(Received for publication, March 4, 1997, and in revised form, May 1, 1997)
From the Department of Molecular Biology, The University of Medicine and Dentistry of New Jersey, School of Osteopathic Medicine, Stratford, New Jersey 08084
We have previously identified distinct nuclear
and mitochondrial isoforms of dUTPase in human cells, reporting
the cDNA sequence of the nuclear isoform (DUT-N). We now
report a cDNA corresponding to the mitochondrial isoform (DUT-M).
The DUT-M cDNA contains an 252-amino acid open reading frame,
encoding a protein with a predicted Mr of
26,704. The amino-terminal region of the protein contains an
arginine-rich, 69-residue mitochondrial targeting presequence that is
absent in the mature protein. In vitro transcription and
translation of the DUT-M cDNA results in the production of a
precursor protein with an apparent molecular mass of 31 kDa as judged
by SDS-polyacrylamide gel electrophoresis. The DUT-M precursor is
enzymatically active and immunoreacts with a dUTPase-specific monoclonal antibody. Mitochondrial import and processing studies demonstrate that the DUT-M precursor is processed into a 23-kDa protein
and imported into mitochondria in vitro. Isoelectric
focusing experiments demonstrate that the DUT-N has a pI of 6.0, while the processed form of DUT-M has a more basic pI of 8.1, measurements that are in agreement with predicted values. Studies aimed at understanding the expression of these isoforms were performed utilizing
quiescent and replicating 34Lu human lung fibroblasts as a model cell
culture system. Northern blot analysis, employing an isoform-specific
probe, demonstrates that DUT-N and DUT-M are encoded by two distinct
mRNA species of 1.1 and 1.4 kilobases, respectively. Western and
Northern blot analysis reveal that DUT-M protein and mRNA are
expressed in a constitutive fashion, independent of cell cycle phase or
proliferation status. In contrast, DUT-N protein and mRNA levels
are tightly regulated to coincide with nuclear DNA replication status.
Because DUT-N and DUT-M have identical amino acid and cDNA
sequences in their overlapping regions, we set out to determine if they
were encoded by the same gene. The 5 region of the gene encoding
dUTPase was isolated and characterized by a combination of Southern
hybridization and DNA sequencing. These analyses demonstrate that the
dUTPase isoforms are encoded by the same gene with isoform-specific
transcripts arising through the use of alternative 5
exons. This
finding represents the first example in humans of alternative 5
exon
usage to generate differentially expressed nuclear and mitochondrial
specific protein isoforms.
dUTPase (EC 3.6.1.23) catalyzes the hydrolysis of dUTP to dUMP and pyrophosphate, simultaneously removing dUTP from the DNA biosynthetic pathway and providing substrate (dUMP) for the de novo synthesis of thymidylate (1). The importance of the dUTPase function in prokaryotic, eukaryotic, and viral systems has been firmly established in recent years. dUTPase is essential for viability in Escherichia coli and Saccharomyces cerevisiae and was shown to be required for efficient DNA replication in several viral systems (2-5). The importance of the dUTPase function in the human system has been further distinguished by its importance in anti-thymidylate chemotherapy.
Thymidylate metabolism has been an important target for the development of widely utilized chemotherapeutic agents such as 5-fluorouracil, fluorodeoxyuridine, and ZD1694 that are used in the treatment of breast and gastrointestinal tumors. The enzymatic target of these compounds, thymidylate synthase (TS),1 generates dTMP from dUMP and a folate cofactor. Investigations by several laboratories suggest that the basis of cytotoxicity caused by inhibition of de novo thymidylate metabolism may be the accumulation of excessive dUTP pools (6-9). TS inhibition induces a dramatic elevation of dUTP pools resulting in chronic dUMP misincorporation into DNA during replication and repair, leading to DNA fragmentation and cell death (7). Work published by Curtin et al. (7) demonstrates a significant correlation between intracellular dUTP levels and the magnitude of DNA damage resulting from TS inhibition. dUTPase (the major regulator of dUTP pools in humans) plays a protective role by limiting dUTP accumulation in the cell and countering the cytotoxic effect of TS inhibition. Recently, direct evidence supporting this role was demonstrated. Overexpression of the E. coli dUTPase in HT29 human colorectal tumor cells resulted in the induction of resistance to the TS inhibitor fluorodeoxyuridine (10). These studies provide substantial evidence suggesting that levels of the dUTPase enzyme may be a critical factor in determining fluorodeoxyuridine toxicity in certain cancer types. The essential nature of dUTPase during DNA replication and its role in anti-thymidylate chemotherapy has led many authors to speculate on the use of human dUTPase as a target for future chemotherapeutic design (6, 11, 12). Indeed, the recent determination of the human dUTPase crystal structure represents a clear step toward the development of structure-based dUTPase inhibitors (13).
