(Received for publication, October 2, 1996)
From the UNIGEN Center for Molecular Biology, The Medical
Faculty, Norwegian University of Science and Technology, N-7005
Trondheim, Norway,
Johns Hopkins Oncology Center, Radiobiology
Laboratory, Baltimore, Maryland 21287-5001, the § Department of
Physics, Norwegian University of Science and Technology, N-7005
Trondheim, Norway,
INSERM U.332, 75014 Paris, France, and ** Drexel
University, Department of Bioscience and Biotechnology, Philadelphia,
Pennsylvania 19104
Uracil-DNA glycosylase releases free uracil from DNA and initiates base excision repair for removal of this potentially mutagenic DNA lesion. Using the yeast two-hybrid system, human uracil-DNA glycosylase encoded by the UNG gene (UNG) was found to interact with the C-terminal part of the 34-kDa subunit of replication protein A (RPA2). No interaction with RPA4 (a homolog of RPA2), RPA1, or RPA3 was observed. A sandwich enzyme-linked immunosorbent assay with trimeric RPA and the two-hybrid system both demonstrated that the interaction depends on a region in UNG localized between amino acids 28 and 79 in the open reading frame. In this part of UNG a 23-amino acid sequence has a significant homology to the RPA2-binding region of XPA, a protein involved in damage recognition in nucleotide excision repair. Trimeric RPA did not enhance the activity of UNG in vitro on single- or double-stranded DNA. A part of the N-terminal region of UNG corresponding in size to the complete presequence was efficiently removed by proteinase K, leaving the proteinase K-resistant compact catalytic domain intact and fully active. These results indicate that the N-terminal part constitutes a separate structural domain required for RPA binding and suggest a possible function for RPA in base excision repair.
Uracil-DNA glycosylase (UDG)1 is the
first enzyme in base excision repair for removal of uracil from DNA and
its main function is probably to remove mutagenic uracil residues
resulting from deamination of cytosine in DNA (1). The subsequent steps
in the base excision repair pathway include, as the minimal enzymatic requirement in vitro, an apurinic/apyrimidinic endonuclease,
a deoxyribophosphodiesterase activity (which may be contributed by DNA
polymerase ), DNA polymerase
, and a DNA ligase (2). In analogy
to the complexity of the nucleotide excision repair pathway, base
excision repair is likely to be more complex in vivo. This
is in fact supported by the finding of an alternative, short patch
pathway, requiring proliferating cell nuclear antigen and DNA
polymerase
(3, 4). A catalytically fully active form of human UDG
has been expressed in Escherichia coli (5) and
structure-function relationships determined by site-directed mutagenesis and x-ray crystallography (6). These studies identified this form of human UDG as a one domain structure with a positively charged DNA-binding groove. UDGs are relatively small monomeric enzymes
that, at least in vitro, do not require cofactors. However, UDG is preferentially associated with replicating SV40 minichromosomes, indicating a possible interaction with components of the replication machinery (7). The gene encoding the major human UDG, UNG, is transcribed predominantly late in the G1-phase,
resulting in a 2-3-fold increase in UDG activity early in the S-phase
(8). The cell cycle regulation is consistent with the presence of
several putative regulatory elements detected in the UNG
gene (9), including a putative element for binding of replication
protein A (RPA) (10) reported previously in DNA repair genes in yeast (11). RPA is a trimeric protein required for initiation of DNA replication (12, 13), in the initial steps of nucleotide excision repair in physical interaction with XPA (14), as well as in recombination repair (15). RPA interacts with XPA both via the p70
subunit (RPA1) and the p34 subunit (RPA2) (16). RPA from non-mammalian
species substitutes poorly for human RPA during initiation of SV40 DNA
replication (17, 18), and C-terminally deleted human RPA2 is only
marginally active (19). RPA2 is phosphorylated within the replication
initiation complex early in the S-phase (20) or following DNA damage
caused by ultraviolet light (21) or ionizing radiation (22), but the
role of phosphorylated RPA2 remains unclear. A human homolog of RPA2,
called RPA4, has been identified (23), but its function is not known.
The tumor suppressor p53 physically interacts with and inhibits the
function of RPA, and this interaction may be important for regulating
the onset of the S-phase (24).
