(Received for publication, March 11, 1997, and in revised form, May 15, 1997)
From the Department of Biology and Center for
Molecular Genetics, University of California at San Diego, La
Jolla, California, 92093-0322 and § Department of Medical
Oncology, Fox Chase Cancer Center,
Philadelphia, Pennsylvania 19111
The largest subunit of RNA polymerase II contains a C-terminal repeated domain (CTD) that is the site of phosphorylation by serine (threonine) and tyrosine kinases. Phosphorylation of the CTD is correlated with transcription elongation. A number of different kinases have previously been shown to phosphorylate the CTD; among them is a nuclear tyrosine kinase encoded by the c-abl proto-oncogene. The processive and high stoichiometric phosphorylation of RNA polymerase II by c-Abl requires the tyrosine kinase, the SH2 domain, and a CTD-interacting domain (CTD-ID) in the Abl protein. The physiological tyrosine phosphorylation of RNA polymerase II by c-Abl in DNA damage response has previously been demonstrated. Basal tyrosine phosphorylation of RNA polymerase II, however, is observed in cells derived from abl-deficient mice, indicating the existence of other CTD tyrosine kinases. In this report, we show that the tyrosine kinase encoded by an Abl-related gene (Arg) also phosphorylates the CTD in vitro and in transfected cells. The SH2 and kinase domain of Arg are 95% identical to that of c-Abl. However, these two proteins share only 29% identity in the large C-terminal region. Interestingly, a CTD-ID is also found in the C-terminal region of Arg. Mapping studies and sequence analysis have led to the identification of the CTD-ID that is highly conserved among the divergent C-terminal regions of Abl and Arg. These results indicate that tyrosine phosphorylation of RNA polymerase II CTD could be catalyzed by either c-Abl or Arg kinase.
The product of proto-oncogene c-abl and its related gene, arg, encode nonreceptor tyrosine kinases that are ubiquitously expressed in mouse and human cells (1-3). The two proteins are 95% identical in the N-terminal SH3, SH2, and tyrosine kinase domains. The C-terminal regions that constitute more than one-half of both proteins exhibit an overall identity of only 29% (2). This large C-terminal region that distinguishes c-Abl from the other family of nonreceptor kinases is required for the proper biological function of c-Abl. Truncation of the C terminus causes neonatal lethality in mice, a phenotype similar to the one observed with mice carrying homozygous null mutation for c-abl (4, 5). Several functional domains have been identified in the C-terminal region of c-Abl, including three nuclear localization signals (6), a DNA binding domain composed of three high mobility group-like boxes (7, 8), and binding domains for G- and F-actin (9, 10). Two physiological substrates for c-Abl have been identified for which substrate binding sites have also been identified in the C-terminal region. They are the CRK family of SH2/SH3 adapter proteins (11, 12) and RNA polymerase II (13-15).
The largest subunit of RNA polymerase contains a unique C-terminal
domain that is composed of a seven-amino acid repeat with the consensus
sequence YSPTSPS (16, 17). The heptad sequence is repeated 52 times in
mammals, 44 times in Drosophila melanogaster, and 26 times
in Saccharomyces cerevisiae (17). The
CTD1 of RNAP is essential for cell growth
because truncation of more than half of the repeats in yeast causes
cold sensitivity and inability to induce specific gene expression such
as INO 1 and GAL 10 (18). In mouse, a similar
truncation in an -amanitin-resistant RNA polymerase caused inability
of the polymerase to confer
-amanitin resistance (19).
Because the CTD is rich in serine, threonine, and tyrosine, it serves as a substrate for both serine (threonine) and tyrosine kinases. Several CTD kinases have been identified in yeast and mammals. In yeast, the cyclin-dependent kinase Kin 28, a component of the holo-TFIIH, has been shown to phosphorylate the CTD (20, 21). The mammalian homologue of TFIIH-associated CTD kinase has been shown to be the cdk7/cyclin H kinase pair (22). Another yeast CTD kinase, SRB10/11 kinase-cyclin pair, has been identified as the mammalian cdk8/cyclin C (23). Thus far, the only kinase known to phosphorylate RNAP-CTD on tyrosines is c-Abl (13-15). c-Abl can phosphorylate the CTD to high stoichiometry with the incorporation of >30 mol of phosphate/mol of CTD. Such a high stoichiometric phosphorylation by Abl requires binding of the SH2 domain of Abl to partially tyrosine-phosphorylated CTD (13). Furthermore, CTD phosphorylation by Abl both in vivo and in vitro requires a CTD-interacting domain (CTD-ID) present at the C terminus of Abl (15).
