ABSTRACT
NPM (nucleophosmin/B23) is a major nucleolar protein which is 20 times more abundant in tumor or proliferating cells than in normal resting cells. Recently, it was found that NPM gene is
located at the breakpoints of the t(2:5), t(3:5) and t(5:17) chromosome
translocation. To understand the human NPM gene's structure and regulation,
four genomic clones were isolated from the human chromosome 5 library and their DNA sequences analyzed. The human NPM gene has 12 exons of sizes ranging from 58 to 358 bp.
The chromosome breakpoint for t(2:5) and t(5:17) translocation is within intron 4 and the breakpoint for t(3:5) translocation is within intron 6. The initiation
site is located 96 bp upstream from the ATG site. A typical TATA box (at -25 nt) and a GC box (at -65 nt) were identified in the promoter region. We identified two
gel-shift bands (A and B) with DNA fragment E (-741/-250 nt) by EMSA. A DNA footprint was observed at (-371/-344 nt) with the nuclear extract. A double stranded DNA with the
footprint sequence (-371/-344 nt) competed the formation of gel-shift bands A and B in EMSA suggesting that proteins A and B
bind to the footprint region. We confirmed that protein A is transcription
factor YY1. These results suggest that YY1 may play a role in NPM gene
expression. This is the first report on human NPM gene structure and sequence.
Nucleophosmin (NPM, also called protein B23, numatrim, NO38) is a nucleolar
phosphoprotein 20 times more abundant in cancer cells than in normal resting
cells (
1
). Its putative function is the assembly and/or transport of ribosome. NPM is a
mobile protein; its cellular localization is affected by growth conditions and
influenced by certain cytotoxic drugs. During serum starvation (
2
) or treatments with anticancer drugs-daunomycin, actinomycin D, camptothecin or toyocamycin-it shifts from nucleoli to nucleoplasm (NPM-translocation) (
3
-
7
). The cellular location of NPM is also dependent on GTP and ATP levels. When
the
de novo
synthesis of GTP is inhibited and cells are depleted of GTP, NPM-translocation occurs (
8
). On the other hand, when ATP levels are depleted, NPM-translocation is blocked and newly synthesized ribosomal RNA accumulate in
nuclei and cannot be transported to the cytoplasm (
9
-
10
). NPM is associated with pre-ribosomal particles and other nuclear proteins as well. The proteins it
associates with include transcription factor YY1 (
11
), nucleolar p120 protein (
12
) and HIV protein Rev/Rex (
13
-
14
).
NPM biosynthesis is related to cell proliferation and mitogenesis. We found that NPM mRNA is 50-fold higher in Novikoff hepatoma and 5-fold higher in hypertrophic rat liver than in normal rat liver (
1
). Feuerstein
et al
. (
15
-
16
) reported that when B cells, T cells and Swiss 3T3 cells were stimulated with
various mitotic agents, NPM synthesis increased. On the other hand, down
regulation of NPM was observed in Jurkat T-lymphoblasts during apoptosis (
17
). These studies indicate that NPM expression is associated with cell growth.
The rat NPM gene has been isolated and characterized (
18
). However, the structure and sequence of the functional human NPM gene is not
yet known. Previously, we identified seven human processed pseudogenes of NPM (
19
). Genomic blot analysis indicated that there are at least 10 copies of NPM gene per haploid human genome. With the recent information that the chromosome breakpoints of t(2:5), t(5:17) and t(3:5) chromosome translocation associate with NPM gene (
19
-
25
), we searched for the functional gene in the human chromosome 5 library and
isolated four clones. Here, we report the gene's physical structure and DNA
sequence. The 5' region and the DNA sequence of the intron 4 where t(5:2) chromosome
break occurs were characterized. This is the first report on human NPM gene
structure and sequence. We also found that the transcription factor YY1 (
26
-
29
) binds to the 5' region of the NPM gene, suggesting that YY1 may play a role in NPM gene
expression.
A human chromosome 5 genomic library was obtained from the American Type Culture
Collection (cat. no. LA05NL03). The library was screened with cDNA probes (
1
) and the first intron probe (p16-3/1.2S) derived from the chromosome walk (
20
). Screening libraries, Southern blot analysis, DNA cloning, PCR and DNA sequencing by dideoxynucleotide termination reactions were performed according to the standard methods described in Sambrook
et al
. (
30
). Primers used in PCR and in cloning were derived from various regions of the
cDNA.
