ABSTRACT
MSSP proteins have been identified by their binding to an upstream element of
c-myc
. Independently, two different approaches yielded two cDNA clones highly
homologous to the MSSP cDNAs, suggesting an involvement of MSSP in the
regulation of the cell cycle (scr2) and in the repression of HIV-1 and ILR2
[alpha]
-promoter transcription (human YC1). Screening human genomic libraries, we
have isolated clones belonging to two different gene loci. Whereas the human
MSSP
gene 1 turned out to be intronless, the organization of the coding sequence
within gene 2 is more complex. It spans more than 60 kb and contains 16 exons
(including two alternative first exons), ranging from 48 to 287 bp,
respectively. The intron sizes vary from 0.1 to more than 13 kb. Gene 1 has
been completely sequenced. A deletion series of its upstream region was
conjugated to the luciferase gene, but the transfection of the constructs did
not display any promoter activity. Moreover, compared with gene 2 and the cDNA
sequences known so far, about 20 point mutations as well as flanking direct
repeats have been detected in the
MSSP
gene 1, showing that it possesses all the characteristics of processed
retropseudogenes. Sequence analysis of a 1.7 kb fragment of the 5
'
flanking region of the
MSSP
gene 2 revealed that the promoter of gene 2 lacks consensus sequences for TATA
and CCAAT boxes, is GC-rich, and contains numerous potential transcription factor binding
elements including an Sp1 binding site. DNase I footprinting experiments showed
that the putative Sp1 site was bound by proteins. The results of primer
extension and S1 mapping analyses suggested the transcription of the gene
starts at multiple positions upstream from the initiator methionine codon.
Luciferase assays employing progressive deletions of the 1.7 kb promoter region
allowed us to define the minimal promoter region of 428 bp (-488/+) and revealed a complex pattern of the transcriptional regulation
the human
MSSP
gene 2. Furthermore, it can be concluded that the
MSSP
gene 2 encodes both MSSP-1 and MSSP-2, and moreover scr2 and human YC1.
The involvement of the proto-oncogene
c-myc
in crucial functions such as cell proliferation and differentiation requires a
careful control in every cell type. Its delicately regulated expression and the
elements involved therein have been the subject of numerous studies (
1
-
3
). A sequence of 21 bp about 2 kb upstream of the human
c-myc
gene has been shown to be essential for replication and transcription and to
constitute both a putative DNA replication origin and a transcriptional
enhancer (
4
,
5
). Its stimulation of SV40 DNA replication (
6
) and the functional substitution of its core sequence for the AT-stretch of the SV40 origin (
7
), in addition to the binding of a
c-myc
protein complex to it, suggested the role of the 21 bp sequence as a target for
DNA-protein interaction (
8
). Indeed, several proteins showing direct binding to either of its strands
could be identified and were named MSSP (
c-
To shed full light upon this complex matter, the knowledge of the gene structure
and mode of expression of MSSP is a prerequisite. In the present report, human
genomic clones representing two gene loci have been isolated and characterized
in order to gain insight into the regulation of the
MSSP
gene(s). Whereas
MSSP
gene 1 shares all the characteristics of processed pseudogenes,
MSSP
gene 2 codes for all MSSP cDNAs, as well as scr2 and YC1. Besides the
description of the detailed structure of
MSSP
gene 2, we have also investigated its 5'-flanking region to identify
cis
elements important to drive transcription by deletion analysis.
A human placental genomic library in EMBL3-SP6/T7 and a human leukocyte genomic library in EMBL3 were purchased from
Clontech. The phages were propagated in host bacteria NM538 and LE392. The
infection and plating procedures were according to the recommendations of the
manufacturer. Plaques were screened under stringent conditions with [[alpha]-
32
P]dCTP- labelled gel-purified DNA fragments (10
5
-10
6
c.p.m./ml of hybridization solution) as probes. Human MSSP-1 was labelled with [[alpha]-
32
P]dCTP using a random primer labelling kit from Boehringer Mannheim. MSSP-2 cDNA and genomic DNA fragments (Fig.