In the present work, we demonstrate an interaction between human UDG from the UNG gene (UNG) and RPA using the two-hybrid system in yeast, enzyme-linked immunosorbent assay (ELISA), and a UDG activity assay. The results show that RPA2, but neither the homolog RPA4 nor RPA1, binds to UNG. This interaction is dependent on the N-terminal presequence in UNG. The presequence, which is not necessary for the catalytic activity of UNG, contains a region of 23 amino acids with strong homology to the conserved RPA2-binding region of XPA. These results indicate a possible function for RPA in base excision repair.
Yeast reporter strain HF7c, used in the
two-hybrid system, contains two Gal4-inducible reporter genes,
HIS3 and LacZ (MATCHMAKER Two-hybrid system,
Clontech Laboratories Inc.). Plasmid vectors, pGBT9 and pGADGH,
encoding the Gal4 DNA-binding domain, Gal4-BD, (Gal4 residues 1-147)
and Gal4-activating domain, Gal4-AD (Gal4 residues 768-881),
respectively, were used to express hybrid proteins. pGBT9 and pGADGH
also contain the yeast Trp1 and the Leu2 genes, respectively, as selectable markers. To screen for protein interactions in the two-hybrid system, we used pGBTUNG28, constructed by
insertion of UNG
28-cDNA, which lacks the 28 N-terminal amino acids of the open reading frame, into
EcoRI/SalI-digested pGBT9, and a Jurkat cell
cDNA library, constructed in fusion with Gal4-AD in pGAD1318. The
yeast reporter strain HF7c was sequentially cotransformed with the
UNG
28 hybrid expression plasmid and the Jurkat cell cDNA library
(100 µg) according to Bartel and Fields (25). Interactions with the
UNG
28 hybrid protein were assayed as described by Clontech Laboratories Inc. Positive clones were further tested for specificity by retransformation into HF7c either with pGBTUNG
28 or with
extraneous targets as yeast pGBTSNF1 or pGBT9. The
SNF1-Gal4-BD/SNF4-Gal4-AD interaction in the two-hybrid system was used
as positive control (26), and plasmid vectors without insert were used
as negative control. The cDNA inserts from positive clones were
sequenced using primer walking and TaqPRISMTM Ready Reaction DyeDeoxyTM
terminator cycle sequencing kit on an Applied Biosystems model 373A DNA
sequencing system (Applied Biosystems). The human RPA2 hybrid with the
Gal4-AD in pGADGH (isolated from a HeLa cDNA library, Clontech
Laboratories Inc.) was kindly supplied by K. Tanaka (27). In order to
identify amino acids involved in the interaction between UNG and human RPA2 (p34), pGBTUNG
75 and pGBTUNG
84 were constructed the same way
as pGBTUNG
28. Sequencing of the vectors were performed in order to
ensure in frame reading. Yeast cells were cotransformed with pGAD-RPA2
and pGBT-UNG
28 or pGBT-UNG
75, and necessary controls were
performed. The filter assay was performed according to Clontech Laboratories Inc. pACTRPA1 was constructed by cloning the RPA1 NcoI-SalI fragment from pRPA70 into the
NcoI-XhoI sites of pACT2 (28). pACTRPA2 was
constructed by cloning the RPA2 XhoI-EcoRI fragment from pJGRPA2 into the XhoI-EcoRI sites
of pACT2. pACTRPA4 was constructed by cloning a PCR-generated fragment
encoding the open reading frame of RPA4 into the BamHI site
of pACT11. UNG
28 was fused to the Gal4 DNA-binding domain in the pAS
vector and called pASIUNG
28. The
-galactosidase assay was
performed as described by Harper et al. (28).
A 1-kilobase pair genomic
fragment from the uracil-DNA glycosylase gene was amplified by PCR in
the presence of [3H]dUTP (15 Ci/mmol) (Amersham Corp.,
Little Chalfont, Buckinghamshire, United Kingdom) instead of dTTP. PCR
products were purified by gel filtration followed by adsorption to
glass milk and subsequent elution according to the Geneclean II
protocol (BIO 101, Inc., Vista, CA) in order to remove excess primers
and free nucleotides. [3H]Uracil-containing DNA (60,000 dpm/µl and ~ 1 ng of DNA/µl) was stored in 1 × TE
buffer (10 mM Tris-HCl, pH 7.5, 1 mM EDTA) at 80 °C.
Heterologous
expression and purification of UNG84 (lacking 84 amino acids of the
N-terminal part) have been described previously (5). UNG
28 (lacking
the N-terminal 28 amino acids in the presequence) was expressed using a
Baculovirus system2 and purified to
homogeneity as described for UNG
84 (5). The UDG activity of UNG
28
is severalfold lower than that of UNG
84 on the substrate used in the
present experiments.2 Purified proteins were stored in 50 mM Hepes, pH 8.0, 1 mM dithiothreitol, 0.125 M NaCl, and 50% glycerol at
20 °C.