In recent years, there have been major advances in our understanding of the protein biochemistry of human dUTPase. Several laboratories have described the cloning of dUTPase cDNAs (11, 12, 14), and the crystal structure of a recombinant form of human dUTPase has recently been solved (13). The resulting structural data demonstrate that human dUTPase exists as a homotrimer with active site residues contributed by adjacent subunits (13). Our laboratory has recently discovered that distinct mitochondrial (DUT-M) and nuclear (DUT-N) isoforms of this enzyme exist in human cells (11). The isoforms are separable by SDS-PAGE, with DUT-N having an apparent molecular mass of 22 kDa and the processed form of DUT-M having an apparent molecular mass of 23 kDa. To determine the structural differences between these isoforms, a combination of NH2-terminal protein sequencing and mass spectrometry was utilized to characterize each form in detail, demonstrating that the two isoforms are largely identical, differing only in a short region of their amino termini. Analysis of dUTPase phosphorylation demonstrated that DUT-N is serine-phosphorylated at a consensus cyclin-dependent protein kinase phosphorylation site (Ser11 of DUT-N), suggesting a link to cell cycle regulation (15). While DUT-M contains the identical site, it does not undergo phosphorylation in HeLa cells. Site-directed mutagenesis analysis demonstrates that serine phosphorylation of DUT-N does not regulate enzymatic activity, and the biological significance of DUT-N phosphorylation remains unclear (15). Kinetic analysis of each isoform demonstrates that they retain identical affinities for dUTP despite their structural differences.
Our current knowledge of dUTPase expression in higher eukaryotes is limited. Studies of dUTPase in plants suggest that expression is linked to the rate of proliferation, as demonstrated by a close association of dUTPase activity and replicative vegetative and floral meristem regions (16, 17). Recently, the dUTPase gene was isolated from the human fungal pathogen, Candida albicans, the expression of which is governed in part by an MluI cell cycle box element located in the promoter region (18). This element was shown to be critical for the proliferation-dependent transcriptional regulation of the dUTPase gene. In the human system, studies utilizing peripheral blood lymphocytes suggest that dUTPase expression is also associated with proliferative status (14). However, this previous study is unclear with respect to the expression of distinct nuclear and mitochondrial isoforms. The present investigation was undertaken to extend our previous work characterizing human dUTPase isoforms and elucidate the mechanisms governing isoform expression.
34Lu human lung fibroblasts (CRL 1491) were obtained from the American Type Culture Collection and maintained in Eagle's minimal essential medium supplemented with 10% fetal calf serum purchased from Life Technologies, Inc. Cells were synchronized by incubation in serum-free minimal essential medium for 72 h. Cells were then released from G0 by the addition of minimal essential medium containing 10% fetal bovine serum. Progression through the cell cycle was monitored in serum-stimulated and -starved cells by [3H]thymidine incorporation. At 3- or 6-h intervals after serum stimulation/starvation, cells were labeled for 20 min with 10 µl of [3H]thymidine (1 mCi/ml). The remainder of the procedures were performed as described previously (19).
Isolation and Sequencing of DUT-M cDNA ClonesA gt10
cDNA library was constructed from mRNA derived from
serum-starved 34Lu human lung fibroblasts. Total RNA was isolated using
the reagents and protocols of TRIZOL reagent (Life Technologies, Inc.).
Poly(A)+ RNA was purified with the Poly(A)Ttract mRNA
isolation system (Promega). cDNA construction was performed using
the Superscript Choice system for cDNA synthesis (Life
Technologies). The library was screened using a 935-bp
SphI/EcoRI fragment of the DUT-N cDNA cloned
by this laboratory (11). This DNA fragment contains sequence common to
both the DUT-N and DUT-M isoforms. The library screening protocol was
based on standard procedures as described by Sambrook et al.
(20). Several cDNA isolates were subcloned into the
EcoRI site of pGEM-3Z (Promega). These clones were sequenced
using the Sequenase dideoxy chain termination kit (U.S. Biochemical
Corp.) according to the manufacturer's recommendations. A series of
subclones were generated using internal restriction sites to facilitate sequencing of the cDNA isolates in their entirety. The sequences were determined from both strands.
An EMBL-3 human
leukocyte genomic library (CLONTECH) was screened
with the DUT-N cDNA (11). The library screening protocol was based
on standard procedures as described by Sambrook et al. (20).
Several clones were isolated, and the three largest (9-15 kb) were
further characterized by a combination of restriction digestion,
Southern hybridization, and sequence analysis. An oligonucleotide probe
(5-GGTGTCTCTTCAGAGCAGG-3
) corresponding to the extreme 5
end of the
DUT-N cDNA was synthesized and used to identify a 2.7-kb
EcoRI/BamHI genomic fragment. This fragment was
subsequently sequenced using a series of nested sequencing primers
designed from newly obtained sequence information.