To identify the initiation site, the primer extension method was employed.
Antisense primer P5 (1-17 from the ATG site of cDNA) was end-labeled with [[gamma]-
32
P]ATP by polynucleotide T4 kinase. RNAs were prepared from HeLa cells using the
GIBCO/BRL Trizol reagent. Poly-A RNA were purified with oligotex-dT gel. Annealing of the primer to RNA was conducted at 43oC for 8 h. Primer extension was carried out by AMV reverse
transcriptase at 42oC for 90 min. The primer extended fragment was analyzed in an 8% sequencing
gel with both markers and a sequence ladder derived from the P5 primer.
The procedure of EMSA was according to `Current Protocol, vol II' (
31
) with slight modifications. DNA fragments, P (-249 to +465) and E (-741 to -250), were prepared by cleavages of the 5' DNA with
Mlu
I (which cleaves at -249) and
Sac
I (which cleaves at -741 and +465) restriction enzymes. The reaction mixture (25 [mu]l) containing the MCF-7 or HeLa cell nuclear (
32
) extract (1-3 [mu]g protein), end-labeled DNA fragments (0.5-5 ng), binding buffer [12 mM HEPES, pH 7.9, 4 mM Tris-HCl, 60 mM KCl, 1
mM DTT, 1 mM EDTA, 0.05% NP-40, 5 mM MgCl
2
, (MgCl
2
is required for YY1 binding), 12% glycerol], poly dI-dC (1.5 [mu]g per reaction), BSA (300 [mu]g/ml) " competitors, were incubated at 25oC for 15 min. The mixture was loaded onto a 4%
polyacrylamide gel for electrophoresis. Retardation of DNA fragments due to
binding of protein factors was detected by autoradiography.
DNA fragment E was end-labeled either by T4 kinase or Klenow reaction. The labeled DNA was
incubated with nuclear extracts (3-15 [mu]g) in a final volume of 50 [mu]l containing 12 mM HEPES, pH 7.9, 4 mM Tris-HCl, 60 mM KCl, 1 mM
DTT, 1 mM EDTA, 12% glycerol and 2 [mu]g of poly(dI-dC) and BSA (100 [mu]g/ml). After 15 min of incubation at 25oC, DNase I (0.5-1.5 unit) was added to the mixture and allowed to
digest for 1-1.5 min at 0oC. The reaction was stopped and precipitated by adding
ethanol/tRNA/ammonium acetate buffer mixture (
31
). The reaction product was analyzed by a 6% polyacrylamide-8 M urea gel. The position of the footprint was determined by an adjacent sequencing gel ladder (
33
).
Three DNA fragments from the 5' region of the NPM gene: I (-741/+93), II (-405/+93) and III (-43/+93, which is the basal promoter with a TATA box)
were prepared by PCR with appropriate primers. The 3' end of these clones is located 3 bases upstream from the ATG site. These
DNA fragments were subcloned into the luciferase reporter gene vector pXP1 (
33
) and were transfected into HeLa cells by electroporation (
34
) under the following conditions: 1.5 kV/cm, 1 ms pulse width, 2 trains with 15 pulses.
Cells were cultured in Petri dishes (35 mm) for 48 h before harvest. Under these conditions, ~20% of cells were transfected. Luciferase activity (RLU, relative luciferase unit) was determined with a Luminometer using D-lucuferin as substrate. The luciferase activity for each sample was normalized with their co-transfected [beta]-galactosidase activity (RLU/[beta]-gal unit) which was determined by the ONPG
colorimetric assay.
About 2 * 10
5
plaques from the human chromosome 5 library were screened (see Materials and
Methods). Four clones were identified and purified which have inserts of 13-20 kb. Restriction mapping and Southern blot analysis with probes derived from the 5' and 3' region of the NPM cDNA indicated that three clones (A, B
and C) contained the 5' portion of the NPM gene and one clone (D or clone 21-3) (
20
) contained the 3' region. The inserts of these clones were fragmented with restriction
enzymes and subcloned into Bluescript SK vectors. Subcloning of individual
introns was also achieved through PCR techniques using cDNA primers. The DNA
sequences of these clones were determined.