To test the effect of trimeric RPA
on UDG activity, UNG28 or UNG
84 (10 ng) and recombinant trimeric
RPA (200 ng), kindly provided by Dr. M. S. Wold (29), were incubated
for 30 min on ice in 200 µl of 2 mM NaCl, 40 mM Tris acetate, pH 7.0, 1.6 mM MgCl2, 0.2 mg/ml bovine serum albumin, and 0.2 mM dithiothreitol. Then aliquots of the preincubation
mixture were diluted in the same buffer to equal enzyme activities,
mixed with double-stranded [3H]uracil-containing DNA
prepared by PCR, and UDG activity measured as described (30).
To investigate whether RPA bound to single-stranded DNA could recruit
UNG to DNA, 100 ng of RPA diluted in assay buffer was preincubated on
ice for 30 min with rate-limiting amounts of heat-denatured [3H]uracil-containing single-stranded DNA prepared by
nick translation of calf thymus DNA (30). This was done by diluting the
assay volume to 200 µl such that the final concentration of
[3H]uracil (in DNA) was approximately 0.27 µM, which is severalfold below Km for
both forms of UNG. Then 40 pg of UNG84 or 3.5 ng of UNG
28 were
added, and the assay was mixture incubated at 37 °C for time periods
specified in the results. UDG activity was determined as described in
Ref. 30. To investigate whether RPA in solution could recruit UNG to
DNA, UNG proteins and RPA were preincubated and then substrate was
added under otherwise similar conditions.
Wells of immunoplates
(Nunc Maxisorb, Roskilde, Denmark) were coated with 1 µg of RPA in
100 µl of phosphate-buffered saline (PBS), pH 7.4, at 4 °C
overnight. All subsequent incubations were done at room temperature.
The wells were blocked by incubation with 1 mg of bovine serum albumin
in 200 µl of PBS for 1 h and washed with PBS (three times with
200 µl/well). Then varying concentrations of UNG28 or UNG
84 in
100 µl of PBS with 2% fetal calf serum were added to the wells, and
the plates incubated for 1 h. After washing as above, protein
A-purified rabbit antibodies raised against UNG
84 (5) were added (15 ng/well). The anti-UNG antibodies (PU101) detect UNG
28 and UNG
84
equally well in an ELISA system when they are coated directly onto
immunoplates.2 After further incubation for 1 h and
subsequent washing with PBS, horseradish peroxidase-conjugated swine
anti-rabbit IgG (Dacopatts A/S, Glostrup, Denmark) was added and
binding quantitated using a colorimetric method
(o-phenylenediamine, Dacopatts A/S) according to the
manufacturer's instructions.
UNG proteins were
treated with proteinase K, separated on sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), blotted, and
visualized as follows: 3.2 pmol of UNG28 (calculated molecular mass
30.9 kDa, SDS-PAGE molecular mass 33 kDA) or UNG
84 (calculated
molecular mass 25.5 kDa, SDS-PAGE molecular mass 27 kDa) diluted in UDG
buffer (10 mM NaCl, 20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM dithiothreitol, and 0.5 mg/ml
bovine serum albumin) were incubated with 400 ng of proteinase K
(Sigma) in a total volume of 100 µl for 0 and 160 min at room temperature. At 160 min 100 µl of UDG buffer or UDG
buffer containing 1 mM phenylmethylsulfonyl fluoride
(Sigma) were added to the tubes which were immediately placed on ice. To test whether RPA protected UNG
28 against protease digestion, 3.2 pmol of UNG
28 diluted in UDG buffer were preincubated on ice with 8.6 pmol (1 µg) RPA in a total volume of 100 µl for 30 min. Then the sample was split in two equal parts, incubated with 400 ng of proteinase K or buffer for 160 min at room temperature, and
digested UNG proteins analyzed by Western blotting using PU101 antibodies for detection (5).
