protein kinase activities,
we carried out protein phosphorylation assay by adding [[gamma]-
32
P]ATP to aliquots of the fraction. As shown in Figure
1
, we observed two labeled bands (of 43 and 60 kDa) on SDS-PAGE after incubation, indicating the presence of a protein kinase in the
affinity-purified E4TF3 fraction. The affinity-purified fraction contained at least eight distinct polypeptides as
described previously (Fig.
1
, lane 2) (
18
). Interestingly, one of the phosphorylated proteins seemed to correspond to the
43 kDa protein known to be ATF1. We have not yet characterized the 60 kDa
protein. This result suggested that the copurified kinase activity with the
ATF/E4TF3 family probably phosphorylated the 43 kDa protein.
We usually washed the beads five times with buffer containing 0.1 M KCl to purify E4TF3 from nuclear extracts (
18
), although the washing step reduced a yield of the kinase activity (data not
shown). Competitors [single strand DNA and poly(dI-dC):poly(dI-dC)], were also omitted, because they also reduced the activity
(data not shown). The purification of the kinase activity was monitored using
GST-ATF1 as the substrate in the protein phosphorylation assay. The kinase
activity phosphorylated efficiently GST-ATF1 protein but not GST protein (data not shown).
Next, the eluate from beads was dialyzed and loaded on a Mono Q column. Protein kinase activity was eluted with a 10 ml linear gradient of 0.1-0.5 M KCl (Fig.
2
). The ATF1 kinase activity peak eluted between 0.22 and 0.28 M KCl; most of the
DNA binding activity, however, was detected in the flow-through fractions (data not shown), suggesting that the kinase loosely
associated with ATF/E4TF3.
The kinase activity was isolated from the E4TF3 fraction purified using beads, presumably through an ability to interact with members of the
ATF/E4TF3 family. To confirm the interaction
in vitro
, the kinase fraction was incubated separately with GST and GST-ATF1. After extensive washing, the bound proteins were released from the
beads by the addition of high salt. The input samples and materials that eluted
from the beads were subsequently assayed for ATF1 kinase activity and the
products were analyzed by SDS-PAGE and autoradiography. Figure
3
A shows that the kinase activity was selectively retained on GST-ATF1, but not on GST columns, indicating that the kinase activity
associated with ATF1.
Recently, Masson
et al
. have demonstrated that the critical serine residues map to a putative
transcription activation domain of ATF1 using various ATF1 mutants (
20
). They showed that serine 36, 38 and 41 were involved
in vitro
phosphorylation of ATF1. Therefore, we used some of the ATF1 mutants to
identify the site(s) of phosphorylation catalyzed by the kinase activity. Histidine tagged fusion ATF1 (H6-ATF1) and mutants proteins were isolated (
20
) and incubated with the kinase fraction (Fig.
4
A). Changing serine 36 to alanine (Mut 36) caused a reduction of the
susceptibility of the kinase activity and Mut 36/38 showed a greatly reduced susceptibility of the
kinase activity compared with Mut 36. However, the kinase activity efficiently
phosphorylated Mut 38/41 as well as wild-type ATF1, strongly suggesting that the major phosphorylation site of the
kinase activity was serine 36 in the P box. We also observed that
phosphorylations decreased the mobility of ATF1 in SDS gels (Fig.
4
A). This was also represented by using
in vitro
translation products of ATF1 mRNA in rabbit reticulocyte lysate (
20
). Significantly, mutation of serine 36 alone had a dramatic effect on mobility
(Fig.
4
A), and double mutants Mut 36/38 and Mut 36/41 both gave a greater effect as
compared with Mut 36. Interestingly, the mobility of Mut 38/41 on SDS gels was
similar to that of ATF1, suggesting that serine 36 is most important for
phosphorylation of the kinase activity and a phosphorylation-induced conformational change in ATF1. Furthermore, serine 36 which is
lacking in CREB and CREM, is known to be phosphorylated by CKII
in vitro
as reported previously (
20
). To compare the phosphorylation states of these mutants, therefore, we used
CKII purified from pig testes (
35
). As seen in Figure
4
A, phosphorylation states of mutants by pig CKII were almost the same as that of
mutants by the kinase activity. This result suggests that phosphorylation site
sequences are highly homologous between the kinase activity and CKII.
