(Received for publication, October 30, 1996, and in revised form, December 19, 1996)
From the Institut für Biochemie der
Ludwig-Maximilians-Universität München, Genzentrum,
Feodor-Lynen-Strasse 25, D-81377 München, Germany, the
§ Gesellschaft für Strahlen forschung-National
Research Center for Environment and Health, Institute of Immunology,
Marchioninistrasse 25, D-81377 München, Germany, and the
¶ Max-Delbrück-Center for Molecular Medicine, Robert
Rössle Strasse 10, D-13122 Berlin-Buch, Germany
In this article we describe the molecular cloning
of Pirin, a novel highly conserved 32-kDa protein consisting of 290 amino acids. Pirin was isolated by a yeast two-hybrid screen as an
interactor of nuclear factor I/CCAAT box transcription factor
(NFI/CTF1), which is known to stimulate adenovirus DNA replication and
RNA polymerase II-driven transcription. Pirin mRNA is expressed
weakly in all human tissues tested. About 15% of all Pirin cDNAs
contain a short 34-base pair insertion in their 5-untranslated
regions, indicative of alternative splicing processes. Multiple Pirin
transcripts are expressed in skeletal muscle and heart. Western blots
and immunoprecipitations employing monoclonal anti-Pirin antibodies reveal that Pirin is a nuclear protein. Moreover, confocal
immunofluorescence experiments demonstrate a predominant localization
of Pirin within dot-like subnuclear structures. Homology searches using
the BLAST algorithm indicate the existence of Pirin homologues in mouse and rat. The N-terminal half of Pirin is significantly conserved between mammals, plants, fungi, and even prokaryotic organisms. Genomic
Southern and Western blots demonstrate the presence of Pirin genes and
their expression, respectively, in all mammalian cell lines tested. The
expression pattern, the concentrated localization in subnuclear
structures, and its interaction with NFI/CTF1 in the two-hybrid system
classify Pirin as a putative NFI/CTF1 cofactor, which might help to
gain new insights in NFI/CTF1 functions.
Nuclear factor I (NFI)1 consists of a
family of cellular sequence-specific DNA binding proteins cloned from
mammalian organisms and chicken (1-4). Sequence homologies led to the
identification of four different genes, termed as NFI-A, NFI-B, NFI/CTF
(identical to NFI-C), and NFI-X. This diversity is further increased by
alternative RNA splicing generating numerous polypeptides from a single
NFI gene (1, 5, 6). All NFI/CTF proteins share a large and highly
conserved N terminus, which forms dimers in solution and mediates
sequence-specific DNA binding (7-10). The N terminus is able to
stimulate adenovirus DNA replication by up to a factor of 50, and it
has been shown that this stimulation does not require the NFI/CTF C
terminus (9, 11, 12). NFI/CTF directly binds to the adenoviral
DNA-polymerase, which is assumed to stabilize the preinitiation complex
(13, 14).2 However, the cellular function
of the NFI/CTF proteins has not yet been completely elucidated. It has
been suggested that NFI/CTF might be involved in cellular DNA
replication. Experimental data indicate that NFI/CTF binds to DNA
polymerase in vitro (16). Until now further support for
an implication of NFI/CTF in cellular DNA replication is lacking.
A more thoroughly studied function of NFI/CTF is its potential to stimulate RNA-polymerase II-driven transcription. A number of cellular promoters contain NFI/CTF binding sites, and it has been shown that NFI/CTF can activate and repress transcription of these genes (1, 17-20). Five different polypeptides of the NFI/CTF family have been isolated, designated as NFI/CTF1, NFI/CTF2, and NFI/CTF3 (1); NFI/CTF5 (21); and NFI/CTF7 (6). Splicing does not affect their conserved N termini, and consequently all these polypetides specifically bind to the NFI/CTF consensus sequence. Recent studies revealed that these variants display a broad activation spectrum of transcriptional activation, ranging from none (NFI/CTF2) to intermediate (NFI/CTF1) and strong (NFI/CTF5 and NFI/CTF7) stimulation rates (21). These different activities may be linked to their C termini, which are highly diverse as a consequence of differential splicing. Through linker scanning mutagenesis within the transactivation domain of NFI/CTF1, a core activation domain was identified, which is essential for the activator function of NFI/CTF1 (22). However, in NFI/CTF7, alternative splicing removes most of this element, and it is completely lacking in NFI/CTF5 (21). Consequently, this suggests that NFI/CTF proteins, to activate gene expression, may also contact different components of the transcription machinery. The mechanisms by which different NFI/CTF proteins activate transcription might therefore be complex and variable. For example, it has been described that the proline-rich domain of NFI/CTF1 binds to the TATA binding protein (23). This is in contrast to other reports that demonstrate a NFI/CTF1-mediated recruitment of TFIIB but exclude binding to TATA binding protein (24). A biochemical approach using a NFI/CTF1 affinity column was performed to isolate a set of cofactors from HeLa nuclear extracts (25). These cofactors possessed an inhibitory activity for basal in vitro transcription, which was relieved by addition of the NFI/CTF1 activator. This suggests an antirepressor function of NFI/CTF1. To our knowledge, the molecular cloning of such repressing cofactors has not yet been accomplished.