Coupled in vitro transcription and translation was performed using the TNT system purchased from Promega. The open reading frames of both DUT-M and DUT-N were cloned into pGEM-3Z in the T7 and SP6 orientations, respectively. The reaction mixture (50 µl) contained 25 µl of rabbit reticulocyte lysate; 2 µl of TNT reaction buffer; 20 µM concentration of each amino acid minus methionine; 45 µCi of [35S]methionine (Amersham, SJ235, 15 mCi/ml, >1000 Ci/mmol); 2 units of either SP6 or T7 RNA polymerase; and 2 µg of plasmid (pGEM-3Z) containing either DUT-N or DUT-M. Reactions were incubated at 30 °C for 90 min.
Cellular Fractionation: Isolation of Functionally Intact Mitochondria from HeLa Cells for Mitochondrial Import StudiesMitochondria were isolated by the procedure of Rickwood et al. (21) using differential centrifugation after cells were disrupted by homogenization. For complete details of this procedure, see Caradonna et al. (23). The purified mitochondria were washed twice in MESH buffer (20 mM Hepes-NaOH, pH 7.4; 220 mM mannitol; 70 mM sucrose; and 0.1 mM EDTA) and used directly for import studies.
In Vitro Mitochondrial Import of DUT-MMitochondrial import of in vitro expressed proteins was performed as described previously (22). Briefly, 25 µl of radiolabeled translation reaction was mixed with 25 µl of the prepared mitochondria (at 5 mg/ml total protein). Reactions were allowed to proceed for the indicated times at 30 °C and then chilled on ice. The sample was either disrupted by sonication followed by dUTPase protein immunoprecipitation or further processed as described next. The mitochondria were pelleted by centrifugation and washed twice in 500 µl of MESH buffer. The mitochondria were subsequently resuspended in 20 µl of MESH buffer and incubated with 20 µg/ml trypsin (L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated and affinity-purified; Promega) for 30 min on ice. The mitochondria were washed twice more with 500 µl each of MESH buffer and sonicated, and dUTPase protein was immunoprecipitated. dUTPase protein was subsequently fractionated by 15% PAGE, and the gel was dried and exposed to x-ray film to visualize the radiolabeled protein.
Antibodies, Immunoprecipitation, SDS-PAGE, Two-dimensional PAGE, Western Blot Analysis, and Enzyme AssaysdUTPase-specific monoclonal antibodies (mAb 415) were generated and prepared as described previously (23). For immunoprecipitation, 5 µl of mAb (1 mg/ml) was used. Immunoprecipitation was performed as described previously (23). dUTPase-specific polyclonal antibodies were raised against recombinant DUT-N protein (expressed in the baculovirus system) as described previously (11). The polyclonal antibodies were purified by dUTPase protein affinity chromatography and used for immunoblot analysis at a dilution of 1:1000. Enzyme assays were performed as described previously (10). Two-dimensional PAGE was performed using a Mini-PROTEAN II cell obtained from Bio-Rad. Procedures were performed according to the manufacturer's recommendations and as described in Strahler et al. (14). Western blot analysis was performed as described previously (11).
Northern Hybridization AnalysisTotal RNA was isolated from 34Lu cells using the Trizol Reagent method (Life Technologies), and poly(A)+ RNA was isolated with the Poly(A)Ttract mRNA isolation system (Promega) using the protocols supplied by the manufacturer. In each case, 1 µg of mRNA was electrophoresed in a 1.5% agarose formaldehyde gel and transferred to nitrocellulose. Blots were probed with 32P-labeled random-primed cDNA probe (SphI/EcoRI DUT-N fragment or GAPDH cDNA) or with a DUT-M-specific probe (290-bp AluI/EcoRI fragment). Blots probed with random-primed cDNAs were hybridized at 42 °C overnight in 50% formamide, 5 × SSPE (1 × SSPE: 0.18 M NaCl, 10 mM NaH2PO4, and 1 mM EDTA (pH 7.7)), 5 × Denhardt's solution, 1% SDS, 100 µg/ml denatured salmon sperm DNA. The filters were washed twice at room temperature in 2 × SSC, 0.1% SDS (5 min); twice in 0.2 × SSC, 0.1% SDS at 50 °C (15 min); and twice in 0.2 × SSC at 65 °C (15 min). The blots were subsequently exposed to Kodak X-Omat film. Blots were stripped by boiling for 10 min in 0.5% SDS prior to reprobing.