The strongest evidence that c-Abl phosphorylates RNA polymerase II in vivo came from the study of cellular response to DNA damaging agents such as methyl methanesulfonate (MMS). We have found that MMS activates c-Abl tyrosine kinase in S phase cells (24). We have also shown that MMS can cause an increase in the phosphotyrosine content of RNA polymerase II, but only when c-Abl is present (24). In Abl-null cells or in cells reconstituted with a kinase-defective c-Abl, MMS did not induce a significant increase in the tyrosine phosphorylation of the largest subunit of RNA polymerase II (24). However, Abl-null cells or cells expressing only the kinase-defective Abl contain basal levels of phosphotyrosine in the largest subunit of RNA polymerase II (14, 24). This suggested that basal tyrosine phosphorylation could be catalyzed by another tyrosine kinase. To identify the alternative CTD-tyrosine kinase, we tested the Arg gene product and showed that this tyrosine kinase can indeed phosphorylate the CTD.
COS cells were cultured at 37 C° in Dulbecco's modified Eagle's medium containing 10% supplemented calf serum (Hyclone Laboratories). Cells were transfected using LipofectAMINE (Life Technologies, Inc.) according to manufacturer protocol or using the DEAE-dextran method. Two days after transfection, the cells were incubated for an additional 24 h with 1 µM ouabain before harvesting (15).
PlasmidsThe wild type and mutant Arg cDNA (human type
IB) were cloned in pBS (Bluescript, Stratagene) vector. For transient
cotransfection studies, wild type and mutants of Arg were expressed
from a cytomegalovirus promoter-based vector CB6+ (25). The
Sma construct removed amino acids 1128-1182, whereas
Acc deleted
1043-1182,
Cla-Pflm1 removed 547-913,
Pflm1 deleted 913-1182,
and
Pflm1-Acc removed amino acids 913-1043. The cDNA of murine
CTD was subcloned into pEBG to obtain pEBG-CTD, which expresses the
GST-CTD in eukaryotic cell lines (15).
Monoclonal (8WG16) and polyclonal antibodies against RNAP and c-Abl have been described previously (15, 26). Polyclonal anti-Arg antibody has been described (25, 27). Monoclonal anti-Tyr(P) was purchased from ICN Pharmaceuticals. Horseradish peroxidase-conjugated secondary antibody was obtained from Life Technologies, Inc.
Immunoprecipitation and Kinase ReactionWild type and
mutant Abl or Arg proteins were obtained by in vitro
translation using TNT reticulocyte lysate (Promega) with either T3 or
T7 RNA polymerase. The translation mixture was diluted to give a final
concentration of 10 mM Tris-HCl (pH 7.4), 5 mM EDTA, 130 mM NaCl, 1% Triton X-100, 1 mM
phenylmethylsulfonyl fluoride, and 10 µg each of phenanthroline,
aprotinin, leupeptide, and pepstatin. Immunoprecipitation with anti-Arg
or anti-Abl antibody was carried out as described previously (15). The
mixture was incubated for 2 h at 4 °C, after which Protein
A/G-Sepharose beads were added and incubated for an additional 1 h. The immunoprecipitants were washed twice with Buffer A containing
500 mM NaCl, twice with Buffer A containing 100 mM NaCl, and twice with Buffer A alone. The immune
complexes containing similar amounts of wild type and mutant Arg or Abl
proteins were rinsed and suspended in 20 µl of kinase buffer (20 mM Tris-HCl (pH 7.4), 10 mM MgCl2, and 1 mM dithiothreitol) containing 0.2 µg of GST-CTD or
0.05 µg of purified HeLa RNAP IIA or IIB. Kinase reactions were
initiated by the addition of 10 µM ATP plus 25 µCi of
[-32P]ATP (7000 Ci/mmol; ICN). Reaction mixtures were
incubated at room temperature for 10 min (RNAP II) and 30 min (GST-CTD
substrates) before the addition of SDS sample buffer. The samples were
resolved on 8% SDS-polyacrylamide gel electrophoresis, transferred to
Immobilon-P, and exposed for autoradiography. Quantitation was done
using PhosphorImager (Molecular Dynamics).