Using a genetic approach we searched for cellular proteins that by interaction with NFI/CTF1 might act as potential cofactors. NFI/CTF1 was used as a "bait protein" to screen a HeLa cDNA library in the yeast two-hybrid system. Here, we describe the molecular cloning of Pirin, the most predominant NFI/CTF1 interactor of this screening. Pirin is a novel 32-kDa protein that is expressed in the same human tissues as NFI/CTF1. Pirin is highly conserved among mammals, and its N terminus shares significant homologies with novel not yet cloned proteins from plants, fungi, and even prokaryotes. Using monoclonal anti-Pirin antibodies, we show that Pirin is exclusively localized within cell nuclei, predominantly in subnuclear dot-like structures that might represent loci of specific nuclear processes, like DNA replication and/or RNA polymerase II transcription.
The following oligonucleotides were used
to perform 5 RACE RT-PCR, which is described below. Pirin-specific
primers are designated as follows: PSP1,
5
-CCGGCCCGCAGTCATCCACTGCAAATCTCCTGG-3
; PSP2, 5
-TGATCAGGAAATCCTCCTGGTCTACCTC-3
; and PSP3,
5
-ACTCGGGTCTGCCAATGCTTCTCCGGAC-3
.
A HeLa cDNA library cloned in DR2 (26)
was screened with a labeled 1057-bp Pirin cDNA probe isolated from
two-hybrid clone TH 101 by partial EcoRI/XhoI
digestion. The screening was done as described previously (27, 28).
Positive phages were plaque-purified. Two clones containing the largest
Pirin cDNA inserts but differing in their 5
-untranslated regions
(see Fig. 2) were designated pDR-Pirin 1 (GenBankTM accession number
Y07867[GenBank]) and pDR-Pirin 17 (GenBankTM accession number Y07868[GenBank]).
Isolation of HeLa RNA and 5
Approximately
2 × 107 HeLa cells grown to confluency were washed
with cold PBS, collected with a rubber policeman, and harvested (1100 rpm, 5 min, 4 °C). Cells were lysed in 2 ml of TRISOLV reagent (Biotecx) by repetitive pipetting and stored at room temperature for 5 min. Cell debris and nuclei were pelleted (14000 rpm, 15 min, 4 °C),
and supernatant containing whole cell RNA was purified by
phenol/chloroform extraction. After precipitation and washing in 70%
ethanol, RNA was air-dried briefly and redissolved in water treated
with diethylpyrocarbonate. This RNA was used for 5 RACE RT-PCR which
was performed according to the instructions of the supplier (Life
Technologies, Inc.). Briefly, 1 µg of RNA was mixed up with 2.5 pmol
of Pirin-specific primer (PSP1) and denaturated at 70 °C for 10 min.
After chilling on ice, first strand cDNA synthesis was performed
with 200 units of SuperScript II reverse transcriptase at 42 °C for
30 min. RNA was degraded with 2 units of RNase H for 10 min at
55 °C, and cDNA was purified on a GlassMax spin cartridge and
tailed for 10 min at 37 °C using 2 mM dCTP and 10 units
of terminal deoxynucleotidyl transferase. Second strand cDNA
synthesis and subsequent PCR amplification was done using 2.5 units of
Taq DNA polymerase (Stratagene), 10 pmol of anchor primer
(matching to the tail), and 10 pmol of Pirin-specific primer PSP2.
