Cell, Vol. 90, 809±819, August 22, 1997, Copyright 1997 by Cell Press
Monoallelically Expressed Gene Related to p53
at 1p36, a Region Frequently Deleted
in Neuroblastoma and Other Human Cancers
Mourad Kaghad,*Helene Bonnet,*Annie Yang,²
Laurent Creancier,*J ean-Christophe Biscan,*
Alexandre Valent,³Adrian Minty,*Pascale Chalon,*
J ean-Michel Lelias,*Xavier Dumont,*
Pascual Ferrara,*Frank McKeon,²
and Daniel Caput*§
Innopole B.P. 137
31676 Labege CEDEX
²Department of Cell Biology
Harvard Medical School
Boston, Massachusetts 02115
³Lab. de Cytogenetique et Genetique
URA 1967 CNRS
Institut Gustave Roussy
p53 acts as a tumor suppressor; its loss of function
appears to conferselective advantages oncells through
deregulated growth and resistance to cell death (Grae-
ber et al., 1996; Kinzler and Vogelstein, 1996).
Despite the widespread presence of p53 mutations
in human malignancies, many tumors develop in the
absence of p53 abnormalities or obvious tumor anti-
gens, most likelydue to aloss of othertumorsuppressor
genes (Weinberg, 1993). Many of these tumor suppres-
sor genes, including Rb in retinoblastoma, NF1 in neu-
rofibramatosis, and DCC and APC in colon carcinomas,
were initially identified through cytogenetic evidence of
loss of heterozygosity (LOH) (Benedict et al., 1983;
Cavenee et al., 1983; Ballester et al., 1990; Buchberg et
al., 1990; Fearon et al., 1987).
Similarly, extensiveinvestigations ofneuroectodermal
tumors, including neuroblastoma, melanoma, and multi-
ple endocrine neoplasm, have suggested the presence
of multiple tumor suppressors at the subtelomeric re-
gion of chromosome 1 (Brodeuret al., 1977; Balaban et
al., 1986; Ross et al., 1995; Sozzi et al., 1988; Dracopoli
et al., 1989). Neuroblastomas with1p LOH canbe subdi-
vided into two classes as having either small deletions
(5 to 10 Mb) at 1p36.2-3 or a clinically more aggressive
form characterized by N-myc amplification and larger
deletions of chromosome 1 including subbands p36 and
p35 (Brodeur et al., 1984; Takeda et al., 1994; Caron et
al., 1995). Notably, the chromosome that sustains the
discrete1p36deletionofthe firstclass ofneuroblastoma
is almost exclusively of maternal origin, indicating that
the putative tumorsuppressorin this regionis imprinted
(Barlow, 1995; Caron et al., 1995). In contrast, N-myc-
amplified neuroblastomas show a LOH at 1p35-1pter
fromeitherchromosome (Caronetal.,1995;Cheng etal.,
1995). These observations suggest that neuroblastoma
develops through different mechanisms of inactivating
alleles of putative tumor suppressors at 1p36 and that
additional genes, at 1p35 and at other loci, influence
tumorigenesis. Theetiology of neuroblastomais compli-
cated by an additional class of neuroblastoma, referred
to as 4S, that initially appears as a widely disseminated,
aggressive disease (Ambros et al., 1995). Remarkably,
a majority ofthe 4S neuroblastomas suddenly and spon-
taneously regress in the absence of treatment. Under-
standing these complex pathways of neuroblastoma in-
duction and progression will require the identification
of the provisional tumor suppressors located on the
short arm of chromosome 1.
We have identified a novel gene encoding a protein,
termed p73, with remarkable sequence similarity to the
DNA-binding, transactivation, and oligomerization do-
mains of p53. We show that p73 has oligomerization
and transactivation properties similar to p53 and that
the p73 gene maps to the 1p36.33 region frequently
deleted in neuroblastoma and other tumors. Inaddition,
we provide evidence to support the notion that alter-
ations in p73 gene expression may be one factor in the
development of neuroblastoma and other tumors.