Using the hybrid protein Gal4-BD-UNG28, a human Jurkat
cDNA library fused to Gal4-AD in pGAD1318 (gift from J. Camonis,
U248, INSERM, Paris) was screened in the yeast two-hybrid system with a
transformation efficiency of 2.5 × 106
transformants/100 µg of library DNA used. From a 5.6-fold
amplification, 62 positive clones were identified. All these clones
indicated an interaction between UNG
28 and a protein closely related
to RPA2 as determined by nucleotide sequencing. Initially we assumed that this was a new homolog of RPA2. However, this potential RPA2 homolog hybridized strongly to genomic DNA from rat and not to human
DNA. Furthermore, the nucleotide sequence of this cDNA clone was
identical to a partial cDNA clone for rat RPA2 given the accession number H32647[GenBank] in GenBankTM. Therefore this new RPA2 cDNA clone most
probably results from contamination of the human cDNA library with
cDNA from rat and has consequently been reported to GenBankTM
(accession number X98490[GenBank]) as a putative rat homolog of human RPA2
showing 89% identity with the human protein. Using this information,
interaction between the human hybrid proteins Gal4-BD-UNG
28 and
Gal4-AD-RPA2 (authentic human RPA2) was demonstrated with both
His3 and LacZ reporter genes (Fig.
1, panel A). No interaction was detected
between the Gal4-AD-RPA2 and Gal4-BD-UNG
75 (Fig. 1, panel
B) or Gal4-BD-UNG
84 hybrid proteins (data not shown), indicating that the binding of RPA2 to UNG is dependent on the presequence of UNG. This interaction between RPA2 and UNG is specific, since the RPA2 hybrid did not interact with other irrelevant Gal4-BD hybrid proteins such as Gal4-BD-SNF1 or the unfused Gal4BD (Fig. 1A). In addition to the controls specified in Fig. 1,
panels A and B, we tested for interaction between
Gal4-BD-UNG
28 and Gal4-AD or Gal4-AD-SNF4, but no such unspecific
binding was observed (data not shown).
Although no other interactions were observed after screening with the
cDNA library, we investigated the possible interaction of UNG28
with RPA1 (known to interact with XPA) and RPA4 using a modified
version of the two-hybrid system (28). Human RPA1, RPA4, and RPA2 (as a
positive control) were each fused to the Gal4-AD in the pACT vector,
and UNG
28 was fused to the Gal4-BD in the pAS vector. As shown in
Fig. 1C, UNG
28 interacted only with RPA2 and not with
RPA1 or RPA4. We did not directly test for possible interaction with
RPA3, but such an interaction may not be likely since it could not be
demonstrated when human UNG was screened against the Jurkat T-cell
cDNA library.
The longest
and shortest RPA2-cDNA clones obtained from the screening, started
at positions corresponding to amino acids 4 and 136 in human RPA2,
indicating that the C-terminal region of RPA2 from amino acid residue
136-271 is sufficient for binding to UNG. UNG probably interacts with
an amino acid region of RPA2 that is divergent in RPA4. Although RPA4
has a 47% overall identity with RPA2, a region located between
positions 184 and 210 in RPA2 is distinctly different in the
corresponding region in RPA4. Alignment of human RPA2 with RPA2 from
other species shows that this region appears to be conserved in
mammals, but not in lower eukaryotes or the human RPA2 homolog RPA4
(Fig. 2).
Binding of Trimeric RPA to UNG in Vitro and Effect on UDG Activity of UNG Proteins
To examine whether physiologically active
trimeric form of RPA interacted with UNG in vitro, ELISA
assays using UNG-Ab and purified recombinant trimeric RPA together with
purified UNG28 or UNG
84 were performed. Data shown in Table
I support the UNG
28-RPA interaction, while the
severalfold lower absorbance detected with UNG
84 may represent
unspecific interaction close to the detection limit or very weak
specific interactions. This experiment demonstrated that the RPA2-UNG
interaction is direct and does not require a yeast intermediate
protein. Furthermore, UNG binds to RPA2 in its heterotrimeric
physiological context.
|
We also tested the effect of trimeric RPA on UDG activity of UNG
proteins on double-stranded DNA substrate. Since RPA binds strongly to
single-stranded DNA, a UDG substrate without detectable gaps or
single-stranded regions (generated by PCR) was used for this purpose.
Using this substrate in the presence of excess trimeric RPA, UDG
activity of UNG28 was inhibited by approximately 25%, while the UDG
activity of UNG
84 remained unaffected (Table II). These in vitro assays support the view that the interaction
between the two proteins is specific and not an artifact observed with the hybrid protein in the two-hybrid system. Furthermore, the lack of
effect of RPA on UNG
84, which does not contain the presequence, is
consistent with the results of the two-hybrid experiments.
|
In order to investigate whether the strong DNA binding capacity of RPA
could mediate recruitment of UNG to single-stranded DNA, UDG assays
were performed under conditions where the concentration of
single-stranded DNA substrate was limiting. Neither preincubation of
trimeric RPA with substrate nor preincubation of trimeric RPA with
UNG28 or UNG
84 resulted in consistently increased UDG activity. Rather, a weak inhibition of UDG activity was mostly observed under the
varying conditions used; thus we could not demonstrate significant
recruitment of UNG by RPA in these in vitro experiments (Table III).
|
The region of XPA required for interaction with
RPA2 is located within the N-terminal 58 amino acid residues (16).