Because no PCR products were obtained after a first round of 35 PCR
cycles, 0.5 reaction volumes were reamplified in 20 PCR cycles using
anchor primer and Pirin-specific primer PSP3. PCR products were cloned
directly into plasmid pCRII (Invitrogen) using the TA cloning system
(Invitrogen). The cDNA 5
end points were determined by standard
DNA sequencing methods.
Commercially available
Northern blots (Clontech) containing approximately 2 µg of
poly(A)+ RNA/lane were hybridized with a radioactively
labeled 1.1-kb HinfI cDNA fragment isolated from clone
pDR-Pirin 1. Hybridization was carried out for 1 h at 68 °C
using ExpressHyb solution (Clontech). Blots were washed either at low
or high stringency conditions as recommended by the supplier and
exposed for up to 14 days at 80 °C using two intensifying screens.
To quantify relative amounts of Pirin transcripts, blots were also
exposed to a phosphoimaging screen (Storage phosphor screen GP, Eastman
Kodak) for 24 h and analyzed by phosphoimaging (STORM 860).
For genomic Southern blots we used a premade blot, each lane containing 4 µg of EcoRI-digested genomic DNA isolated from nine different eukaryotic species (ZOO-blot, Clontech). The blot was hybridized with a 1.3-kb Pirin cDNA insert derived from clone pDR-Pirin 1. Hybridization and washing was done according to the instructions of the supplier, and signals were detected by phosphoimaging. Digitized images were printed on a Sony videoprinter (UP-D 8800).
Expression and Purification of GST-Pirin 22-290 and His-tagged PirinA cDNA fragment encoding Pirin residues Ala22-Asn290 was isolated from clone TH 5 by partial EcoRI/XhoI digestion and cloned into plasmid pGEX4T-1 (Pharmacia Biotech Inc.), which was digested in the same way. Recombinant plasmids were sequenced and transformed into expression strain Escherichia coli JM 109. Protein expression and purification on glutathione-Sepharose 4B (Pharmacia) were carried out as described elsewhere (29).
To express full-length Pirin, clone pDR-Pirin 1 was digested with HpaI to isolate a 1.6-kb fragment covering the complete Pirin cDNA. The HpaI fragment was redigested with HinfI, blunted by Klenow polymerase, and cloned into plasmid pQE31 (Qiagen), which was digested with BamHI and blunted by Klenow fragment. Recombinant plasmid pQE-Pirin was sequenced and transformed into expression strain E. coli M15[pREP4] (Qiagen). Protein expression and purification on nickel nitrilo triacetate agarose was performed as described by the manufacturer (30).
Expression of Recombinant NFI/CTF1 in Baculo/Sf9Recombinant NFI/CTF1 protein was expressed in Spodoptera frugiperda cells (Sf9, ATCC CRL 1711) and purified to homogenity as described elsewhere.2
Production of Rat Anti-Pirin Monoclonal AntibodiesFollowing purification approximately 50 µg of the GST-Pirin 22-290 fusion protein dialyzed in PBS were emulsified with Freund's complete adjuvant and injected both intraperitoneally and subcutaneously into Lou/C rats. After 4 weeks a final boost with 50 µg of protein without adjuvant was given intraperitoneally and subcutaneously 3 days before fusion. Fusion of the myeloma cell line P3X63-Ag8.653 with the rat immune spleen cells was performed essentially as described previously (31). Hybridoma supernatants were tested in a solid phase immunoassay using the GST-Pirin 22-290 fusion protein or an irrelevant GST fusion protein (negative control) absorbed to polystyrene microtiter plates. Following incubation with culture supernatants for 1 h, bound monoclonal antibodies were detected using peroxidase-labeled goat anti-rat IgG antibodies and o-phenylenediamine as chromogen in the peroxidase reaction. Solid phase enzyme-linked immunosorbent assay on microtiter plates coated with mouse anti-rat Ig antibodies was used to determine the immunoglobulin type with rat Ig class (anti-IgM; Zymed) and IgG subclass-specific mouse monoclonal antibodies (32).