Previous studies by this laboratory have
demonstrated that DUT-N and DUT-M proteins are nearly identical,
differing only in a short region of their amino termini (11). Based on
this evidence, we speculated that their respective mRNAs would also
be identical in their overlapping regions. We therefore utilized the 3
end of the DUT-N cDNA (SphI/EcoRI, 935-bp
restriction fragment) to screen for a cDNA isolate corresponding to
the DUT-M isoform. This restriction fragment removes the 5
end of the
DUT-N cDNA coding for DUT-N-specific amino acids, leaving the 3
end containing sequence shared between the isoforms. A cDNA library
was constructed from mRNA derived from 34Lu human lung fibroblasts
that were serum-starved. Quiescent 34Lu cells were shown to express
only the mitochondrial isoform, thus providing a greater likelihood of
identifying a DUT-M-specific cDNA (described hereafter).
Screening the 34Lu library resulted in the identification of 10 positive clones, four of which were chosen for plaque purification and
subsequent characterization. The isolates were subcloned into the
EcoRI site of pGEM-3Z, and subsequent sequence analysis
indicated that all four isolates encode the DUT-M isoform based on
comparison of the deduced amino acid sequence with the experimentally
determined protein sequence of DUT-M (11). The nucleotide sequence and open reading frame of the largest DUT-M cDNA (960 bp) is presented in Fig. 1. The DUT-M open reading frame (starting at
nucleotide 63) corresponds to a 252-amino acid polypeptide with a
predicted molecular weight of 26,704. The first 342 nucleotides of the
DUT-M cDNA are unique. Nucleotides 343-960 are identical to
residues 46-664 of the DUT-N cDNA previously reported by this lab
(Fig. 2A). The deduced amino acid sequence
indicates that the DUT-M protein possesses an arginine-rich, 69-amino
acid leader sequence, which is removed in the mature protein. Amino
acid residues 1-93 are unique to DUT-M, while residues 94-252 are
common to both nuclear and mitochondrial forms (Fig. 2B).
The cDNA does not contain a consensus polyadenylation signal and
appears to be incomplete at the 3 end as evidenced by a lack of a
poly(A) tail. During cDNA library construction, random hexamers
were used to prime during first strand cDNA synthesis instead of
oligo-dT; this may account for the lack of a complete 3
end.
Isoelectric Points of DUT-N and DUT-M
Current models of
mitochondrial protein import suggest that a positively charged
targeting presequence is required for protein import into mitochondria
(24). In agreement with this model, an arginine-rich, 69-residue
presequence is predicted by the DUT-M cDNA. The 69-residue
presequence has a calculated pI of 12.9, while the full-length DUT-M
sequence has a calculated pI of 9.97. Reports of homologous nuclear
encoded mitochondrial and cytosolic/nuclear isoproteins confirm that
mitochondrial proteins have a more basic overall charge than their
cytosolic/nuclear counterparts (25). DUT-N has a predicted pI of 6.16, while the processed form of DUT-M has a predicted pI of 8.23. To
experimentally determine the pI values of the dUTPase isoforms,
two-dimensional SDS-PAGE was employed. Total HeLa cell extract and
extract from purified mitochondria were fractionated separately by
two-dimensional PAGE and transferred to nitrocellulose, and dUTPase
isoforms were detected by Western blotting (Fig. 3,
A and B). Detection of dUTPase from total cell
extract (Fig. 3A) reveals two proteins with isoelectric points of 6.0 and 8.1. These pI values correlate with the predicted values for DUT-N and DUT-M, respectively. Fractionation of protein derived from purified mitochondria confirms that the protein with the
pI of 8.1 is in fact DUT-M (Fig. 3B).
In Vitro Characterization of DUT-M
To confirm the identity of
the DUT-M cDNA, a plasmid construct containing the DUT-M open
reading frame was subjected to in vitro transcription
and translation in the presence of [35S]methionine.
dUTPase protein was immunoprecipitated from the lysate with
dUTPase-specific monoclonal antibody, fractionated by SDS-PAGE, and
exposed to x-ray film. Alternatively, the immunoprecipitated protein
was assayed for dUTPase activity. The DUT-N open reading frame was used
as a positive control in these experiments (Fig. 4,
lane 1). As demonstrated in Fig. 4, lane 2,
in vitro transcription/translation of the DUT-M cDNA
results in the production of a precursor protein with an apparent
molecular mass of 31 kDa that immunoreacts with dUTPase
monoclonal antibody. A second minor protein species is also evident of
slightly lower molecular weight. We believe this is an artifact of the
in vitro transcription/translation system, possibly due to
the use of a second in-frame methionine codon at position 129 of the
DUT-M nucleotide sequence. In vitro dUTPase assays
demonstrate that the DUT-M precursor protein retains dUTPase activity
despite an extended amino terminus contributed by the mitochondrial
targeting presequence (data not shown).
In Vitro Mitochondrial Processing of DUT-M
To further
establish the mitochondrial localization of the DUT-M protein, in
vitro import experiments were performed to demonstrate that the
DUT-M precursor is taken up by mitochondria and proteolytically cleaved
into a lower molecular weight, mature form. DUT-M was transcribed and
translated in the presence of [35S]methionine in
vitro and incubated for various times with mitochondria isolated
from HeLa cells. DUT-N was used as a negative control in these
experiments. The results of these experiments are seen in Fig.