We describe a gene encoding p73, a protein that
shares considerable homology with the tumor sup-
pressor p53. p73 maps to 1p36, a region frequently
deleted in neuroblastoma and other tumors and
thought to contain multiple tumor suppressor genes.
Our analysis of neuroblastoma cell lines with 1p and
p73 loss of heterozygosity failed to detect coding se-
quence mutations in remaining p73 alleles. However,
the demonstration that p73 is monoallelically ex-
pressed supports the notionthat itis a candidate gene
in neuroblastoma. p73 also has the potential to acti-
vate p53 target genes and to interact with p53. We
propose that the disregulation of p73 contributes to
tumorigenesis and that p53-related proteins operate
in a network of developmental and cell cycle controls.
Spontaneous lesions in the gene encoding the tumor
suppressorp53 havebeenimplicated inthe progression
of a wide range of human tumors (Nigro et al., 1989;
Hollsteinet al., 1991; Levineet al., 1995).The prevalence
of tumors in individuals or mice bearing constitutional
p53 mutations suggests that loss of p53 activity also
contributes to the generation of tumors (Li and Fraumeni,
1969; Malkin et al., 1990; Donehower et al., 1992; J acks
et al., 1994). Wild-type p53 can also be neutralized
through direct interaction witheither cellularproteins or
viral tumor antigens (Lane and Crawford, 1979; Linzer
and Levine, 1979; Werness et al., 1990). p53 appears
to induce cell cycle arrest or apoptosis in response to
cellular stresses such as DNA damage and hypoxia
(Kastan et al., 1992; Livingstone et al., 1992; Lowe et
al., 1993; Hartwell and Kastan, 1994). By this means,
§To whom correspondence should be addressed.
Figure 1. Homology between p73 and p53
Primary amino acid sequences of human p73? and ? were deduced from cDNAs from normal human tissues.
(A) Comparative alignments of coding sequence and gene structure of human p73? and p53. Identities are enclosed in shaded boxes, with
p53 residues frequently mutated in tumors presented in bold. The identity and position of introns within p73 and p53 genes are denoted
above and below the respective sequences.
(B) C-terminal sequence of human p73?, a splicing variant lacking exon 13, is aligned with its site of divergence from p73?.
(C) Homology between the C-terminal sequence of p73? and that of a p53-like protein from squid.
(D) Summary of yeast two-hybrid interaction assays between p53, p73?, and p73?, presented as ordinate values relative to B-galactosidase
activity of p53±p53 interactions.
ResultsPrives, 1996) are conserved and occupy identical posi-
tions in p73 (Figure 1a). No significant homology was
detected between the C-terminal domain (364±393) of
mammalian p53 and p73. However, the C-terminal do-
main of human p73? shows homology with recently dis-
covered invertebrate p53 homologs (Figure 1c), sug-
gesting the possibility that p53 may have evolved from
a more primitive, p73-like gene. In support of this con-
cept, the intron±exon organization of the p73 gene was
found to be similar to that of the p53 gene (Figure 1a).
p73? is encoded by transcripts lacking the 96 nucleo-
tides corresponding to exon 13. This deletion interrupts
the open reading frame, yielding a polypeptide of 499
amino acids (Figure 1b). Both p73? and ? transcripts
were detected by PCR in all human tissues tested, in-
cluding brain, kidney, placenta, colon, heart, liver, spleen,
and skeletalmuscle (data not shown), indicating a wide-
spread, albeit low level, expression of these proteins.