Interestingly, within the 49 amino acids of the presequence of UNG,
which interacts with RPA2, a 23-amino acid region with strong homology
to the RPA2-binding region of XPA is found. The
hydrophobicity/hydrophilicity profiles are also very similar. When
allowing for two gaps of 2 amino acids, 11 out of 23 amino acids are
identical (47.8% identity), and in addition 3 are conserved (60.9%
similarity), (Fig. 3, panel A). XPA from
several other species also show homology to this region of human UNG
with chicken showing the highest homology (60.9% identity and 78.3%
similarity). SV40 T-antigen, which also binds to the C-terminal region
of RPA2 (19), has a 26-amino acid region with limited homology to human
UNG. In addition, a protein in the cyclin family (UNG2) apparently
having UDG activity also has a region of homology to UNG and XPA (Fig.
3B). The presequence of mouse
UDG3 is also strongly homologous to human
UDG and contains the putative RPA2-binding motif. The corresponding
region is also conserved in UDG from the fish Xiphophorus,
although more distantly related,4 but is
absent in UDG from yeast or animal viruses (various herpes simplex
viruses and pox viruses) (data not shown). Although limited to a small
region, this is, to our knowledge, the first homology demonstrated
between damage recognizing proteins in the nucleotide excision repair
and base excision repair pathways.
In conclusion, these results support the biochemical data and suggest that residues 57-79 in human UNG and 20-46 in XPA interact with RPA2.
The RPA2-binding Presequence in UNG Probably Constitutes a Separate Structural DomainThe crystal structure of UNG84, a
catalytically fully active enzyme, has revealed a globular one-domain
protein with the N-terminal region localized on the surface. Using
proteinase K digestion, the presequence is easily removed from
UNG
28, resulting in a processed form slightly longer than UNG
84
(Fig. 4, lanes 2 and 3). No
further digestion of UNG
28 is observed. Purified UNG
84 is
proteinase K-resistant, probably due to its compact structure
(lanes 2 and 4). In an attempt to examine whether
binding of RPA to UNG
28 affected the proteinase K digestion,
UNG
28 was preincubated with RPA and then incubated with proteinase
K. Lanes 2 and 6 show that RPA did not protect
UNG
28 from protease digestion. Proteinase K treatment of UNG
28
removed the presequence without any loss of activity, in fact the
activity is increased (data not shown), whereas the catalytic activity
of UNG
84 was not affected by the proteinase K treatment. The shorter
form of UNG resulting from proteinase K treatment could not have
resulted from digestion at the C-terminal end, because deletion of only
the last 4 amino acids at the C-terminal end by site-directed
mutagenesis where Trp-301 was converted to a stop codon resulted in
complete loss of activity (data not shown). These results, together
with the results obtained in the two-hybrid studies, indicate that the presequence constitutes a separate structural domain required for RPA
binding, but not for catalytic activity.
In this report we demonstrate an interaction between human UNG and
human RPA in vivo, using the yeast two-hybrid system, and in vitro by an ELISA assay using purified trimeric RPA and
purified forms of human UDG (UNG28 or UNG
84). Only the 34-kDa
subunit of RPA2 interacts directly with UNG, and the interaction
results in a weak inhibitory effect on UDG activity rather than
stimulation. The weak inhibitory effect may be an in vitro
effect, and we do not suggest that RPA inhibits UDG activity in the
much more complex in vivo situation; this result merely
strengthens the other results showing that the presequence is involved
in an interaction between UNG and RPA, since a possible functional
consequence on UDG activity should be confirmed inside cells.
Previously, RPA has been shown to be involved in the damage recognition
step of nucleotide excision repair (14), in recombination repair (15),
and in replication (13), but a possible involvement in base excision
repair has not been reported, except for a possible role of RPA as a
transcription factor in the regulation of expression of DNA repair
genes (11). A putative element for binding of RPA is located in an
inhibitory region of the UNG promoter (9, 10). In nucleotide
excision repair, XPA interacts with both RPA2 and RPA1, but whereas
RPA1 is essential for cell survival after UV light exposure, RPA2 has a
more modest although significant effect (16). A region between amino
acids 28 and 78 in UNG is necessary for interaction with RPA2, but we
have not formally shown that this region is sufficient for interaction.