Preparation of HeLa Nuclear Extracts and ImmunoprecipitationNuclear extracts of HeLa cells were prepared as described elsewhere (33). For immunoprecipitations 15 µl of protein G-Sepharose 4 Fast Flow (Pharmacia) were mixed with 1 ml of hybridoma supernatant and incubated at room temperature for 1 h shaking continuously. Unbound antibodies were removed by washing twice with 15 ml of PBSTX (PBS + 0.1% Triton X-100). To minimize unspecific protein binding, the material was washed once with PBS supplemented with 10% fetal calf serum (FCS). Nuclear extracts were thawed on ice, and 10-50 µl (approximately 50-250 µg) were incubated under continuous shaking with antibody-Sepharose complexes at 4 °C for 1 h. Immunocomplexes were washed three times with 15 ml PBSTX, boiled 5 min at 95 °C, separated on a 12% SDS-polyacrylamide gel (34), and analyzed by Western blotting (35).
Immunostaining and Confocal Laser Scanning MicroscopyHeLa cells and mouse fibroblast 3T6 cells were grown to confluency and 1 × 107 cells were seeded onto glass slides precoated with 0.1% gelatin (Petri dishes, 140 mm in diameter). Cells were cultivated for 4 h until they attached to the glass slide surface. Cells were washed once in PBS and fixed with paraformaldehyde solution (4% (w/v) in PBS) for 30 min on ice. Samples were washed in PBS, and cells were permeabilized with digitonin solution (0.05% (w/v) in PBS) for 15 min on ice. After washing in PBST (PBS + 0.05% Tween 20), fixed and permeabilized cells were incubated for 1 h at room temperature with monoclonal anti-Pirin antibody 1E8 (200 µl of hybridoma supernatant) supplemented with 2.5% goat serum added to minimize unspecific antibody binding. Cells were washed twice in PBST and incubated with 100 µl of fluorescein isothiocyanate-labeled goat anti-rat IgG antibody (Jackson) diluted 1:80 in PBST containing 2.5% goat serum. After 1 h at room temperature samples were washed three times in PBST and mounted in Moviol (Calbiochem). Laser confocal fluorescence microscopy was performed with a Leica TCS 4D-I confocal microscope. Digitized images were printed on a Sony videoprinter (UP-D 8800).
Cell Lines and Cell CultureThe following cell lines were cultured in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% FCS (Life Technologies, Inc.): HeLa cells, a human cervix carcinoma (Flow), COS7, an African green monkey kidney cell line (DSM ACC 60, Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany) and 3T6, Swiss albino mouse fibroblasts (DSM ACC 202, Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH). MDBK, a bovine kidney cell line (DSM ACC 174, Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH), was grown in Dulbecco's modified Eagle's medium supplemented with 5% FCS. Cell line C6, a rat glial tumor (ATCC CCL 107) was cultured in Ham's F10 medium supplemented with 15% FCS, and the myelomonocytic dog cell line K1 (36) was grown in RPMI 1640 with 10% FCS.
Computer ProgramsWe used the programs FASTA, MOTIFS, BLAST, PILEUP, and MFOLD, which are part of the Wisconsin Sequence Analysis Package (37).
The yeast
two-hybrid system is widely and successfully used to analyze
protein-protein interactions in vivo (38, 39). We used the
LexA-based system developed by Brent and co-workers (38) to screen a
HeLa cDNA library for proteins interacting with
replication/transcription factor NFI/CTF1. Isolated interactors were
tested extensively in yeast to eliminate false positive binders and to
identify proteins specifically binding to NFI/CTF1. Details of this
screen will be described elsewhere. In this article, we present the
molecular cloning of Pirin, the most predominant NFI/CTF1 interactor.
To analyze whether Pirin clones isolated in the two-hybrid screen
encode the full-length protein, we first sequenced the largest
two-hybrid clone (TH 101) to determine its correct open reading frame.
Subsequently a 1.05-kb insert was isolated from TH 101 by partial
EcoRI/XhoI digestion and used as a probe to screen a HeLa cDNA library cloned in DR2 (26).
A total of 17 clones were isolated (pDR-Pirin 1 to pDR-Pirin 17) and
sequenced. The sequence of one of these clones, pDR-Pirin 1, containing
a 1277-bp cDNA insert is depicted in Fig. 1. The translational start codon of Pirin is located within the sequence 5-CCATATGGG-3
(underlined G
residues correspond to positions
3 and +4). This fits with the Kozak
consensus ATG, typically flanked by a neighboring purine at position
3 and/or a G at position +4 (40). There is no other putative start
codon within the whole 5
-untranslated region (5
-UTR) of the Pirin cDNA. An in-frame stop codon is located at position
168,
confirming the indicated translational startpoint. The 5
-UTR is very
rich in G and C residues (64, 7%). We have performed RNA secondary structure analysis using the program MFOLD (37), which predicted various extended and stable stem structures within the 5
-UTR of Pirin
(data not shown).