5A. Lane 1 shows DUT-M protein
without treatment, and lanes 2-5 show DUT-M protein that
was incubated with purified mitochondria for 10, 20, 30, and 40 min,
respectively. These data clearly demonstrate that, upon incubation with
mitochondria, the DUT-M protein is proteolytically cleaved into a
23-kDa form. This processed molecular weight is in agreement with
observations of the mature DUT-M protein in vivo (11).
Lane 6 was left blank. Lane 7 shows DUT-N protein
untreated. Lane 8 shows DUT-N protein that was incubated
with mitochondria for 60 min; no proteolytic processing is evident.
To demonstrate that the DUT-M processing event correlates with import into the mitochondria, radiolabeled DUT-M protein was incubated with purified mitochondria for various times, and mitochondria were subsequently harvested by centrifugation. The mitochondria were washed and then treated with trypsin to degrade protein that was not imported into the organelle. The mitochondria were washed and subsequently fractionated by SDS-PAGE followed by autoradiography. DUT-N protein was used as a negative control. The results of these experiments are shown in Fig. 5B. Lane 1 represents DUT-M without treatment. Lanes 2-5 represent DUT-M protein incubated with mitochondria for 10, 20, 30, and 40 min, respectively, followed by treatment with trypsin. Lane 6 was left blank. Lane 7 shows DUT-N protein that was left untreated. Lane 8 represents DUT-N protein that was incubated with mitochondria for 60 min and subsequently treated with trypsin. These results indicate that the processed form of DUT-M (23 kDa) becomes resistant to tryptic proteolysis, indicating that import into the mitochondria results in protection from digestion. The residual DUT-M precursor protein remaining outside of the mitochondria and the treated DUT-N protein were subsequently degraded by treatment with trypsin.
Differential Expression of Nuclear and Mitochondrial Isoforms of dUTPase in Serum-stimulated and Serum-starved Human FibroblastsIt has been shown previously that members of the
de novo thymidylate biosynthetic pathway, including TS,
dihydrofolate reductase, and thymidine kinase, are induced in response
to serum stimulation of quiescent human fibroblasts (Ref. 26 and
references therein). To investigate dUTPase isoform expression in a
similar manner, we decided to use 34Lu human lung fibroblasts as a
model cell culture system. 34Lu cells were serum-starved for 72 h
and subsequently serum-stimulated by the addition of 10% fetal bovine
serum to the media. DNA replication status was monitored by
[3H]thymidine incorporation at the indicated time points.
Cells were harvested at time points correlating with G0,
G1, and S phase (as monitored by
[3H]thymidine incorporation), and equivalent amounts of
total cellular protein were fractionated by 15% SDS-PAGE. The protein
was transferred to nitrocellulose, and dUTPase isoforms were detected
by immunoblot analysis. Time course experiments presented in Fig.
6, A and B, illustrate dUTPase
expression in response to serum stimulation. The Western blot presented
in Fig. 6A illustrates dUTPase expression status in cells
0 h post-stimulation (lane 1), in cells 12 h
post-stimulation (lane 2), and in cells 30 h
post-stimulation (lane 3). It is clear from this data that
the lower molecular weight DUT-N isoform is induced in response to
serum stimulation, correlating with the onset of DNA replication (see
Fig. 6B). In contrast, the higher molecular weight DUT-M
isoform is expressed in a constitutive manner, independent of
proliferation status.
A similar time course experiment was performed to establish dUTPase isoform expression as a consequence of serum starvation (Fig. 6, C and D). Cycling cells were serum-starved, and dUTPase expression was monitored by Western blot. In Fig. 6C, lane 1 illustrates dUTPase expression at 0 h, lane 2 shows expression at 24 h post-starvation, and lane 3 shows expression at 48 h post-starvation. These data indicate that DUT-N is preferentially degraded within 48 h post-starvation, a rate that corresponds with the cessation of nuclear DNA replication (Fig. 6D). DUT-M levels, however, remain constant as the cells transition into the G0 state.
To determine if a similar regulatory phenomenon was occurring at the
mRNA level, Northern blot analysis was performed on mRNA derived from serum-stimulated and serum-starved 34Lu cells. Cells were
either serum-starved for 72 h or serum-stimulated for 36 h,
and poly(A)+ RNA was isolated. The RNA (1 µg) was
fractionated by agarose gel electrophoresis and then transferred to
nitrocellulose. The resulting Northern blots were probed with the 3
end (935-bp SphI/EcoRI fragment) of the DUT-N
cDNA, a region shared between both isoforms (Fig.
7A). To ensure equal loading of mRNA
samples, a GAPDH cDNA was used to probe the identical blot (Fig.