1
) were labelled with [[alpha]-
32
P]dCTP by nick translation (
16
). Prehybridization and hybridization were performed in 50% formamide at 42oC (
16
). Membranes were washed twice in 3* SSC/0.1% SDS at 37oC and twice in 0.1* SSC/0.1% SDS at 50-68oC (depending on the background of the respective
probe) for 30 min. The filters were then autoradiographed and analyzed with a
bioimaging analyzer (BAS 2000, Fuji Film Co.) and/or placed in contact with a
Fuji X-ray film and an intensifying screen and exposed at -70oC for 1-3 days. Selected recombinants that hybridized to the
screening probes were rescreened and purified. Phage DNA from the purified
positive plaques was prepared by the bacteriophage lysate method (
16
). Inserts were excised and subcloned into pUC19 and pBluescript (Stratagene).
Subcloned fragments were analyzed and intron lengths were determined by a
combination of restriction endonuclease digestion, Southern blotting, PCR
analysis, and nucleotide sequencing. PAC clones were obtained from Genome
Systems [St. Louis, MO; clone addresses PAC-85-H1 (
MSSP
gene 2), PAC-280-C14 (
MSSP
gene 1) and PAC-320-C24 (
MSSP
gene 1)] after screening with a probe spanning exons II-IV, which was produced by PCR with exon-specific primers and MSSP-2 cDNA as a template. With these clones PCR and sequence
analysis were performed.
Nucleotide sequencing was performed both manually using the chain termination
method of Sanger
et al.
(
17
) with a Sequenase 2.0 kit (US Biochemical Corp.) and
Bca
Best kit (Takara Shuzo Co., Ltd.) and automatically on a model 373A DNA sequence
(Applied Biosystems) using a fluorescent dideoxy terminator kit. In the case of
the
MSSP
gene 1, nested deletions were performed on appropriately double-digested subclones with
Exo
III for different time points, mung bean nuclease-treated, ligated and transformed in
Escherichia coli
strains DH5[alpha] or C600. The resultant plasmids were sequenced with universal primers.
To determine the sequence of each exon and adjacent sequence of the
MSSP
gene 2, synthetic oligonucleotides corresponding to known cDNA and genomic
sequences were used besides universal primers (M4, RV, SP6, T3, and T7).
HeLa cells were washed with phosphate-buffered saline (PBS) and lysed with ISOGEN (Nippon Gene) according to the
instructions of the manufacturer. After purification by an additional phenol
extraction and ethanol precipitation, RNA was resuspended in sterile water.
Primer extension analysis was performed using a 20mer oligonucleotide
complementary to position +39 to +20 of the
MSSP
gene 2 (GCCGTGCAGGGTCGCGGACA). The primer was labelled at the 5' end with [[gamma]-
32
P]ATP and T4 polynucleotide kinase. The labelled nucleotide was purified on a
Sephadex G-50 spin column. A quantity of 6 * 10
5
c.p.m. of the primer nucleotide was co-precipitated with 120 [mu]g of HeLa total RNA and resuspended in 30 [mu]l hybridization buffer (40 mM PIPES, pH 6.4, 0.4 M NaCl, 1 mM
EDTA, and 0.2% SDS). The mixture was heated to 55oC during 10 min and annealed at 37oC for more than 3 h. The RNA and the annealed oligonucleotide were
precipitated with isopropanol and rinsed with 70% ethanol. The pellet was
resuspended in reverse transcriptase buffer with 1 mM dNTP, 130 U of RNase
inhibitor, and 25 U Moloney Murine Leukemia virus reverse transcriptase.
Elongation was carried out for 2 h at 37oC. The reaction products were phenol/chloroform-extracted, ethanol-precipitated, and resuspended in formamide loading buffer. One
third was electrophoresed on an 8% polyacrylamide-8 M urea sequencing gel along with a sequencing ladder using the same
primer for the dideoxy sequencing of a 5'-genomic clone.