Pirin is a very hydrophilic protein encoded by 290 residues with a
predicted molecular mass of 32118 Da. Within the 3-untranslated region, which is 215 bp in length, the sequence element 5
-AATAAA-3
is
localized at positions 1019-1024 and 1050-1055 (Fig. 1). During 3
processing of pre-mRNA the RNA motif 5
-AAUAAA-3
is required for
proper RNA cleavage and subsequent polyadenylation (41). The spacing
between the Pirin AATAAA element at position 1050 and the poly(A) tail
is 18 bp in length, indicating that this element is most likely
recognized during RNA processing.
To confirm the isolated 5 end of clone pDR-Pirin 1, we have carried
out 5
RACE RT-PCR. Whole cell RNA was isolated from HeLa cells, and
first strand cDNA synthesis was performed using the Pirin-specific
primer PSP1. In a first round of 35 PCR cycles using anchor primer and
Pirin-specific primer PSP2, no bands could be amplified. After PCR
reamplification with anchor primer and the flanking Pirin-specific
primer PSP3, a single specific band appeared that was cloned directly
into plasmid pCRII. Comparing the sequence of pDR-Pirin 1 isolated from
the
DR2 library with isolates obtained by 5
RACE RT-PCR, we
identified an 15-bp elongated 5
cDNA end (Fig. 1).
Sequencing of Pirin cDNA clones
derived from DR2 HeLa-cDNA library and from 5
RACE RT-PCR
revealed clones that differ in their 5
-untranslated regions. These
clones, of which pDR-Pirin 17 is a representative (GenBankTM accession
number Y07868[GenBank]), carry an additional 34-bp element inserted 52 bp
upstream of the translational start codon (Fig. 2).
Approximately 15% of all Pirin cDNA isolates contained this short
additional element. We completely sequenced several clones with this
insertion. Except for the additional 34 bp, no further differences are
discernible. The 34-bp element resembles a mammalian intron because its
sequence 5
-GTGAGT ... TAG-3
fits well with the consensus intron
sequence 5
-GT(A/G)AGT ... (C/T)AG-3
(Fig. 2, B and
C). However, the intron-like motif in Pirin cDNAs is
much shorter than typical introns, and it is lacking an extended
polypyrimidine stretch required for efficient RNA splicing. It is
notable, however, that the 34-bp insertion does not contain a putative
translational start site but encodes for an in-frame stop codon. These
results indicate that both types of Pirin transcripts differ in their
5
region, but they encode for the same protein.
HeLa cells
expressing Pirin are derived from a human cervical carcinoma. To
analyze whether Pirin is also expressed in primary human tissues, we
used a 1.1-kb HinfI cDNA fragment derived from clone
pDR-Pirin 1 to probe Northern blots containing equal amounts of
poly(A)+ RNA isolated from eight different human tissues.
We found Pirin expression in all tissues tested, and we estimate the
size of the Pirin transcript between 1.4 and 1.6 kb (Fig.
3A). This corresponds well with the lengths
of isolated Pirin clones of more than 1.3 kb when they contain the
intron-like element. We detect the highest transcript levels in heart
and liver (Fig. 3A, lanes 1 and 5). All other tissues express Pirin at significantly lower levels with a
minimum in brain and pancreas (Fig. 3A, lanes 2 and 8). Interestingly, larger transcripts appear in some
tissues, and these are detected even after high stringency washing.
Heart and skeletal muscle contain a 3.0-kb transcript that can be seen
best in skeletal muscle (Fig. 3A, lane 6).
Skeletal muscle might contain three more transcripts of approximately
4.4, 7.0, and 8.5 kb. Due to their very weak expression, the indicated
sizes were calculated by a signal profile scan on a PhoshorImager (data
not shown). Northern blots using mRNA isolated from human fetal
organs indicated that in embryonic tissues Pirin is expressed in the
same relative amounts as in adult organs. Expression was very low in
brain (Fig. 3B, lane 1), low in lung (Fig.