Figure
1
shows the restriction enzyme sites, the relative size and the location of exons
in clones C and D. More than 60% of the gene sequence has been determined
(sequences have been submitted to the GenBank). Since some of the intron
sequences (except introns 3, 4 and 5) were not fully determined, the relative
position between exons was determined by PCR with cDNA primers and by
restriction mapping. Clone C contains exons 1-6 and clone D contains exons 7-11. The combined sequence of the NPM gene in clones C and D span ~25 kb. Since the overlapping sequence has not been found
between the 3' of the C clone and the 5' of the D clone, the estimated total length of the NPM gene could
be >25 kb. There are 12 exons with sizes ranging from 58 (exon 7) to 358 bp
(exon 12). The DNA sequences of the exons are identical to the cDNA sequence (
1
) except that the exon 11 is not identified in the major gene product. Table
1
shows the exon/intron junctions of the human NPM gene. The same exon/intron junctions are observed in
the rat NPM gene (
18
). Figure
2
A shows the DNA sequence of the 5' region of the NPM gene and Figure
2
B shows the sequence of the intron 4 where the chromosome breaks occur (discussed
in the following).
Table 1
It was recently discovered that the 5' region of the NPM gene is fused with other genes in lymphoma cells
having t(2:5), t(5:17) or t(3:5) chromosome translocation (
20
-
25
). Based on the RT-PCR transcript sequence of the fusion protein derived from
cells with t(2:5) and t(5:17) chromosome translocation (
20
,
24
), the N-terminal 117 amino acids of NPM are fused to anaplastic lymphoma kinase
(ALK) or retinoic acid receptor alpha (RARA). In t(3:5) chromosome
translocation, the N-terminal 175 amino acids of NPM are fused to myepodysplasia leukemia
factor 1 (MLF) (
25
). Compared to the cDNA sequence (
1
) and exon/intron positions of human NPM gene (Table
1
and Fig.
1
), the 117 and 175 amino acids of NPM are located at the junctions of exon 4/exon 5 and exon 6/exon 7, respectively. Accordingly, the chromosome breakpoint
for t(2:5) and t(5:17) translocation is within intron 4 and the breakpoint for
t(3:5) translocation is within intron 6. The precise breakpoint location within
these introns is not known. Intron 4 is 911 bp long (Fig.
2
B). Sequence analysis indicated that this intron contains two
Alu
sequences (dotted underlines) which have 70-80% sequence homology to the
Alu
sequences observed in human ABL and BCR (break point cluster regions) genes (
36
). We also found that the 3' end of this intron is particularly enriched in T (denoted with italic).
To identify the initiation site for NPM transcription, the primer extension method was employed. Figure
3
shows the autoradiograph of the primer extended DNA fragments. We found one major and three minor primer
extended fragments with lengths of 111, 112, 113 (major) and 114 bp. Compared
to the adjacent sequencing gel, the major fragment (113 bp long) extended to
the cytosine (C) at 96 bp upstream from the ATG site. Accordingly, the
initiation site (+1) is assigned to this position (Fig.
2
A). A well defined TATA box (TATATAA) and a GC box (GGCG) were observed at -25 nt and -66 nt, respectively (Fig.
2
A). Sequence analysis with a computer search program identified four potential
cis
-elements upstream from the promoter region. They are
(i)
E1A-F (AGGACGT) (
37
) at -293; (ii) UCR core (CGCCATTTT) (
26
) at -352; (iii) ZRE-1 (TTACACA) (
38
) at -377; and (iv) UBP-1 (CTCTCTGG) (
39
), located at -414 and -550.
To identify proteins that bind to the 5' region of the NPM gene, an electrophoretic mobility shift assay (EMSA)
was employed. Two DNA fragments (P and E) were used (Fig.
1
). Fragment P (-249 to +465) contains the TATA box, the GC box and part of the first
intron. Fragment E (-741/-250) is located 5' from fragment P. Nuclear protein extracts of HeLa or MCF-7 cells were incubated with the DNA fragments.
Retardation of these fragments by proteins was identified (Fig.
4
). One gel-shift band (arrow) was identified using fragment P (Fig.
4
A), and two gel-shift bands (A and B) were identified using fragment E (Fig.
4
B). Formation of these gel-shift bands was not affected by competition with non-specific competitors: calf thymus DNA (which was sheared to sizes ~500 bp), poly dI-dC/poly dI-dC and BSA, but was competed out by the unlabeled
DNA fragments.