pMSSP-Luc was digested with
Hin
dIII and treated with bacterial alkaline phosphatase before digestion with
Sma
I. The
Sma
I-
Hin
dIII fragment of 545 bp containing MSSP-1 promoter was end-labelled with [[gamma]-
32
P]ATP and T4 polynucleotide kinase and was used for a probe. Cytoplasmic RNA
(100 [mu]g) and the labelled DNA probe (10
5
c.p.m.) were co-precipitated with ethanol, suspended with 50 [mu]l of hybridization buffer containing 80% formamide, 40 mM PIPES (pH
6.4), 400 mM NaCl and 1 mM EDTA, heated at 100oC for 8 min, and hybridized overnight at 42oC. The RNA mixtures were then mixed with 400 [mu]l of ice-cold S1 nuclease buffer containing 0.25 M NaCl and 300 mM
sodium acetate, 3 mM ZnSO
4
, 100 [mu]g/ml of salmon sperm DNA, and 10 U S1 nuclease (Takara), and incubated 25oC for 30-60 min. The S1-resistant DNA hybrids were precipitated and
electrophoresed on a 10% polyacrylamide denaturing gel.
The promoterless plasmid pGV-B (PicaGene
TM
, TOYO INK) served as the vector backbone for all the luciferase expression constructs. pMSSP-Luc: Nucleotide sequences of the PCR primers used were 5'-GCTCGAGGTCTAAACCATAGAAC-3' for MSSP-N and 5'-GAAGCTTCATGAAGCTGGAAGGG-3' for MSSP-C. After
the PCR reaction with the above primers on the [lambda] clone containing the upstream region of MSSP gene 2 as a template, the
product was digested with
Xho
I and
Hin
dIII and was inserted to the
Xho
I-
Hin
dIII sites of pGV-B. p[Delta]X-Luc: The
Xba
I-
Hin
dIII fragment of pMSSP-Luc was first inserted to the
Xba
I-
Hin
dIII site of pBluescript SK(-). The
Sac
I-
Hin
dIII fragment from the construct was then inserted to the
Sac
I-
Hin
dIII sites of pGV-B. p[Delta]B-Luc: The
Bam
HI-
Hin
dIII fragment of pMSSP-Luc was inserted to the
Bam
HI-
Hin
dIII site of pBluescript SK(-). The
Sac
I-
Hin
dIII fragment from the construct was then inserted to the
Sac
I-
Hin
dIII sites of pGV-B. p[Delta]S-Luc: pMSSP-Luc was digested with
Sma
I and the larger fragment yielded was self-annealed. p[Delta]P-Luc: pMSSP-Luc was digested with
Xho
I and
Pst
I and the larger fragment yielded was treated with Klenow fragment prior to self-ligation.
Human HeLa cells were cultured in Dulbecco's modified Eagle's Medium (DMEM)
supplemented with 10% calf serum. Five [mu]g of the respective reporter plasmid and 2 [mu]g of the [beta]-galactosidase expression vector (pCMV-[beta]-gal), carrying the cytomegalovirus (CMV)
promoter, were co-transfected to the cells (60% confluent) by the calcium phosphate co-precipitation method (
18
). Four to five hours after transfection, the cells were boosted with 20%
glycerol for 2-3 min at room temperature, then incubated for 48 h.
The transfected cells were washed with PBS and lysed in the plates using 200 [mu]l of a detergent solution (lysis buffer PicaGene
TM
, TOYO INK). The cell extract was then centrifuged for 5 s in an Eppendorf
microcentrifuge and the supernatant was collected. The transfection efficiency
was normalized by a [beta]-galactosidase assay, set up in 300 [mu]l according to the standard procedure (
16
), by incubation at 37oC until a yellow colour developed. The absorbance of the solution was then
measured on a double beam spectrophotometer at 420 nm. The luciferase activity
of the extract was determined by mixing standardized aliquots in a total of 20 [mu]l lysis buffer with 100 [mu]l of luciferase substrate (PicaGene
TM
, TOYO INK) in a vial. Immediately after mixing, the light intensities emitted
by the samples were measured on a luminometer (lumicounter ATP-300, Advantec Toyo Ltd.). Background luciferase activity was assessed in
assays from parallel cultures transfected with the promoterless plasmid pGV-B.
A human placental genomic library in EMBL3 SP6/T7 was screened with labelled
full-length MSSP-1 cDNA (
9
) using the plaque hybridization method. Three intense clones (A2, A24, and A26)
were isolated, subcloned and subjected to further analysis. The restriction
analysis and the sequencing of the ends of the subclones revealed that these
clones represented two different genomic loci, termed
MSSP
gene 1 (A2 and A26) and
MSSP
gene 2 (A24) (Fig.
