Nuclear and mitochondrial uracil-DNA glycosylases are generated by alternative splicing and transcription
from different positions in the
UNG
gene
Nuclear and mitochondrial uracil-DNA glycosylases are generated by alternative splicing and transcription from different positions in the UNG gene
Hilde
Nilsen
,
Marit
Otterlei
,
Terje
Haug
,
Kristin
Solum
,
Toril A.
Nagelhus
,
Frank
Skorpen
and
Hans E.
Krokan*
UNIGEN Center for Molecular Biology, The Medical Faculty, Norwegian University
of Science and Technology, N-7005
Trondheim
,
Norway
Received November 6, 1996;
Revised and Accepted December 23, 1996
DDBJ/EMBL/GenBank accession nos
+
ABSTRACT
A distinct nuclear form of human uracil-DNA glycosylase [UNG2, open reading frame (ORF) 313 amino acid residues]
from the
UNG
gene has been identified. UNG2 differs from the previously known form (UNG1,
ORF 304 amino acid residues) in the 44 amino acids of the N-terminal sequence, which is not necessary for catalytic activity. The rest
of the sequence and the catalytic domain, altogether 269 amino acids, are
identical. The alternative N-terminal sequence in UNG2 arises by splicing of a previously unrecognized
exon (exon 1A) into a consensus splice site after codon 35 in exon 1B
(previously designated exon 1). The UNG1 sequence starts at codon 1 in exon 1B
and thus has 35 amino acids not present in UNG2. Coupled
transcription/translation in rabbit reticulocyte lysates demonstrated that both
proteins are catalytically active. Similar forms of UNG1 and UNG2 are expressed
in mouse which has an identical organization of the homologous gene. Constructs
that express fusion products of UNG1 or UNG2 and green fluorescent protein
(EGFP) were used to study the significance of the N-terminal sequences in UNG1 and UNG2 for subcellular targeting. After
transient transfection of HeLa cells, the pUNG1-EGFP-N1 product colocalizes with mitochondria, whereas the pUNG2-EGFP-N1 product is targeted exclusively to nuclei.
INTRODUCTION
Uracil-DNA glycosylase is the first enzyme in base excision repair (BER) 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
). A catalytically fully active recombinant form of human uracil-DNA glycosylase lacking an N-terminal sequence encoded by the open reading frame (ORF) (
2
) has been used to study structure-function relationships as determined by site directed mutagenesis and X-ray crystallography (
3
). These studies identified this form of human uracil-DNA glycosylase as a one domain structure with a positively charged DNA-binding groove and a deep uracil-binding pocket (
3
,
4
) which is accessed by flipping of uracil with its deoxyribose and 5'-phosphate out of the helix (
5
). The complete sequence of the
UNG
gene, which encodes human uracil-DNA glycosylase (UNG)
(
6
), and comparison with
UNG
cDNA from human placenta (
7
) indicated that the gene contained six exons and a TATA-less promoter. This cDNA contains an ORF of 304 amino acids (
7
), but a homogeneous form of uracil-DNA glycosylase purified from human placenta (
8
) lacked 77 N-terminal amino acids predicted from the ORF (
7
). Subsequent work demonstrated that the
UNG
gene encoded both the mitochondrial and nuclear forms of UNG-proteins and that this N-terminal sequence was required for transport of the enzyme to
mitochondria, but not for nuclear import as determined by transfection studies
(
9
). The similarity of the mitochondrial and nuclear forms was also demonstrated
by subcellular fractionation and Western blotting, as well as by immunostaining
(
9
,
10
). It was thus assumed that proteolytic removal of the N-terminal sequence could explain the differential transport of UNG either
to mitochondria or nuclei, although artificial cleavage during purification
could not be excluded.
Here we report a distinct form of human nuclear UNG. We designate this form UNG2
and the previously known human form (
7
) UNG1. Both forms are encoded by the
UNG
gene, but have different N-terminal sequences. In addition we have isolated cDNAs for mouse homologs
of UNG1 and UNG2. Transient transfection experiments with constructs that
directed expression of fusion products of either UNG1 or UNG2 and a variant of
green fluorescent protein indicated that the N-terminal sequences of UNG1 and UNG2 direct targeting to mitochondria and
nuclei, respectively.