The footprint (NPM-
cis
-1) has the sequence CCATTTTG which is a potential binding site for the
zinc finger DNA binding protein, YY1 ([delta] factor, UCRBP, NF-E1) (
26
-
29
). Table
2
shows the comparison of this sequence to other sequences that associate with
protein YY1. To confirm whether YY1 is involved in the binding of DNA and
causes the gel-shift, recombinant His-YY1 (a gift from Dr E. Seto) was employed in EMSA. Purified His-YY1 (a single band in SDS gel with >95% pure) was incubated
with the end-labeled oligo NPM-
cis
-1. As shown in Figure
7
, YY1 caused the gel-shift of NPM-
cis
-1 (lane 2). Compared to the gel-shift band positions induced by the nuclear protein extract which
produced a major and a very weak minor band (lane 3), the band produced by YY1
corresponded to the major band. The slightly slower mobility of the YY1 shifted
band could be due to a small difference in the charge and M.W. of the
recombinant His-YY1. However, this difference in mobility was not observed with the longer
DNA fragment E. Addition of YY1 antibody to the reaction mixture containing
nuclear protein extract (lane 5) or purified YY1 (lane 10) produced a weak
supershift band. Only a small fraction of band was supershifted. This could be due to a partial dissociation of the immuno-complex under high ionic strength PAGE conditions (
31
). The supershift is specific because normal rabbit serum or purified rabbit IgG (lanes 6,7) had no such effect. These results indicate that YY1 binds to the
footprint sequence NPN-
cis
-1.
A luciferase reporter gene assay was employed to study the 5' DNA domain(s) for NPM expression. The 5' deletion mutants of NPM gene were subcloned into the luciferase
reporter gene vector pXP1 (
34
). These constructs were transfected into HeLa cells and the expression of luciferase activity was determined (Materials and
Methods). As shown in Figure
8
, the promoter/ enhancer activity with the DNA sequence up to -405 (fragment II) including the YY1 binding site (-361/-353) is 93.19 " 19.45 * 10
6
RLU/[beta]-gal unit. This activity is 157 times higher than the basal promoter (fragment III) activity (0.59 " 0.12 * 10
6
RLU/[beta]-gal unit). It was also found that the construct with fragment I (which extended from -405 to -741) has a lower activity (48.01 " 9.8 * 10
6
RLU/[beta]-gal unit), indicating a possible negative regulator element in the
region between -741 and -405.
The exon/intron junctions for the human NPM gene are identical to those of the
rat NPM gene (Table
1
). The number and the size of exons between human and rat NPM are the same,
although the human gene spans >25 kb while the rat gene spans ~11 kb. Many repetitive sequences were found in the introns of the human
gene. While the sequence homology between the coding portions of human and rat
NPM is ~94%, we observed only 75% sequence homology in the 5' noncoding region, the promoter and the enhancer regions. Chang and
Olson (
40
) identified two NPM cDNA clones (B23.1 and B23.2) from rat tissues. The long
form (B23.1) contains exons 1-9 and 11-12, and the short form (B23.2) contains exons 1-10. The short form of NPM (B23.2) is a result of
alternative splicing of the NPM gene (
18
). A sequence corresponding to the rat's exon 10 (expressed in the short form)
was also identified in the human NPM gene (with 80% sequence homology). The
expression of exon 10 in human tissue is not known and remains to be
investigated. Intron 4 of the human NPM gene where chromosome break occurs
(discussed below) is longer (911 bp) than the corresponding rat sequence (558
bp), but there is no sequence homology between them.
NPM gene rearrangement is observed in certain types of non-Hodgkin's lymphoma (
20
-
25
). The 5' portion of the NPM gene (in chromosome 5) is fused to other genes in
t(2:5), t(5:17) and t(3:5) chromosome translocations. These findings indicate
that within the NPM gene, there are regions vulnerable to chromosome breaks.
Our results indicate that the chromosome break occurs [for t(2:5) and t(5:17)
translocation] within intron 4 of the NPM gene. Intron 4 is relatively smaller
(911 bp long) than those in Philadelphia chromosome translocation where the
breakpoints are located in a 20 kb region inside a 70 kb intron (
41
). The highly consistent breakpoint of NPM observed in t(2:5) and t(5:17)
suggests that there may be a `fragile site' in intron 4 vulnerable to cleavage
instead of rejoining to the correct chromosome. Intron 4 contains two
Alu
sequences and T-stretches (9-17 Ts) at the 3' end (Fig.