1
). Partial sequence analysis of the 5'-end of the genomic clone A26 down to the unique
Sal
I site (223 bp downstream from the putative translation start in MSSP-1 cDNA) (Fig.
1
A) revealed an intronless region identical to the 5'-end of the MSSP-1 cDNA (except for three mismatches) and indicated that this
genomic clone extends about 3.5 kb upstream of the start of the cDNA. Attempts
to screen for a sequence homologous to the head sequence of scr2 (
11
) (exon I[alpha] in Fig.
2
) in the far distal 5'-region of gene 1 with the genomic fragment B (Fig.
1
A) and by genomic PCR did not yield the expected result. Following genomic
Southern experiments showed the uniqueness of this sequence in the human genome
as well as the uniqueness of another short region specific for MSSP-2 cDNA (
10
) (exon X in Fig.
2
), missing in both MSSP-1 cDNA and gene 1 (data not shown). The discovery of the co-occurrence of these two stretches in one cDNA clone (MSSP-2) theoretically excluded the possibility of the existence of
a sequence homologous to the scr2 head upstream of gene 1 (Fig.
2
).
Since A24 encompassed only the last eight exons of the human
MSSP
gene 2 (Fig.
1
B), further screenings with genomic fragments (C and E) and a fragment,
containing ~500 bp of the 5'-end of the MSSP-2 cDNA down to the unique
Pst
I site (
10
) (D series), were performed and yielded clones falling into three non-overlapping regions. As neither the screening of the placental genomic
library and of a human leukocyte genomic library with other genomic fragments
(F, G, and H) nor PCR amplification across the gaps using genomic DNA as
template proved a successful strategy for obtaining the missing sections, a PAC
library was screened with a probe spanning exons II-IV. PCR and sequence analysis revealed that two of these clones belong to
gene 1 (clone addresses PAC-280-C14 and PAC-320-C24) and one to gene 2 (clone address PAC-85-H1). Using exonic, intronic and I[alpha] 5'-flanking primers, it
could be demonstrated that this clone encompasses at least the region from the
second
Eco
RI restriction site upstream of exon I[alpha] (Fig.
1
B) down to exon VII, and the existence of the exons I[beta], II and IV therein was confirmed. These exons had not been covered by the
bacteriophage clones. Thus, a contig spanning the whole genomic locus of MSSP
gene 2 was finally established with the overlapping bacteriophage and PAC
clones.
The sequences in and around the exons of the
MSSP
gene 2 and those of the junctions of bacteriophage and plasmid subclones were
determined. The comparison of the sequence of the human
MSSP
gene 2 (Fig.
3
) with those of the cDNAs published so far shows that the human
MSSP
gene 2 is organized into 16 exons (including two alternative first exons, two
optional ones and one with an internal splice site) (Fig.
2
) and 15 intervening sequences, spanning a total of more than 60 kb. The exons
are distributed sparsely at the 5'-end of the gene but rather densely at the 3'-end (Fig.
1
B). Their sizes are rather small (Table
1
). All the exon-intron junction sequences conform to the GT/AG rule (
19
) (Table
2
). The sequence of the known exons coincides with those of all the known cDNAs,
except for some mismatches most probably introduced by cloning procedures or
sequence misreadings.
Screening human genomic libraries with the MSSP-1 and MSSP-2 cDNAs and several genomic fragments, clones from two different
genomic loci were obtained. Whereas the human
MSSP
gene 1 turned out to be intronless, the organization of the coding sequence
within gene 2 is more complex. Gene 1 has been completely sequenced. The
alignment of gene 1 with the upstream region of the alternative first exon I[beta] and the downstream region of the last exon of gene 2 shows homology close
to identity (Fig.