MATERIALS AND METHODS
Materials
Mouse embryonic carcinoma cDNA library, human liver cDNA library and NT2
neuronal precursor cell cDNA library were from Stratagene (La Jolla, CA, USA).
All libraries were propagated in the Uni-ZAPTXR vector using XL-1 blue as host. [[alpha]-
32
P]dCTP, [
35
S]methionine, Rediprime random labelling kit and Hybond N+ filters were all from
Amersham (UK). All sequencing primers were from MedProbe (Oslo, Norway). Dye
terminator cycle sequencing ready reaction kit was from Applied Biosystems
(Foster City, CA). The Dynazyme PCR kit was purchased from Finnzymes Oy (Espoo,
Finland). TNT
in vitro
transcription/translation rabbit reticulocyte lysate system kit, pGEM-T TA cloning kit, Alter Sites II
in vitro
Mutagenesis System, primers for sequencing from T3 and T7 promoters and T3 RNA
polymerase were from Promega (Madison, WI). The plasmid encoding the red-shifted variant of green fluorescent protein (pEGFP-N1) was from Clontech (Palo Alto, CA, USA). Restriction enzymes were
from New England Biolabs Inc. (Beverly, MA, USA).
Screening of cDNA libraries
All libraries were screened as recommended by the manufacturer, using
32
P-labeled
UNG40
cDNA (
7
) as probe. Hybridization was carried out at 65oC overnight in 6* SSC, 5* Denhardt's solution and 0.1% SDS. Filters were washed in 0.1* SSC/0.5% SDS at 65oC and autoradiographed. Three rounds of screening
were done.
In vivo
excision of pBluescript phagemids from the Uni-ZAPTXR vector was performed as recommended by the manufacturer.
Sequence analysis of clones
Sequencing was performed on an Applied Biosystems Model 373A DNA Sequencing
System using the Dye terminator cycle sequencing ready reaction kit as
recommended by the manufacturer. The sequences were analyzed using the Auto
Assembler software (Applied Biosystems).
In vitro
transcription and uracil-DNA glycosylase assays
In vitro
transcription/translation was performed with the TNT transcription/translation
system with [
35
S]methionine as recommended by the manufacturer, using 200 ng of the expression
constructs per 10 [mu]l reaction volume. The mouse UNG1- pBluescript construct was transcribed from the T3 promoter in the
pBluescript vector. The insert of mouse UNG2-pBluescript was amplified by the polymerase chain reaction using Dynazyme
PCR kit, ligated into the pGEM-T vector and transcribed from the T7 promoter. The human UNG2-pBluescript was transcribed from the T3 promoter after
Sac
I/
Nhe
I excision of a 79 bp fragment from the polylinker and the 5'-end of cDNA for UNG2. Human UNG1 cDNA was transcribed from the T7
promoter as previously described (
2
). The samples were run on a 12% denaturing sodium dodecyl sulfate
polyacrylamide gel (SDS-PAGE). The gel was dried, autoradiographed overnight and scanned on an
LKB Ultroscan XL Enhanced Laser Densitometer. Uracil-DNA glycosylase activity was measured in parallel samples of the
in vitro
transcription/translation assay mixture containing unlabelled amino acids (
2
).