3B, lane 2), intermediate in kidney (Fig.
3B, lane 4), and high in liver (Fig.
3B, lane 3). Pirin expression is generally quite
weak, and the blots shown in Fig. 3 were exposed for 14 days.
Pirin Is a Nuclear Protein
To clarify the cellular
localization of Pirin, we produced monoclonal anti-Pirin antibodies in
rats. Antibodies 1E8 and 5C2, which belong to the immunoglobulin
subclass IgG 2a, bound to Pirin with high affinity in a solid phase
enzyme-linked immunosorbent assay and were chosen for further
characterization in Western blot experiments. Both antibodies were
specific for Pirin because they recognized GST-Pirin 22-290 (Fig.
4, A and B, lanes 3),
full-length Pirin expressed as a histidine-tagged fusion in E. coli (Fig. 4, A and B, lanes 6),
and Pirin obtained by in vitro translation (data not shown).
Neither 1E8 nor 5C2 recognized negative controls such as E. coli crude extract (Fig. 4, A and B,
lanes 2), GST (Fig. 4, A and B,
lanes 4), or recombinant NFI/CTF1 expressed in the
Sf9/Baculo system (Fig. 4, A and B, lanes
7). To check whether or not Pirin is a nuclear protein, we
prepared nuclear extracts from HeLa cells. Pirin could be easily
detected in these extracts with both monoclonal antibodies, clearly
classifying it as a nuclear factor (Fig. 4, A and
B, lanes 5). Titration experiments revealed that
1 µg of HeLa nuclear extract contains approximately 0.5 ng of Pirin
(compare lanes 5 and 6 in Fig.
4A).
Immunoprecipitation experiments further corroborated the nuclear localization. Monoclonal antibody 1E8 specifically precipitated nuclear HeLa Pirin (Fig. 4C, lane 4), which was neither bound by protein G-Sepharose alone (Fig. 4C, lane 6) nor by a rat monoclonal control antibody of the same IgG subtype (Fig. 4C, lane 5).
Interestingly, monoclonal antibody 5C2 predominantly bound to Pirin but also recognized significant amounts of a nuclear protein with a molecular mass of approximately 58 kDa (Fig. 4B, lane 5). Further experiments are currently under way to isolate and characterize this Pirin-related protein. In this context it is important to remember that larger Pirin-related transcripts were expressed in skeletal muscle and heart (Fig. 3A), and it is conceivable that at least one of these transcripts may encode the 58-kDa protein.
HeLa Pirin was running as a 38-kDa protein in denaturating polyacrylamide gels. Because the predicted mass of Pirin is only 32.1 kDa, this discrepancy may be due to post-translational modification. However, a comparison of the migration behaviors of HeLa Pirin (Fig. 4A, lane 5) and the recombinant His-tagged Pirin with a predicted mass of 33.9 kDa (Fig. 4A, lane 6) demonstrates that both polypeptides are running at 38 or 40 kDa, respectively, indicating that Pirin may possess some peculiar migratory properties in SDS gels.
Pirin Shares Homologies with Novel Proteins of Mammals, Plants, and ProkaryotesWe have performed extensive computer searches using
programs like FASTA or MOTIFS to identify Pirin-related proteins or to find known sequence patterns within Pirin such as nuclear localization signals. Except for a "hypothetical 26.3-kDa protein" from E. coli (GenBankTM accession number P46852[GenBank]), we failed to detect any
homologies. Systematic searches using the program BLAST, however,
revealed significant homologies between Pirin and novel proteins, most
of them described as expressed sequence tags (ESTs). Combining results
of these BLAST searches we have done a realignment using the program
PILEUP (Fig. 5, A and B). We found
the most striking homologies between human Pirin and ESTs of mouse
(W08720[GenBank]) and rat (AA012706). Because the 144 residues of the mouse EST matched to human Pirin throughout its complete length with 95.8% identity, we postulate the existence of a Pirin homologue in mouse. The
rat clone consisting of 63 residues was 93.6% identical over its
complete length to human Pirin (Fig. 5B). Interestingly, the rat EST was identified by a differential display approach using myometrium from pregnant rats. This might indicate an up-regulated expression of rat Pirin in myometrium during pregnancy.