7B). Fig. 7A, lane 1, contains dUTPase
mRNA isolated from serum-starved cells, and lane 2 contains dUTPase mRNA isolated from serum-stimulated cells. In
serum-starved cells (lane 1), a single 1.4-kb dUTPase
transcript is evident. Upon serum-stimulation, however, a 1.1-kb
message is induced, while the levels of the 1.4-kb transcript, observed in quiescent cells, remain the same (lane 2). This pattern
of dUTPase mRNA expression mimics dUTPase protein isoform
expression, with the inducible 1.1-kb transcript corresponding to DUT-N
and the constitutive 1.4-kb transcript corresponding to DUT-M. To clearly demonstrate the identity of the transcripts, an
isoform-specific probe was utilized for Northern blot analysis. The
extreme 5
end of the DUT-M cDNA sequence is unique, providing for
an isoform-specific probe. This 5
DUT-M-specific sequence (250-bp
EcoRI/AluI fragment) was used to probe 34Lu
mRNA isolated from cycling cells. The common probe (935-bp
SphI/EcoRI fragment) was used as a control on an identical blot (Fig. 7C). As previously shown, the common
probe hybridizes to both the 1.1- and 1.4-kb dUTPase transcripts (Fig. 7C, lane 1). The DUT-M-specific probe, however,
specifically hybridizes to the 1.4-kb transcript, indicating that the
1.4-kb transcript encodes DUT-M (Fig. 7C, lane
2). Together, these data indicate that the DUT-N and DUT-M
isoforms are transcribed as distinct mRNAs of 1.1 and 1.4 kb,
respectively, in 34Lu cells. In addition, the 1.1-kb transcript is
induced as a consequence of serum stimulation, while the 1.4-kb DUT-M
transcript is maintained at constant levels regardless of cell growth
status.
Genomic Organization of the DUT-N and DUT-M Isoforms
Because
the cDNA and amino acid sequences of the dUTPase isoforms are
identical in their overlapping regions, we set out to determine if they
were encoded by the same gene. An EMBL-3 human leukocyte genomic
library was screened with the DUT-N cDNA to isolate the dUTPase
gene. Three genomic isolates were subcloned and characterized by
restriction digestion and Southern hybridization analysis. The three
genomic isolates contained identical restriction fragment patterns in
their overlapping regions, suggesting that the three isolates contained
identical nucleotide sequence (data not shown). An oligonucleotide was
designed from the 5 region of the DUT-N cDNA and was used as a
probe to identify a restriction fragment containing the extreme 5
region of the dUTPase gene (see "Experimental Procedures" for
details). A 2.7-kb BamHI/EcoRI fragment from the
largest genomic clone (15 kb) was isolated and subcloned into pGEM-3Z.
Sequence analysis of a region of the dUTPase gene spanning 1.6 kb is
presented in Fig. 8 and illustrated in schematic form
with respect to the DUT-N and DUT-M cDNA sequences in Fig.
9. The sequence data demonstrate that the dUTPase gene does in fact encode both nuclear and mitochondrial isoforms.
Interestingly, the isoforms appear to arise by the use of alternative
5
exons, which then join the common sequence. The 5
exon of DUT-M is
spliced to the common sequence, while the 5
exon of DUT-N is
contiguous with the common sequence. Exon 1 of DUT-M correlates with
positions 492-833 of the genomic isolate (Fig. 9). There is a
DUT-M-specific intron between positions 834 and 1302, with a splicing
event occurring between exon 1 of DUT-M and the first exon shared
between the isoforms. Consensus donor and acceptor splice sites are
located at positions 833 and 1303, respectively. In contrast, the DUT-N isoform is encoded beginning at position 1259 and is continuous with
the first common exon (see Fig. 9). After the first common exon there
is an intron corresponding to position 1441 of the genomic isolate. A
consensus donor splice site is located at position 1441. An exon
containing a portion of the common carboxyl terminus shared by both
DUT-N and DUT-N lies approximately 15 kb further downstream (data not
shown).
We have set out to extend the characterization of dUTPase isoforms
and begin to elucidate basic molecular mechanisms governing dUTPase
isoform expression in human cells. In this report, we describe the
isolation and characterization of a cDNA encoding DUT-M, elucidate
isoform-specific expression patterns in replicating and serum-starved
human fibroblasts, and describe the isolation and characterization of
the 5 region of the human dUTPase gene that encodes both nuclear and
mitochondrial isoforms.