The
identification of tyrosine-phosphorylated RNA polymerase II in Abl-null
cells indicated the presence of other tyrosine kinases capable of
phosphorylating RNAP-CTD (14). In human and in mouse, a c-Abl-related
gene product, Arg, has been cloned and sequenced (Fig.
1). In humans, the Arg gene product was identified as a
140-kDa protein that is widely expressed in various tissues (28). To
determine if Arg kinase could phosphorylate the CTD of RNAP in
vitro, human type IB Arg was translated in vitro,
immunoprecipitated, and incubated with purified RNAP IIA or IIB in the
presence of ATP (Fig. 2A). In the
lane containing RNAP IIA (lane 1), a radiolabeled band corresponding in molecular mass to subunit IIa could be detected. This suggests that Arg phosphorylated RNAP II. Subunit IIa was not
labeled when RNAP II was incubated with control immunoprecipitates (lane 3). To show that, indeed, it is the CTD of RNAP that
is the target of Arg tyrosine kinase, the C-terminal truncated subunit IIb form was also incubated with the Arg kinase. Under similar reaction
conditions, only IIa, but not the C-terminal truncated IIb, could be
phosphorylated by the Arg kinase (Fig. 2A, compare lane 2 with lane 1). The inability to
phosphorylate subunit IIb, which lacks the CTD but contains 32 tyrosines, strongly suggests that Arg specifically phosphorylates
tyrosine residues within the CTD. To test this directly, a GST-CTD
fusion protein expressed and purified from bacteria was also tested.
The GST-CTD was found to be an effective substrate for the Arg tyrosine
kinase since, under the condition assayed, a smeary band was observed
upon incubation of GST-CTD with Arg (Fig. 2B). The
observation that the GST alone was not phosphorylated indicates that
phosphorylation occurs within the CTD portion of the fusion protein.
Taken together, these results show that the CTD of RNA polymerase II is
a substrate for the Arg tyrosine kinase.
Both c-Abl and Arg Phosphorylate the CTD of RNAP with Equal Efficiency
In an effort to compare the ability of Abl and Arg to
phosphorylate the CTD, both the kinases were immunoprecipitated from in vitro translations and incubated with increasing
concentrations of GST-CTD in the presence of ATP (Fig.
3A). To compare the CTD-specific kinase
activity of these two kinases, a nonspecific substrate, namely
acid-denatured enolase, was also used. Under the conditions assayed,
Arg phosphorylated the enolase substrate 2.2-fold better than Abl
kinase (lanes 11 and 12). Fig. 3B
shows the incorporation of 32P into GST-CTD as a function
of increasing CTD concentrations. After normalizing for the equal
kinase activity input in the kinase assays, it can be seen that both of
these kinases phosphorylated the CTD equally well at the concentrations
of the substrate tested (Fig. 3B). These results show that
both Arg and c-Abl tyrosine kinase have the ability to phosphorylate
the RNAP II with equal efficiency in vitro.
A CTD-interacting Domain in Arg Tyrosine Kinase
Previously,
we have found that a CTD-ID located between the Sal and
Bgl sites in the C terminus of c-Abl is required for RNAP II
phosphorylation both in vitro and in vivo (15).
Deletion of the CTD-ID leads to an increase in the
Km of c-Abl for the CTD (15). The efficient
phosphorylation of the CTD by Arg kinase suggested that domains similar
to those present in c-Abl kinase, which are required for CTD
phosphorylation, must be present in the Arg kinase. In Arg, the SH2 and
kinase domain share >95% homology with Abl (Fig. 1). However, the
C-terminal domain of the Arg kinase differs considerably and exhibits
only 29% homology to the c-Abl C terminus. To determine regions in the
C terminus of Arg that are required for CTD phosphorylation, a series
of truncation and internal deletion mutants in the C terminus of Arg
kinase were constructed and tested for phosphorylation of the GST-CTD
fusion protein (Fig. 4). The CTD-specific kinase activity for each of these mutants was determined by normalizing the
kinase activity for each of these mutants with a nonspecific substrate,
enolase. Although specific mutations in the C terminus affected CTD
phosphorylation, they did not affect enolase phosphorylation. A
compilation of the results obtained with various mutants of Arg after
normalizing for equal protein input is given in Fig. 4. As expected,
C-terminal deletion mutants, Pflm1 (not shown) and
Acc (Fig.