Preparation of pUNG-EGFP-N1 fusion constructs and localization studies
UNG15
cDNA, which encodes UNG1, in pGEM7Zf+ (pUNG15) (
2
,
7
) was digested with
Bcl
I, which cuts at bp 1019 in
UNG15
cDNA, blunted with DNA polymerase I (Klenow fragment), and ligated to an
Age
I linker prepared from the oligonucleotide 5'-ACCGGTGCC-3' and its complementary copy. The religated pUNG15
containing the
Age
I linker correctly ligated into the
Bcl
I site (verified by sequencing) was digested with
Rsr
II, which cuts at bp 49 in
UNG15
cDNA (
7
), blunted as above and finally digested with
Age
I. The fragment was then ligated into pEGFP-N1 digested with
Sma
I (blunt) and
Age
I. The construct was sequenced to verify that the construct was in frame with
the ATG of the EGFP-N1 fusion protein. The TGA stop codon of pUNG15 was changed to GGA by site-directed mutagenesis performed according to the procedure provided
by the manufacturer using ssDNA prepared with R408 phage. Potential pUNG1GGA
-EGFP-N1 constructs were screened by digestion with
Bcl
I (digests only unmutated plasmids) and verified by sequencing. The correct
construct was named pUNG1-EGFP-N1. cDNA for UNG2 (this article) in pBluescript was digested with
Nhe
I, which cuts 54 bp upstream of ATG, and
Eco
NI which cleaves the cDNAs in the sequence that is shared by cDNAs for UNG1 and
UNG2 (positions 529 and 520, respectively). The resulting fragment of interest
(501 bp) was isolated and ligated to the 5155 bp fragment of
Nhe
I/
Eco
NI-digested pUNG1-EGFP-N1 to obtain pUNG2- EGFP-N1. Transient transfections of HeLa cells were
done with the CaPO4
-method (Profection, Promega) according to the manufacturer's
recommendations. Confocal microscopy (BioRad MRC-600) of HeLa cells and staining of mitochondria with mouse anti human
mitochondria antibody (MAB 1273, Chemicon) and Texas Red anti-mouse IgG (Vector) were performed as previously described (
10
). Examination of HeLa cells transfected with expression plasmids pEGFP-N1, pUNG1-EGFP-N1 or pUNG2-EGFP-N1 was carried out using an excitation wave length
of 488 nm and emission wave length >515 nm at 16 h after transfection.
RESULTS
We have screened a human NT2 neuronal precursor cell cDNA library and a mouse
embryonic carcinoma cDNA library and discovered a new form of human uracil-DNA glycosylase (human UNG2) encoded by the
UNG
gene, as well as the homologous cDNA from mouse (mouse UNG2). In addition we
have isolated the cDNA for the mouse homologue (encoding mouse UNG1) of human
UNG1 (
7
). cDNA for human UNG2 has an ORF encoding 44 N-terminal amino acids not found in human UNG1 whereas cDNA for human UNG1
has an ORF encoding 35 amino acids not found in human UNG2 (Fig.
1
). The two forms are identical in the rest of the N-terminal sequence which is not required for enzyme activity, as well as in
the catalytic domain, altogether 269 identical consecutive amino acids. cDNAs
for human UNG2 and its mouse homologue are apparently as abundant as UNG1 in
cDNA libraries from proliferating cells since among 20 cDNA clones that were
sequenced 10 were of the UNG2 type and 10 were similar to the previously known
UNG1 type. Among four mouse cDNAs sequenced, three were of the UNG2 type and
one was of the UNG1 type. However, screening of a human hepatocyte cDNA library
with
UNG40
cDNA (
7
) resulted in the isolation of 80 strongly hybridizing clones and sequencing of
14 of these demonstrated that they were all similar to the previously
characterized cDNA for UNG1 or the cDNA
UNG40
(
7
).
To examine whether human UNG1 and UNG2 were translocated to different
subcellular compartments, we prepared constructs expressing fusion proteins of
the UNG proteins and a red shifted variant of green fluorescent protein (EGFP-N1). These were used for transient transfection experiments with HeLa
cells. The major advantage of the green fluorescent protein (over the use of
antibodies) is that this method relies on the autofluorescence of this protein
alone, and thus possible cross reaction of an antibody with epitopes in
irrelevant proteins is not a problem. The control (pEGFP-N1) shows that the green fluorescent protein displays a homogeneous
staining over the cells (Fig.
4
A). In contrast, the UNG2-EGFP-N1 fusion protein is exclusively located in the nuclei (Fig.
4
C) and the UNG1-EGFP-N1 fusion protein (Fig.
4
D) is mainly, if not exclusively, located in extranuclear spots that have the
same appearance as mitochondria stained with Texas red (Fig.
4
B).
The few spots of green fluorescence visible over the nucleus in Figure
4
D are probably due to the presence of mitochondria over and under the nuclei in
this summation image of optical sections. These results provide convincing
experimental evidence that UNG2 is a nuclear protein and UNG1 a mitochondrial
protein.