Mass spectrometry and NH2-terminal protein sequencing demonstrate that the nuclear and mitochondrial isoforms of dUTPase are identical except for a short region of their respective amino termini (11, 15). Their common COOH-terminal regions contain five highly conserved motifs, characteristic of dUTPase proteins from various species, that contribute to active site formation (13, 27). Here, we describe the cloning and characterization of a cDNA encoding the DUT-M isoform. The DUT-M cDNA encodes a 252-amino acid precursor protein with a calculated molecular weight of 26,704 that migrates with an apparent molecular weight of 31,000 as judged by SDS-PAGE. Similar to many other mitochondrially targeted proteins, the DUT-M precursor contains a positively charged mitochondrial targeting presequence (24). In vitro mitochondrial import studies demonstrate that the 31-kDa precursor protein is proteolytically cleaved into a mature 23-kDa form and transported into mitochondria. Reports comparing mitochondrial and cytosolic isoproteins suggest that the mitochondrial counterparts generally have a more basic net charge (25). In agreement with this observation, the mature DUT-M protein has a measured pI of 8.1, while the nuclear isoform has a pI of 6.0. A report by Strahler and co-workers suggests that distinct phosphorylated and unphosphorylated forms of dUTPase exist in stimulated peripheral blood lymphocytes and are separable by two-dimensional PAGE (14). The pI values reported by these authors were 6.0 for the phosphorylated form and 8.06 for the unphosphorylated form. These observations are consistent with the pI measurements of DUT-N and DUT-M in this report and are in agreement with the phosphorylation status of DUT-N (phosphorylated) and DUT-M (unphosphorylated) determined in previous reports (15). In addition, the phosphorylated form of dUTPase is induced upon mitogenic stimulation of peripheral blood lymphocytes, similar to the induction of DUT-N observed in serum-stimulated 34Lu cells (14). Together, these observations suggest that the nonphosphorylated form of dUTPase described by Strahler et al. (14) is in fact DUT-M (pI 8.1), while the phosphorylated form is DUT-N (pI 6.0).
Expression of DUT-N and DUT-M IsoformsdUTPase is the major regulator of dUTP pools in human cells, and recent evidence suggests that relative levels of dUTPase protein may significantly influence the efficacy of chemotherapeutics that target thymidylate synthase (7, 10). In light of these observations, we have set out to establish the basic mechanisms of dUTPase isoform expression in humans.
Western blot analysis of dUTPase in serum-starved and serum-stimulated human fibroblasts suggests that DUT-N protein expression is tightly linked to nuclear DNA replication status, being induced during the G0 to S phase transition. During exit from the cell cycle into G0, DUT-N protein is degraded. In contrast, DUT-M protein is expressed in a constitutive manner with no regard for cell cycle status, mimicking the pattern of mitochondrial DNA replication.
Northern blot analysis utilizing common and isoform-specific probes demonstrates that DUT-N and DUT-M are encoded by distinct transcripts of 1.1 and 1.4 kb, respectively. Northern analysis of cycling, serum-starved, and serum-stimulated cells demonstrate that DUT-M-specific mRNA is expressed in a constitutive fashion, while DUT-N-specific mRNA expression is growth-dependent.
Analysis of DUT-N expression suggests that it belongs to a class of
S-phase-specific genes that are expressed in a
growth-dependent fashion such as members of the nucleotide
precursor biosynthesis pathway (TS, dihydrofolate reductase, and
thymidine kinase) in addition to members of the DNA replication pathway
(DNA polymerase and proliferating cell nuclear antigen) (26, 28).
Expression of S-phase genes is controlled at the transcriptional as
well as post-transcriptional levels, and regulatory mechanisms appear to be complex and multileveled (26, 29-31). However, certain common
features are emerging from both structural and functional characterization of these genes. For instance, it appears that the E2F
transcription factor is responsible for a certain percentage of
transcriptional activation as a consequence of growth stimulation. The
above mentioned genes all retain E2F consensus binding sites in their
promoter regions, and each gene is induced, to varying degrees, in
response to overexpression of E2F1 in a cell culture model system (26).
The growth-dependent increase of both DUT-N protein and
mRNA levels suggests that DUT-N may also represent a target gene
for E2F-mediated S-phase induction. It is reasonable to expect that,
like many other S-phase genes, growth-dependent expression of DUT-N
will involve several levels of regulation.
In contrast to the growth-dependent expression of DUT-N, the DUT-M isoform is expressed in a constitutive manner at both the mRNA and protein levels. This pattern of expression is consistent with current views of mitochondrial DNA replication. Studies reviewed by Clayton (32) demonstrate that mitochondrial DNA synthesis occurs randomly through the cell cycle, with no regard to nuclear DNA replication status.
The regulation of dUTPase isoforms in 34Lu cells suggests that both isoforms are largely expressed to coincide with the replication pattern of their respective genomes. However, certain developmental conditions break this rule. Strahler and co-workers demonstrated that high levels of the phosphorylated form of dUTPase (presumably DUT-N) are expressed in nonproliferating immature thymocytes (14). These authors speculate that the dUTPase function may be required during T cell development. These authors also note that proliferating cell nuclear antigen, another growth-regulated gene, is also expressed in the immature thymocytes. This may indicate that many S-phase genes are expressed in these nonreplicating cells through a common mechanism. Our data clearly show that DUT-N and DUT-M are regulated by dramatically different mechanisms; this is of particular interest because the isoforms are encoded by the same gene.