3
B). Compared with the exonic sequences of gene 2 and the cDNA sequences known so
far, including those of recently published expressed sequence tags (EST) and
that of MSSP-3, the
MSSP
gene 1 contains about 20 point mutations, none of which interrupts the reading
frame nor causes frameshifts, which demonstrates that none of the cDNAs results
from a transcript of this gene. The region of gene 1 which corresponds to the
cDNA sequences is bounded by 11 bp direct repeats, the homology between the two
genes, however, extends even beyond them. The 5' direct repeat is surrounded by sequences related to exon I[beta]. The situation of the sequence around the 3' direct repeat is more complex. The homology between gene 1
and gene 2 in the 3'-flanking region continues beyond the putative polyadenylation
signal, spanning about 470 bp interrupted after ~250 bp by a 299 bp insert, which comprises the downstream direct repeat
sequence. The lack of this tract in gene 2 could be interpreted by its deletion in gene 2
rather than by its insertion in gene 1 after the formation of gene 1.
Interestingly, the bounding sequences of the gap show a high degree of
homology. As long as no sequence data of exon I[beta] upstream of -200 in Figure
3
B are available, further homology in the 5'-flanking region cannot be excluded, either. The lack of a poly-A stretch before the 3' direct repeat might be due to reverse transcription
before its addition to the fully spliced RNA. Negative Northern data, no
hybridization of the homologous 3'-flanking sequences to HeLa RNA, would support this hypothesis. The
percentage identity to the mRNA at the nucleotide level is 98.6% considering
only the coding sequences, a typical value for a processed retrogene,
suggesting that the
MSSP
gene 1 has arisen relatively recently in evolution. Commonly, the analysis of
retrogenes does not go beyond a sequence comparison. Experiments to demonstrate
the lack of promoter activities or transcription sites are hardly performed. In
view of the absence of deletions, insertions or mutations to stop codons in the
reading frame of gene 1, these approaches were considered necessary to provide
further evidence for a final conclusion. Deletion series (from the 5' and the 3' end) of its upstream region were cloned upstream of the
luciferase gene into vectors with and without the SV40 promoter, but never was
any significant activity above background level obtained with constructs of the SV40-promoterless series. An S1 mapping was started at a very early stage,
before the knowledge of the structure of gene 2, assuming the uniqueness of the
5' sequence of gene 1. We primarily obtained numerous non-specific bands due to S1 nuclease-sensitive sites. After the discovery of a second copy of this
region in the human genome (exon I[beta]) it was realized that an independent analysis of this region is
impossible and that an optimal choice of the probe and an accurate
interpretation of the results has to await the determination of the precise 5' end of the homology between the two genes and the functional analysis of
the I[beta]5'-flanking region. Since only the fortuitous placement of a
retrosequence next to an active promoter could theoretically result in
transcription, the absence of deletions, insertions or mutations to stop codons
within the reading frame is only of secondary importance. In the case of gene
1, even on the assumption that it may be transcribed at a very low level, which
cannot be ruled out completely, the mutation Met
155
-> Thr
155
disrupts an RNP consensus and most probably precludes translation of the
putative transcript into a functional polypeptide. In conclusion, although the
precise reconstruction of its formation may be difficult to accomplish, the
MSSP
gene 1 arguably shares all the hallmarks of retropseudogenes.
A set of recombinant DNA clones that contain the entire
MSSP
gene 2 were isolated, and the gene 2 was shown to span more than 60 kb and to
contain a total of 16 exons (including two alternative first exons), ranging
from 48 to 287 bp. The intron sizes vary from 0.1 to more than 13 kb. The 5'-flanking region of the
MSSP
gene 2 contains a functional promoter: A relatively high promoter activity was
observed in the region from -488 to +61 (relative to the transcription site), which can be divided
into a sequence essential for transcription (the region up to -196) and the other sequence necessary for maintenance of a high
expression level (-488 to -196). The complex patterns of transcriptional activity of the 5' deletion mutants suggests that both positive and negative
elements are involved in the regulation of
MSSP
gene 2 transcription. The presence of consensus binding sites for various
transcription factors implies, but does not prove, their involvement in the regulation of the promoter activity. The presence of overlapping binding
sites (e. g. Sp1, AP-2, GC-box etc. at -460) suggests a balance activity of these transcription
factors in different physiological conditions. Among the putative binding
sites, at least the Sp1 binding site was protected from DNase I digestion by
the nuclear proteins from HeLa cells, suggesting the involvement of Sp1 in the MSSP gene 2 transcription. Other notable features of the
nucleotide sequence in the 5'-flanking region of the
MSSP
gene 2 are the presence of a 26 bp TG repeat (TG-element) at position -1001 to -976 and two series of GGA repeats from -173 to -102. Tandem repeat sequences of 2-5 reiterated nucleotides are frequently
found in eukaryote genomes. The most common type is the dinucleotide CA repeat
which can appear in up to 10
5
different locations, each of which contains up to 60 bp (
31
). Their varying length provides a useful system for the generation of genetic
markers which can be used for mapping and linkage analysis (
32
). These elements can induce a conformational change from right-handed B-DNA to left-handed Z-DNA and negative supercoiling (
33
). It is not clear whether these conformational changes play a role in
transcriptional regulation, either as part of an enhancer or as part of a
silencer. Additional investigations by deletion analysis and site-directed mutagenesis will be required to define more precisely the
cis
-acting elements and
trans
-acting factors important in the cell cycle-regulated expression of the human
MSSP
gene 2.