CKII was monitored during several successive chromatography steps, as outlined
in Figure
5
A, by immunoblotting and protein phosphorylation assays. After gel filtration,
the CKII fractions were pooled and loaded onto a Mono S column using the SMART
system. The CKII activity was observed in two different peaks, in flow-through fractions and gradient fractions which eluted between 0.25 and 0.3
M KCl (Fig.
5
C) (see Discussion). Both active fractions were independently loaded on a Mono Q
column and proteins were eluted by a linear gradient. Surprisingly, in both
cases, the CKII activity was eluted in the same fraction number (Fig.
5
D and E). Proteins included in the fractions were subjected to SDS-PAGE, and stained by silver (Fig.
5
B, lanes 1 and 2). Subunits of CKII were detected in the fractions by an
immunobotting assay (Fig.
5
B, lanes 3 and 4). Three bands with molecular weights of 44, 40 and 28 kDa were observed mainly by silver staining, and these
bands reacted specifically with the anti-CKII polyclonal antibody, indicating that the final fractions contained [alpha], [alpha]' and [beta] subunits of CKII with molecular weight of 44,
40 and 28 kDa respectively (
27
). Therefore, we used the fraction as HeLa cell CKII. To confirm the HeLa cell
CKII activity, we performed the same experiments as described in Figure
4
. The HeLa cell CKII activity specifically inhibited by a low concentration of
heparin and could use GTP as a phosphate donor. Furthermore, phosphorylation
states of ATF1 mutants by HeLa cell CKII were completely the same as those of
mutants by pig CKII (data not shown).
We further carried out an immunoblotting assay using the anti-CKII polyclonal antibody (
38
). As seen in Figure
6
, the kinase fraction contained [alpha], [alpha]' and [beta] subunits of CKII (lanes 1 and 3) as observed in HeLa
cell CKII (lanes 4 and 5), indicating an identity of the kinase with CKII. The
ratio of [alpha]:[alpha]' was different from the result in Figure
5
B. We think that this discrepancy was probably caused by different detection
methods: we used a color detection method using horse radish peroxydase and
diaminobenzine-HCl in Figure
5
, but an ECL method was performed in Figure
6
. However, distribution of all subunits of CKII are completely the same between
HeLa cell CKII and the kinase fraction. These results showed an identity of the
kinase to CKII.
We have found a protein kinase activity in highly purified ATF/E4TF3 samples
capable of phosphorylating the 43 kDa protein, ATF1 which is a member of the
ATF/E4TF3 family (
33
). The kinase activity does not have a DNA binding activity but can interact
with ATF1, and this specific interaction leads to copurification of the kinase
activity with the ATF/E4TF3 family when ATF/E4TF3 was purified by using the DNA
affinity latex beads. The kinase was CKII and one of the CKII subunits, CKII [alpha], bound to ATF1
in vitro
, suggesting a possible involvement of CKII in transcription.
In this study, we purified CKII partially from HeLa cell extracts. At the step
of the Mono S column, the CKII activity was divided into two parts, flow-through and gradient elution fractions (Fig.
5
). CKII is known to associate with a variety of cellular proteins, topoisomerase
I, topoisomerase II, p53, Hsp 70, nuclear matrix proteins and unknown nuclear
proteins (
39
and references therein). These associations probably caused the separation of
the CKII activity. Actually, the CKII activity was recovered in fractions
containing proteins <= 200 kDa molecular weight range by velocity sedimentation centrifugation
(Fig.
2
). It has been reported that the TFIID activity also separately eluted from a
phosphocellulose column by the difference of components of TFIID (
40
).
Various protein kinases including CKII, PKA and DNA-PK are known to phosphorylate ATF1
in vitro
(
20
,
33
). No direct interaction between the kinases and ATF1 has been reported.
Interestingly, our identified protein kinase, CKII, could associate with ATF1
in vitro
(Fig.