Surprisingly we found homologies between human Pirin and data base entries from Arabidopsis thaliana (H36274), Dictyostelium discoideum (Z29535), Bacillus acidocaldarius (X62835), Streptococcus lividans (X60953), and E. coli (P46852). None of these entries described a protein with known function, but all of them share striking homologies with the N-terminal half of Pirin (Fig. 5A). Between Pirin residues Gly52 and Tyr131 we identified 29 amino acids that are highly conserved throughout all these putative proteins. Most of these key residues are localized within two clusters (Gly52-Gly70 and Gly88-Ser107), and we postulate a conserved but as yet unidentified function to be localized within Gly52 and Tyr131.
It should be mentioned that we find no putative Pirin homologue in Saccharomyces cerevisiae. The only yeast protein sharing some similiarity with Pirin is SMC1, a protein known to be involved in chromosome segregation (42). However none of the conserved key residues shown in Fig. 5 matched to SMC1, and the overall identity between Pirin and SMC1 (16.5%) does not seem to be significant.
Monoclonal Anti-Pirin Antibody 1E8 Detects Pirin Homologues in a Variety of Mammalian Cell LinesTo prove the alignment results
experimentally, we probed a DNA blot containing
EcoRI-digested genomic DNA from nine different species with
a 1.3-kb Pirin cDNA fragment derived from clone pDR-Pirin 1. Pirin-specific bands were detected in all lanes containing genomic DNA
from mammalian organisms (human, monkey, rat, mouse, dog, cow, and
rabbit), whereas chicken and yeast DNA did not generate significant
signals (Fig. 6A). To confirm these results
at the protein level, whole cell extracts from the following mammalian cell lines were probed with monoclonal antibody 1E8: HeLa (human cervix
carcinoma), COS7 (monkey kidney cells), C6 (rat glial tumor), 3T6
(mouse fibroblasts), K1 (myelomonocytic dog cell line), and MDBK
(bovine kidney cells). Monoclonal antibody 1E8 detected a single band
corresponding in size to HeLa Pirin in all extracts, indicating that
Pirin homologues are significantly conserved throughout the mammalian
world (Fig. 6B).
Pirin Is Concentrated within Subnuclear Dot-like Structures
By Western blotting and immunoprecipitations we have
shown that human Pirin is a nuclear protein. To address this question in more detail, we used anti-human Pirin antibody 1E8 in
immunofluorescence experiments. Cells from asynchronously growing
cultures of HeLa and mouse fibroblast 3T6 cells were prepared for
immunofluorescence microscopy, which was performed with a confocal
laser scanning microscope. As shown in Fig. 7, Pirin was
exclusively localized within the nucleoplasma, and it was predominantly
concentrated within dot-like subnuclear structures. These dots vary in
number and size, but typically their size was between 150 and 500 nm. A
similiar dot-like structure was observed with a monoclonal antibody raised against NFI/CTF.3 Nucleoli
apparently are devoid of Pirin. As expected, immunolocalizations within
HeLa cells showed a very similiar distribution and the same dot-like
structures (data not shown).
In this article we described the isolation, cloning, and
expression of a novel nuclear protein we designated Pirin. It was identified as the major interacting partner of NFI/CTF1 in a yeast two-hybrid screen. The complete Pirin cDNA encodes a very
hydrophilic polypeptide consisting of 290 residues. Among all isolated
cDNAs a subpopulation of approximately 15% contained a short 34-bp
extra element within the 5-UTR. This might indicate alternative
splicing in the 5
-UTR, a process known to occur in pre-mRNAs
encoding cytosolic proteins (43, 44) as well as nuclear factors such as
HeLa p54nrb (45) or the transcription factor NF-E2 (46). It is
remarkable that the 34-bp insertion in Pirin looks like a mammalian
intron. Nuclear pre-mRNAs typically contain short but strongly
conserved nucleotide sequences at both splice sites, i.e. a
5
-terminal GU and the 3
-terminal AG (Fig. 2), and these features are
observed also in the 34-bp Pirin insert. The lengths of mammalian
introns are known to vary considerably from tens of thousands of
nucleotides down to a minimum of approximately 70 nucleotides (47). It
is not only the shortness of the 34-bp Pirin insert sequence that makes
it rather difficult to understand by what mechanisms this putative
intron could be removed by splicing. The 34-bp element in Pirin
cDNA also completely lacks the characteristic polypyrimidine
stretch, which is typically located 5
upstream of the 3
intron-exon
junction and required for intron recognition by small nuclear
ribonucleoprotein particles (reviewed in Ref. 47). It is possible that
the insertion is missing some typical intron features and as a result
this intron-like element is removed inefficiently by splicing. Our
findings might also be explained by an alternative splicing mechanism
involving two different 5
splice donor sites (spaced by 34 bp), which
are joined to the same 3
splice acceptor site. These two different 5
splice donor sites might be used with different efficiences, and this
might reflect the ratios of the two cDNA types we isolated. Finally it is possible that HeLa cells transcribe two different Pirin genes
resulting in two transcripts that differ only in 34 base pairs.