Interestingly, amino acids 57-79 in UNG show strong homology to amino
acids 20-46 in human XPA. Since XPA residues 1-59 are necessary and
1-74 sufficient for binding to RPA2 (16), it seems likely that the
region of homology between amino acids 57-79 in human UNG and amino
acids 20-46 in XPA are the actual RPA2-binding regions. At the
C-terminal end of the region of homology, the amino acid sequence RLAAR
is present in both UNG and XPA. However, this sequence does not appear
to be sufficient for RPA2 binding, since Gal4-BD-UNG
75 that starts
with this sequence did not interact in the two-hybrid system. Our
results indicate that the interacting region in human RPA2 is located
downstream of the first 136 amino acid residues. A conserved region in
RPA2 from mammals, lacking in RPA4 and RPA2 from lower eukaryotes, is
present between residues 184 and 210. This may be a candidate region
for binding of UNG, and if so, the pattern of conservation would
indicate that only RPA from higher eukaryotes will recognize UNG,
analogous to the ability of RPA from mammals, and inability of RPA from
lower eukaryotes like Saccharomyces cerevisiae to bind to
T-antigen and to stimulate SV40 replication (17, 18).
One possible function of RPA2 could be to recruit UNG to a DNA repair
complex possibly associated with the replication complex for scanning
for uracil prior to replication. A related function has been proposed
for SV40 T-antigen which, like dnaC, P protein, and the T4 phage
gp59 protein, facilitate primase recognition of RPA-coated DNA (Ref. 18
and references therein). Alternatively, RPA2 might assist transport of
UNG either to the nucleus or within the nucleus. Among the three
subunits, only RPA2 is associated with the nuclear matrix and this
association persists throughout the cell cycle (31). In addition, RPA2
and RPA1 undergo a transition from uniform nuclear distribution to the
punctuate nuclear pattern typical for replication foci during the
S-phase (32). Nuclear staining of UNG also displays a nonhomogeneous
distribution reminiscent of replication foci (33), although it has not
been shown that UNG and replication foci colocalize. One of the
functions of RPA2 could be to contribute to the localization of UNG in
replication foci. UDG activity is not essential for DNA replication in
E. coli or S. cerevisiae (34, 35). In contrast,
it is essential for poxvirus replication (36) and for the replication
of a herpes virus (HSV1) in mammalian cells that do not express
cellular UDG (37). In human cells, UNG is most actively
transcribed very late in the G1-phase and UDG activity
subsequently increases 2-3-fold early in the S-phase (8). In addition,
an association of UDG activity with replicating SV40 minichromosomes
(7) may indicate a possible function for UNG in replication or in an
associated repair complex.
The protease digestion in the present study indicates that the first approximately 80 amino acids, comprising the presequence of UNG, constitute a second structural domain separated from the previously reported structure of the catalytic domain by at least one protease-accessible site (6). We have reported previously that the nuclear and mitochondrial form of human UDG have a size corresponding to the "mature" form lacking the presequence of 77 amino acids, while a larger form that reacted with an antibody specific for the presequence was present in the cytosol fraction. In addition, transfection experiments indicated that the presequence was essential for mitochondrial import, but not for nuclear import (38). However, there are some inconsistencies in the literature regarding the size of nuclear UDG, since some reports indicate a size consistent with processing of the N-teminal end (39, 40), whereas others report a size essentially corresponding to the larger "preform" (41, 42). In conclusion, we do not yet know whether processing at the N-terminal end is a physiological process or merely an in vitro artifact during purification. Proteolytic cleavage of UDG during purification has been reported in yeast (43), and a similar digestion of human nuclear UNG may take place. However, the interaction of the "presequence" of UNG with RPA, a truly pleiotropic nuclear protein, necessitates further studies on the biochemistry of nuclear uracil-DNA glycosylase and possible accessory proteins assisting in base excision repair.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) X98490[GenBank] (rat RPA2).
We thank Dr. R. B. Walter and H. Nilsen for communicating unpublished results to us, Dr. K. Tanaka for supplying pGADGH, Dr. M. S. Wold for supplying purified trimeric human RPA, Dr. G. Slupphaug for supplying UNG-antibodies, and Dr. J. Camonis for supplying the Jurkat cDNA library.