Considering the extensive homology between p53
and p73, including that in the oligomerization domain
ofp53, weassessed whetherthese proteins wouldinter-
act inthe contextof the yeasttwo-hybrid system (Gyuris
etal., 1993).Using this assay, we detected strong homo-
typic interactions between p53 molecules, indicative of
their known ability to form oligomers (Ko and Prives,
1996). In contrast, p73? showed a very low tendency
to form homotypic interactions in this assay. However,
Homology between p73 and the p53
A cDNA encoding p73 was fortuitously discovered in a
hybridization screen of a COS cell cDNA library using
degenerate oligonucleotides corresponding to IRS-1-
binding domains. The coding sequence of p73 was
found to lack any homology to IRS-1 binding domains.
Subsequently, libraries of normal human colon tissue
cDNAs were screened by hybridization to yield cDNAs
encoding p73? and p73?, which are splicing variants of
p73 differing at their C termini (Figures 1a and 1b). The
homology between p73 and p53 is extensive within the
most conserved p53 domains (Zambetti and Levine,
1993; Ko and Prives, 1996) involved withtransactivation
(29% identity with p53 amino acids 1±45), DNA binding
(63% identity with p53 amino acids 113±290), and p53
oligomerization (38% identity in p53 sequence from
319±363) (Figures 1a and 1c). While the homology be-
tween the N terminus of p73 and that required for tran-
scriptional activation by p53 is not strong, a sequence
similar to the MDM2-binding domain of p53 (TFSDLW;
Lin et al., 1994a) is present in p73 as TFEDLW. Signifi-
cantly, residues corresponding to those of p53 fre-
quently mutated in tumors (R175, G245, R248, R249,
R273, andR282)andshownto be requiredforsequence-
specific DNA recognition (Lin et al., 1994b; Ko and
p73, a Gene at Chromosome 1p36, Is Related to p53
arm of chromosome 1 appear frequently in neuroblas-
toma (Takeda et al., 1994; Caron et al., 1995; Cheng et
al., 1995), we performed p73 FISH analysis on various
human cell lines established from neuroblastoma tu-
mors. The SK-N-AS has the smallest 1p deletion of de-
fined neuroblastoma cell lines covering approximately
8 Mb, limited distally by D1Z2 (1p36.33) and proximally
by TNFR2 (1p36.2) (Figure 2b; White et al., 1995; Cheng
et al., 1996). Probing of SK-N-AS with the p73 gene
yielded signal from only one of the two chromosome 1
homologs, demonstrating a p73 LOH in this neuro-
blastoma cellline(Figure 2a).Subsequent, moredetailed
mapping has indicated that the p73 gene is very close
to the D1Z2 marker at 1p36.33, thereby placing p73 at
the distal border of the consensus region of deletion in
neuroblastoma (Figure 2b).
A similar FISH analysis of neuroblastoma cell lines
showing N-Myc amplification and a larger 1p deletion,
including IMR-32 and CHP-212, confirmed the LOH of
the p73 gene in these cells (Table 1). Interestingly, both
the SK-N-SHand the SK-N-MC neuroblastoma celllines
show neither1p (Davidoffetal., 1992;Cheng etal., 1996)
norp73LOH (Table 1), indicating thateitherotherlesions
are responsible for these tumors or that p73 is inacti-
vated by mechanisms besides deletions in these cell
Figure 2. Mapping of p73 Gene to 1p36 and LOH in Neuroblastoma
(A)(Left panel) Fluorescence in situ hybridization (FISH)of a cosmid
probe containing the p73 gene on a normal human chromosome
spread. Chromosomes (in red) are counterstained with propidium
iodide, and p73 gene signals appear as green at the telomere of
1p. (Right panel) FISH analysis on chromosome spreads of the SK-
N-AS neuroblastoma cellline reveals two chromosome 1 homologs
using the centromeric D1Z5 probe (red) but only one chromosome
1 with a p73 signal (green). Chromosomes are counterstained with
(B) Summary of cytogenetic data for the localization of the neuro-
blastoma suppressor locus at 1p36, based on overlapping regions
of deletion in neuroblastoma cell lines and detailed mapping of the
1p deletioninthe SK-N-AS cellline (Cheng etal., 1996).Approximate
positions of genes (in red) and polymorphic markers (in black) are
p73 Expression in Neuroblastoma Cell Lines
To examine p73 and p53 expression in neuroblastoma
cell lines, we developed sensitive RT-PCR reactions to
yield amplicons corresponding to the entire coding se-
quences of p73 and p53 transcripts. RT-PCR products
of p73 transcripts were first obtained from IMR-32, SK-
N-MC, SK-N-SH neuroblastoma cell lines, as wellas the
HT-29 colon carcinoma cell line, but not from the SK-
N-AS cell line (Figure 3a). To gain quantitative informa-
tion on levels of p73 transcript expression, Northern
analysis was performed on these same cell lines. Two
p73 transcripts of 4.4 and 2.9 kb, corresponding to poly-
adenylation variants as determined by direct sequenc-
ing, were detected in HT-29, IMR-32, and SK-N-SH
mRNA, while p73 transcripts were either not detected
in RNA from SK-N-AS cells or present at exceeding
low levels in SK-N-MC cells (Figure 3b, upper panel).