7
) and
in vivo
(data not shown). Recently, we have analyzed the interaction between CKII and ATF1 by using GST pull-down assay and Far western blotting. These analyses demonstrated that the catalytic [alpha] subunit of CKII directly interacts with ATF1, ATF2 and CREB (
41
). Furthermore, this interaction is functionally important for ATF1, since CKII-mediated phosphorylation of ATF1 was dependent on the binding of CKII to
ATF1. However, other members of the ATF/E4TF3 family, ATF2 and CREB, were not
efficiently phosphorylated by CKII even though CKII bound directly to them
(data not shown).
Recently, several studies have reported effects of the CKII activity on transcriptional regulation. It has been shown that the specific function of
anti-oncogene product, p53, is to work as a transcription factor (
42
-
44
). Interestingly, the two proteins, CKII and p53, associated in a molecular
complex (
45
), suggesting that p53 can recruit the CKII activity to the initiation complex.
On the other hand, topoisomerase I, which is also involved in the regulation of transcriptional activity, specifically binds to CKII (
46
,
47
), and CKII is likely to regulate function of topoisomerase I (
47
). It is known that topoisomerase I is included in the TFIID complex (
48
). In addition, Robitzki
et al
. have reported that endogeneous [beta] subunit of CKII was increased when the [alpha] subunit of CKII was overexpressed by transfection in HeLa cells
and the CKII [alpha] subunit appeared to bind specifically to the element in the CKII [beta] gene promoter (
49
). They speculated that CKII [alpha] had an ability to stimulate transcription depending on its binding
element. Taken together, these results strongly suggested that action of CKII
is directly or indirectly involved in transcription and we thought that ATF1
may serve as a scaffold for CKII which can phosphorylate proteins in the
vicinity of the ATF1 site and these phosphorylations may function in
transcription. Actually, the target sequence of CKII is observed among several
proteins in general transcription factors, TFIIA [alpha], TFIIE, the large subunit of TFIIF and the largest subunit of RNA
polymerase II (
50
). Interestingly, DRB inhibits RNA polymerase II-driven transcription specifically (
28
,
51
,
52
).
In vitro
experiments have shown a close correlation between the DRB-sensitivity of CKII and transcription (
28
). Recent studies have also shown that transcriptional activators affected the
efficiency of elongation in transcription (
12
) and this activity specifically inhibited by DRB (
13
,
14
), however, the mechanisms remain unclear. In these contexts, our results in this paper may suggest that the transcription activator ATF/E4TF3 can
recruit the CKII activity in the promoter region and then CKII can
phosphorylate factors which affect the efficiency of elongation rather than the
rate of initiation. In addition, our preliminary results indicated that action
of CKII is really involved in transcription and the inhibition by DRB is due to the
inhibition of the CKII activity (
53
). Therefore, we believe that analysis of an inhibition mechanism by DRB helps
us to investigate the nature of CKII in transcription.
Recently, P-TEF (positive transcription elongation factor) was identified by
reconstruction of DRB-sensitive transcription using
Drosophila
transcription factors (
14
). P-TEF was originally identified as a DRB-sensitive factor and its function was necessary for processivity (
14
). By fractionation of
Drosophila
melanogaster
Kc cells, the P-TEF activity were separated into three activities; P-TEFa, factor 2 and P-TEFb and P-TEFb were purified (
54
). P-TEFb was composed of two subunits with apparent molecular masses of 124
and 43 kDa and essential for processivity
in vitro
. They showed that P-TEFb acts after initiation and is the limiting factor in the production of
long run-off transcripts (
54
). Perhaps P-TEFb is a kinase since DRB is canonically a kinase inhibitor, however, the
subunit composition of P-TEFb displays no obvious similarity to known kinases, especially CKII (
54
).
We are indebted to Drs R. Treisman and H. Watanabe for helpful discussions. We
thank Dr Kikkawa for supplying an anti-CKII serum. We are grateful to Dr Nakaya for providing Pig testes CKII.
The work was supported by a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of
Education, Science and Culture of Japan and by a grant of Japan Health Science
Foundation to H.H. and was supported for the Promotion of Science for Japanese
Junior Scientists to T.W.
REFERENCES
Return