As illustrated in Figs. 8 and 9, dUTPase isoforms appear to arise
through the alternative use of 5 exons. An interesting question is
whether the isoform-specific transcripts are regulated through single
or multiple promoters. The genomic organization and the differential
regulation of DUT-N and DUT-M, at both the protein and RNA levels,
suggests that there may be distinct promoters driving differential
transcription. Support for this speculation is found by examination of
the genomic sequence directly upstream of the DUT-N isoform (Fig. 8).
Comparison of this upstream region with known transcription factor
consensus sequences reveals several interesting traits. Consistent with
the growth-dependent control of DUT-N protein and mRNA
expression, several consensus elements that typify
growth-dependent promoters are evident, including E2F, Sp1,
and NF-
B (see Fig. 8). Many S-phase genes contain E2F and Sp1
elements in their promoters that play central roles in their cell cycle
(growth)-dependent expression (28, 36, 37). The
c-myc gene contains all of these putative elements (E2F,
Sp1, and NF-
B) that regulate the transcriptional activation of
c-myc under a variety of physiologic conditions (38, 39).
Recently, the C. albicans dUTPase gene was reported to
contain an MluI cell cycle box element in its promoter that
is critical for cell cycle-dependent transcriptional
regulation (18). This cis-element is analogous to the human E2F
transcription factor in that it confers S-phase-dependent transcriptional activation in many yeast genes associated with DNA
replication (40). All of these examples provide support for the
involvement of these putative elements in DUT-N transcriptional control; however, confirmation awaits formal promoter analysis of the
human dUTPase gene.
A recent report by Cohen and co-workers
assigned the human dUTPase gene to a single locus on chromosome 15 by
fluorescence in situ hybridization (41). Consistent with
this observation, we have discovered that both isoforms are encoded by
the same gene. By comparing the known protein and cDNA sequences of
the dUTPase isoforms with the 5 region of the dUTPase gene, it is evident that both DUT-N and DUT-M are encoded by a single gene with
each isoform arising through the use of alternative 5
exons (see Fig.
9). As illustrated in Fig. 9, the first exon of DUT-M begins 767 base
pairs upstream of the first exon of DUT-N. Exon 1 of DUT-M splices
directly to the first common exon (splice donor site, 833; splice
acceptor site, 1303). In contrast, the first exon of DUT-N is
continuous with the first common exon shared between the isoforms. The
genomic organization of DUT-N and DUT-M is of particular interest
because it represents the first example in humans of a single gene
encoding differentially expressed nuclear and mitochondrial isoproteins
that arise by the alternative use of 5
exons to generate discrete
transcripts. Studies comparing homologous nuclear encoded mitochondrial
and cytosolic isoproteins suggest that only rarely are compartmentally
distinct isoforms encoded by the same gene (25). The citric acid cycle
enzyme fumarase is an important example of a gene that encodes both
cytosolic and mitochondrial isoforms in a variety of different
organisms including human, rat, and yeast. Interestingly, these
isoforms arise through several distinct regulatory mechanisms in
different organisms. In humans, for example, fumarase isoforms are
translated as a single gene product without a targeting presequence;
however, the cytosolic form is acetylated (33). In contrast, rat
fumarase isoforms arise by alternative start codon usage within the
same transcript, the mitochondrial form encoding a cleavable targeting peptide at the NH2 termini (34). Yeast fumarase isoforms
were recently shown to be transcribed and translated as a single
protein, containing a signal sequence that is cleaved in the
mitochondria (35). The authors speculate that the cytosolic form arises
from the release of an import-incompetent form of fumarase from the organelle into the cytoplasm. These examples demonstrate that partitioning of isoproteins in different cellular compartments that are
encoded by the same gene can occur through one or more transcriptional,
translational, or post-translational mechanisms. Human dUTPase
represents a novel example for generating mitochondrial and nuclear
isoforms that are differentially regulated.
In summary, we have confirmed the existence of a mitochondrial specific
isoform of dUTPase by cloning and characterizing a DUT-M-specific
cDNA. By analyzing quiescent and serum-stimulated 34Lu human lung
fibroblasts, we have show that the isoforms are differentially
expressed at both the protein and mRNA levels to coincide with the
DNA replication patterns of their respective target compartments.
Further, we show that DUT-N and DUT-M are encoded by a single gene,
arising through the use of alternate 5 exons. Future studies will
focus on the transcriptional mechanisms contributing to
isoform-specific expression.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U90223 (DUT-M cDNA) and U90224 (human dUTPase gene).
We thank Michael J. Hansbury for review of this manuscript.