Corresponding Northern analysis of p53 transcripts in
these cell lines revealed a common 2.5 kb transcript
except for SK-N-MC, which displays a previously char-
acterized truncation (Figure 3b, lower panel; Davidoffet
al., 1992). Westernblots to detect the p73? protein were
performed using a polyclonal antibody directed against
the C terminus of p73?. Reflecting the Northern data for
p73 transcripts in these cell lines, extracts from HT-29,
IMR-32, and SK-N-SH cells showed easily detectable
levels of the p73? protein. However, SK-N-MC and SK-
N-AS extracts contained significantly reduced levels of
p73?, showing only a faintband thatmigrates somewhat
faster than the main product seen for HT-29, IMR-32,
and SK-N-SH (Figure 3c). To characterize further the
expression pattern of p73, we performed immunolo-
calization of the endogenous p73? in an array of cell
lines, all of which revealed a pattern of small, punctate
dots in the nucleus of some cells in an asynchronous
p73? displayed strong homotypic interactions, equiva-
lent to that of p53. The potential forheterotypic interac-
tions was also assayed by the two-hybrid system. Inter-
estingly, p53 and p73? displayed significant mutual
interactions in either bait or prey configuration, while
p53 showed negligible interactions with p73?.Weak but
detectable interactions between p73? and p73? were
also evidentinthis assay.Thephysiologicalsignificance
of the apparent oligomerization and restricted hetero-
typic interactions amongst p73?, p73?, and p53 deter-
mined in these assays remains to be established.
p73 Gene Localized to Neuroblastoma
Suppressor Locus at 1p36
To test the possibility that the p73 gene was located at
sites ofsuspected tumorsuppressorgenes, we mapped
the p73 gene using fluorescence in situ hybridization
(FISH)on normalhuman chromosome spreads. Thep73
cosmid probehybridized to the subtelomeric p36 region
of chromosome 1 (Figure 2a). As deletions in the short
Table 1. Analysis of p73 in Neuroblastoma and Other Tumor Cell Lines
(N-Myc single copy)
11p?, deletion in short arm of chromosome 1, as cited in text. NR, none reported.
2Determined by fluorescence in situ hybridization (FISH) with a p73 genomic probe on chromosome spreads, performed as described in
Experimental Procedures. NT, not tested.
3Determined by FISH analysis with the chromosome 1 centromeric probe, D1Z5 (Oncor). NT, not tested.
4Refers to the polymorphism of G/C and A/T alleles, as determined by PCR on genomic DNA and PCR-RL (see Figure 4 and Experimental
Procedures) of products. Minus sign (?) indicates p73 loss of heterozygosity, assumed on the basis of known 1p36 deletions. For cell lines
in which chromosome 1p deletions have not been reported (NR), the A/T allele was never detected, though the presence of other non-G/C
or A/T alleles cannot be formally excluded (denoted by question mark).