5, lanes 5-7) in which 269 and 139 amino
acids were deleted could not phosphorylate the CTD efficiently.
However, deletion of 54 amino acids from the C terminus did not affect CTD phosphorylation (Fig. 5, lanes 9-11). This indicated
that the region between Acc and Sma is necessary
for efficient CTD phosphorylation. An internal deletion mutant
Pflm1-Acc in which amino acids 547-913 were removed exhibited CTD
kinase activity similar to that of full-length Arg (not shown). Taken
together, these results indicate that a specific region in the C
terminus of Arg kinase (between Acc and Sma,
amino acids 1043-1128) is required for efficient phosphorylation of
the CTD.
The CTD-ID Is Required for in Vivo Phosphorylation of the GST-CTD
Because Arg phosphorylates the CTD efficiently in
vitro, we tested whether Arg can phosphorylate RNAP-CTD in
vivo. For this purpose, Arg was cotransfected with GST-CTD
construct in COS cells, and the phosphotyrosine content of the CTD was
analyzed by reacting to anti-Tyr(P) antibodies. Anti-GST
immunoprecipitates prepared from cells transfected with GST-CTD in the
presence and absence of Arg kinase were probed with anti-CTD and
anti-Tyr(P) antibody. Cells transfected with GST-CTD alone contained a
protein band of the expected molecular weight for GST-CTD (Fig.
6A, lane 1). However, in the
presence of wild type Arg, a smeary shifted band was observed
(lane 2). Anti-Tyr(P) immunoblotting of a similar blot
indicated that the mobility-shifted band contained phosphotyrosine (panel C, compare lane 1 with lane 2).
The anti-Tyr(P) signal was completely eliminated when phosphotyrosine
was included in the immunoblotting reaction, indicating specificity of
the signals generated by the Tyr(P) antibodies (panel B).
Thus, Arg kinase phosphorylates the CTD in vivo.
To determine if the phosphorylation of the RNAP II CTD by Arg kinase
requires the region between Acc and Sma, the
assay was applied to relevant Arg mutant proteins overproduced in COS
cells. Mutants of Arg such as Sma and
Pflm1-Acc, which
phosphorylated the CTD efficiently in vitro, scored positive
in this assay (Fig. 6A, lanes 4 and
5). However,
Pflm1 and
Acc could not phosphorylate the
CTD (lanes 3 and 6) despite equal levels of
expression in these transfected cells (Fig. 6D). The low
molecular weight bands are most likely due to degradation. Since the
autophosphorylation ability of these mutants was not affected (not
shown), it can be inferred that it is the lack of CTD-ID that prevented
these mutants from phosphorylating the CTD in vivo.
A
comparison of the amino acid sequences between Acc and
Sma in Arg kinase to that of previously mapped CTD-ID
(Sal-Bgl) in Abl kinase indicated the presence of a highly conserved
amino acid stretch (see Fig. 8A). Previously, we mapped the
amino acids 935-1065 (Sal-Bgl) of Abl kinase C terminus to be the
CTD-ID (Fig. 1). The results obtained from Arg CTD-ID suggested that
the CTD-ID in Abl is between amino acids 1002 and 1065 (Sca-Bgl). To
determine if this was the case, an internal deletion mutant of Abl,
Sca-Bgl, was constructed and tested for its ability to phosphorylate
the CTD in vivo. Anti-GST immunoprecipitates from wild type
Abl as well as Bgl, but not
Sal-Bgl- and
Sca-Bgl-transfected
cells, contained the mobility-shifted CTD band (Fig.
7A, compare lanes 2 and
5 with 3 and 4). The presence of
phosphotyrosine in mobility-shifted GST-CTD was confirmed by
anti-Tyr(P) immunoblotting (panel C). The expression levels
of various Abl mutants shown in Fig. 7D indicate the
presence of approximately similar levels of expression of these
mutants. Together, these results allow us to further narrow down the
CTD-ID of Abl to the amino acids 1002-1065 (Sca-Bgl) region of the
kinase.