Experiments are now under way to isolate and sequence genomic Pirin
clones, which will enable us to address this question.
It will be interesting to investigate whether alternative sequence
patterns in the 5-UTR of Pirin transcripts influence its mRNA
stability or result in differentiated protein translation. Initiation
of translation requires that the small ribosomal subunit, in
conjunction with initiation factors, scans the 5
-UTR until it finds an
AUG initiator codon in a favorable sequence context. It has been
demonstrated that the primary sequence of the 5
-UTR can affect the
rate of translation (reviewed in Ref. 48). Ribosome binding is most
sensitive to secondary structures near the 5
cap, but it is known that
stable stem-loop structures anywhere in the 5
-UTR can block or slow
down ribosome scanning (49). The RNA structure program MFOLD predicted
the existence of stable extended stem-loops within the Pirin 5
-UTR,
and we found significant differences in these structures depending on
whether or not the 5
-UTR contains the 34-bp insertion. It is therefore
possible that both types of Pirin transcripts encode the same protein
but differ in stability and/or are translated with different
efficiences.
Northern blotting and RT-PCR experiments indicated that the Pirin
mRNA is transcribed at low levels. We have repeatedly probed different Northern blots with highly labeled probes and even after low
stringency washing it took a minimum of 4 days of standard autoradiography to detect specific Pirin signals. Control genes like
pur (50) and others revealed significantly stronger
mRNA expression, indicating that weak Pirin signals are not simply a result of insufficient amounts of RNA on the Northern blot. Results
from Northern blots are in agreement also with 5
RACE RT-PCR
experiments. A variety of Pirin gene-specific primers failed to amplify
Pirin cDNA from HeLa whole cell RNA during a first round of PCR
cycles, whereas control genes were amplified easily with appropriate
primers in a single PCR round (data not shown).
It is remarkable that both skeletal and heart muscle contained multiple transcripts specifically hybridizing to Pirin cDNA probes even under high stringent conditions. It seems possible that these tissues express larger Pirin-related proteins, and this is in agreement with the results obtained by monoclonal anti-Pirin antibody 5C2 recognizing not only the 32-kDa Pirin but also a larger nuclear polypeptide of approximately 58 kDa. Further experiments are required to elucidate this point.
Genomic Southern blots, Western blots, immunoprecipitations, and immunofluorescence experiments using monoclonal antibodies indicated that Pirin is a highly conserved nuclear protein at least among mammals. Moreover, computer alignments demonstrated significant homologies within the N terminus of Pirin and putative uncloned proteins in plants, fungi, and prokaryotes. Many of these putative proteins have been described as ESTs only. Cloning of these sequences and comparison with the Pirin sequence will help us to clarify whether all aligned sequences found so far are genuine homologues of Pirin or encode other proteins sharing a conserved protein domain with Pirin.
Given its conserved characteristics, its dot-like nuclear localization, and its interaction with the transcription/replication factor NFI/CTF1 in the yeast two-hybrid system, it will now be very interesting to investigate the function of Pirin. We are currently setting up an NFI/CTF1-driven transcription system in mammalian cell lines to find out whether or not Pirin can modulate transcriptional activation by NFI/CTF1 in vivo. In parallel we will analyze whether Pirin is involved in other nuclear processes such as DNA replication or DNA repair, because these events are known to require proteins often conserved significantly during evolution.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) Y07867[GenBank] and YO7868.
We thank Dr. Akis Zorbas and Dr. Horst Ibelgaufts for discussions and for critically reading the manuscript.