2
B). The
Alu
sequences have 70-80% sequence homology to the
Alu
sequences observed in the human ABL and BCR genes (
36
). The T-stretches at the 3' end may facilitate triplet formation (
42
), altering the DNA structure to favor breakage.
It is hypothesized that abnormal expression of genes due to gene rearrangement
may affect normal cell growth and differentiation. The NPM promoter may cause overproduction of hybrid proteins which have tyrosine kinase (ALK) or retinoic acid receptor activities. These hybrid proteins may play a role in tumorigenesis. It was reported that over-expression of the NPM-ALK hybrid protein caused a malignant transformation of NIH 3T3 cells (
43
).
As a first step toward understanding NPM gene regulation and expression, we
investigated protein bindings in the 5' region of the NPM gene with EMSA and footprinting analysis. With DNA
fragments E (-741/-250) and P (-249/+465), we identified three DNA binding proteins (Fig.
4
). Footprint analysis and competition EMSA (Figs
5
and
6
) showed that the binding site for proteins A and B is between -371/-344. We identified that protein A is YY1. The binding of YY1 to
this sequence is specific because a modified oligo (NPM-
cis
-A) with changes in the consensus sequence could not effectively compete
for its binding to YY1 (Fig.
6
). YY1 is a zinc finger transcription factor that binds to promoters and
enhancers of many viral or cellular genes (
27
-
29
). Depending on which proteins it associates with, YY1 can either be a positive
or negative factor (
27
-
29
,
44
). It is also reported that binding of YY1 bends the DNA near the
promoter/enhancer region which facilitates the interaction of protein factors
on either side of its binding site (
45
).
There are three potential factor binding sites around the YY1 binding site (Figs
1
and
2
). They are (i) E1A-F binding site (AGGACGT) (
37
), located at -294; (ii) ZRE-1 site (TTACACA) (
38
), located at -377; and (iii) two copies of UBP-1 binding site (CTCTCTGG), located at -414 and -550 (
39
). E1A-F, ZRE-1 and UBP-1 are cellular proteins which bind to enhancer elements and
activate viral gene expression. One could speculate, based on this observation,
that the factors responsible for viral gene activation may also affect NPM gene
expression. Nonetheless, the bending of DNA (
45
) in the enhancer/promoter region of NPM gene by YY1 could help to facilitate
interactions among factors and eventually control NPM gene expression. Our
finding that YY1 binds to the 5' region of the NPM gene is particularly significant, since YY1 is also
found to associate with NPM protein (
11
). These findings suggest that NPM may be involved a feedback control mechanism.
We thank Dr Edward Seto for the gift of YY1 protein and antibody, and Dr Ram
Reddy for his critical reviewing of this manuscript. This work was supported by
a NIH grant (CA 42476) from the National Cancer Institute.
*To whom correspondence should be addressed. Tel: +1 713 798 7902; Fax: +1 713
798 3145; Email: pchan@bcm.tmc.edu
Exon no.Exon size5" Splicing site3" Splicing site
Exon 1
154
TTC G/gtaactgc
cag/GT TGT
154
TTC G/gtaactgg
cag/GT TGT
Exon 2
80
ACG/gtacttaa
cacag/GTC
80
ACG/gtatgtaa
cacag/GTC
Exon 3
120
ACG/gtaagggca
ttctag/GTT
120
AC
ttttag/GTT
Exon 4
94
GTA G/gtatgt
ttag/CT GTG
94
GTA G/gtagga
ttag/CT GT
Exon 5
107
CAG/gtaga
aaag/AAA
107
CAG/gtaaa
tcag/AAA
Exon 6
65
GAA GA/gtaagt
tgtag/T GAT
62
GAA GA/gtaagt
tgtag/T GAT
Exon 7
58
AAA/gtgagt
tgcag/TCT
58
AAA/gtacat
tgcag/TCT
Exon 8
87
AAA/gtaag
ttcag/GGA
84
AA
ttcag/GG
Exon 9
102
AAA/gtgagt
tgcag/GCG
102
AAA/gtgagt
tgcag/GCG
Exon 10
332
..poly (A) site
aatag/GGT
324
..poly (A) site
aatag/GGT
Exon 11
75
GAG/gtcaactg
tccag/GCT
75
GAG/gtaactga
cctag/GCT
Exon 12
358
..poly (A) site
346..poly (A) site
REFERENCES
Return