A CTD-ID Peptide Can Inhibit both Abl and Arg CTD Kinase Activity
If the highly conserved region in Arg and Abl C termini constitutes the CTD-ID, then addition of exogenous CTD-ID should compete for the CTD and, thus, inhibit Abl as well as Arg phosphorylation of the CTD. A synthetic peptide corresponding to an 18-amino acid stretch in the CTD-ID (Fig. 8A) was synthesized and added to in vitro Abl as well as Arg kinase assays. A nonspecific peptide that contained an identical amino acid composition, but scrambled sequence, was also used as a control. The nonspecific peptide had little effect on the ability of Abl or Arg to phosphorylate CTD (Fig. 8B, lanes 1-3 and 7-9). A small inhibition observed in the CTD kinase activity of Abl at 1:50 molar ratio could be a nonspecific effect. Nevertheless, both Abl and Arg CTD kinase activity was affected at each of the specific peptide concentrations tested (Fig. 8, compare lanes 4-6 with 10-12). These results provide additional support that the conserved C-terminal region between Arg and Abl kinase contains the CTD-interacting domain.
The CTD of RNA polymerase II is found to be a substrate for the Abl-related gene (arg) product, Arg. Full-length RNA polymerase II large subunit purified from HeLa cells and a bacterially expressed GST-CTD could be efficiently phosphorylated by the Arg kinase. On comparison, the CTD kinase activity of Arg was equivalent to that of c-Abl kinase with a similar value of Km in the 0.4-1.0 µM range. In addition, CTD phosphorylation by Arg is shown to require a CTD-interacting domain that has been previously identified in the C terminus of c-Abl kinase. Mapping studies and sequence comparison led to the assignment of the CTD-ID to a highly conserved sequence in the otherwise divergent C-terminal region of Abl and Arg. The identification of the CTD-ID was further confirmed by the inhibition of CTD kinase with a synthetic peptide containing the identified CTD-ID sequence.
The high degree of conservation of amino acid sequence in the SH2, kinase, and the CTD-ID region between Arg and Abl kinases indicates that these two kinases share substrate specificity for RNAP II phosphorylation. The c-Abl kinase has been shown to be present both in the nucleus and in the cytoplasm (9, 10). Wang and Kruh (25) recently reported that when overexpressed in COS cells, Arg is mostly localized in cytoplasm, with minor fractions of the protein present in the nucleus. Due to the lack of specific antibodies that can recognize endogenous Arg, it has not been possible to determine the subcellular localization of Arg in Abl-null cells. In any case, it is possible that Arg may substitute for Abl kinase in these cells in order to catalyze the basal tyrosine phosphorylation of RNA polymerase II. However, we cannot rule out that the basal tyrosine phosphorylation of RNA polymerase II in Abl-null cells is catalyzed by other tyrosine kinases that are unrelated to either Abl or Arg.
Recently, we have shown that several DNA damaging agents such as
cisplatin, IR (-irradiation), MMS, and mitomycin C can activate Abl
kinase activity, and this activation is correlated with an increase in
the phosphotyrosine content of RNAP II (24). When Abl-null cells were
exposed to DNA-damaging agents, no increase in the phosphotyrosine
content of RNA polymerase was observed (24). We have also shown that
that IR activation of Abl is dependent on the function of the
ataxia telangiectasia mutated gene product. In
AT patient cells, IR does not activate Abl and does not induce tyrosine
phosphorylation of RNAP II (29). If Arg is involved in maintaining the
basal tyrosine phosphorylation of RNA polymerase II, this result would
suggest that Arg may not be activated by DNA damage-induced signaling
event. Given the fact that Arg and Abl are highly divergent in the
C-terminal region except for the CTD-interacting domain, it is possible
that these two kinases may transduce different signals to mediate the
tyrosine phosphorylation of RNA polymerase II.
Tyrosine phosphorylation of the CTD has been correlated with the stimulation of promoter activity (15).The c-Abl tyrosine kinase is regulated by at least three signals. Integrin receptor-mediated adhesion to the extracellular matrix is required to activate c-Abl (30). In the nucleus, c-Abl activity is further regulated by retinoblastoma in the cell cycle (31, 32) and by DNA damage (24, 29). At present, it is not known what signals regulate Arg. Because c-Abl and Arg can phosphorylate RNAP-CTD, these tyrosine kinases can transduce signals directly to RNA polymerase II to regulate transcription.
We thank M. E. Dahmus for purified RNA polymerase IIA and IIB proteins, Baolin Wang for the construction of Arg mutants, and Steven Reed for the ouabain-resistant plasmid. We also acknowledge Laura Whitaker, Erik Knudsen, and William Eidenmuller for critical reading of the manuscript.