Potential polyadenylation sequences (AACAAA and AAGAAA) in the 3' end of the human
MSSP
genes differ from the putative poly(A) signal AAUAAA. However, polyadenylation
sites different from the canonical poly(A) signal have been reported (
34
,
35
), suggesting that a perfect hexanucleotide AAUAAA is not an absolutely
essential element in the efficient polyadenylation of MSSP transcripts.
The divergence of the cDNA sequences at the point where the 5' exons splice onto exon III (or also onto exon II in the case of exon I[alpha]) suggested that gene 2 might have an additional 5' exon located somewhere upstream of exon III. The 12 bp
head sequence in the MSSP-1 cDNA (corresponds to the region -25 to -36 in Fig.
3
B), which is only slightly different from the 3' end of exon 1[alpha] (-135 to -149 in Fig.
3
A), also exists in gene 1. Assuming that the homology between the two genes
continues farther towards the 5' end, an 18 bp primer homologous to the 12 bp and 6 additional bp in gene
1 (-25 to -42 in Fig.
3
B) was designed and used for sequencing the PAC clone containing gene 2.
Gratifyingly, this approach finally resulted in the discovery of exon 1[beta] and extended the known scope of homology between the two
MSSP
genes up to -200 in Fig.
3
B. All cDNAs cloned so far are consequently not transcripts of different genes
but alternatively spliced products of the same gene. In other words, gene 2
encodes MSSP-1, MSSP-2 and MSSP-3, as well as scr2 and YC1.
Further studies will be necessary to pinpoint exon I[beta], to perform functional analysis of its 5'-flanking region and to map its transcription start site(s),
although we could not detect any transcript in the region upstream of exon III
under the conditions used in this experiment. These should also include: a
comparison of the expression patterns of I[alpha] and I[beta] transcripts; their relative abundance at different developmental
stages and/or in different tissues; the identification of the set or
combination of transcription factors for each promoter; and a comparison of
their secondary structures. Thus, the meaning and importance of the
differential promoter usage could be determined. The differential utilization
of two alternative promoter sequences may even provide a mechanism for
translational regulation of the
MSSP
gene 2. Due to the non-excludable possibly non-coding nature of exon I[alpha] and I[beta] (all fusion proteins of MSSP used in
in vitro
studies lacked the scr2 head sequence and started at +1 in Fig.
3
B), transcription initiation from both promoters might even result in identical
protein products. Thus, the transcription under the control of multiple
promoters would result in an alternative splicing of sequences within the 5' UTR, which might regulate the expression of the
MSSP
gene 2 at the translational level. Lastly, determining the sequence of the
alternate candidate promoter in the I[beta] 5'-flanking region should reveal the precise range of homology
between the two
MSSP
genes and should eventually allow the development of more specific probes to
prove or disprove the possibility of a transcription of gene 1.
Along with the knowledge of the gene structure, the dissection of the structure
and transcriptional activity of the
MSSP
gene 2 promoter region reported herein will facilitate further study on its
regulation, thus contributing to a better understanding of the physiological
and cellular roles of MSSP, e.g. its involvement in the modulation of the
biological functions of
c-myc
.
This work was supported by grants from the Japanese Ministry of Education,
Science and Culture. We thank Kiyomi Takaya for technical assistance. C. H.
received a Foreign Student Scholarship from the Japanese Ministry of Education,
Science and Culture.
REFERENCES
Return