5Determined by RT-PCR on mRNA and subsequent sequencing of products. Asterisks indicate a second, 30-cycle round of PCR was required
to generate products for sequencing. (-) indicate a lack of detectable product, despite two rounds of RT-PCR.
6Determined by sequence analysis of open reading frame in expressed transcripts. wt, wild type compared with normal human tissue (not
shown); N/A, not available due to lack of mRNA expression, as determined by RT-PCR. ? indicates 10% of transcripts sequenced contained
a deletion in exon 2.
7Detectable by Western blotting of whole cell lysates, as described in Figure 3 and Experimental Procedures.
8Same as 6, performed on p53 mRNA.
population (Figure 3d). A similardistribution of dots was
revealed upon transfecting mammalian cells with a myc
epitope-tagged p73? expression vector, indicating that
p73autonomously targetsto these intranuclearfoci(Fig-
from control, nonneuroblastoma lines, were sequenced
in their entirety and found to be identical to p73 se-
quences derived from normal human tissue (Table 1).
Western blot analysis of the four neuroblastoma lines
that showed extremely low levels of p73 transcript by
RT-PCR revealed a corresponding lack of p73? protein
(Table 1). In addition, two neuroblastoma lines, SK-N-
BE(2) and SK-N-MC, showed detectable levels of p73
transcripts by one round of PCR but no p73? protein.
These results, together with the absence of mutations
in the remaining allele of p73 LOH neuroblastoma cell
lines, suggest that epigenetic mechanisms such as im-
printing and translational suppression act to limit p73
expression in neuroblastoma cells.
Lack of Coding Region Mutations in p73 Gene
As most neuroblastoma cell lines examined displayed
a LOH of the p73 gene, we asked whetherthe remaining
allele sustained genetic changes that might affect p73
function. We used the p73 RT-PCR reaction to amplify
the coding region of p73 from mRNA of an extended
group of neuroblastoma cell lines. As withSK-N-AS, the
neuroblastoma lines SMS-BC, CHP-212, and SMS-KAN
showed no RT-PCR product corresponding to p73 after
30 cycles, suggesting very low expression levels. How-
ever, reamplification for an additional 30 cycles yielded
sufficient product for sequencing p73 transcripts from
these lines. These PCR products, including six from
neuroblastomacelllines displaying a1p36 LOH (IMR-32,
CHP-212, SMS-BC, SMS-KAN, SK-N-BE(2), and SK-N-
AS), two from neuroblastoma cell lines that showed no
1p deletion (SK-N-SH and SK-N-MC), as well as those
Monoallelic Expression of the p73 Gene
Through our analysis of p73 transcripts in neuro-
blastoma cell lines and tumors, we discovered an allelic
polymorphismconsisting of a double nucleotide substi-
tution (G→A) and (C→T) at positions 4 and 14 of exon
2, just upstream of the initial AUG of p73 (Figure 4a).
We denote these two naturally occurring p73 alleles G/C
and A/T. The G/C, A/T polymorphism occurs in a region
p73, a Gene at Chromosome 1p36, Is Related to p53
over the other. We addressed this issue by performing
the assay on G/C and A/T transcripts mixed at known
ratios and showed that this assay was not obviously
biased for one allele. Moreover, this analysis showed
that we could detect both transcripts even when they
are present in ratios of 20:1 (Figure 4d). To examine
further the possibility of monoallelic expression of p73,
we performedPCR-RL ontranscripts ofperipheralblood
cells from five healthy donors determined to be G/C;A/T
heterozygotes. All five PCR-RL assays revealed a p73
pattern corresponding to either, but not both, the A/T
or the G/C allele, supporting the notion of monoallelic
expression of p73, at least within the 20-fold sensitivity
of this assay (Figures 4c and 4d). Thus, monoallelic ex-
pression of p73 may have particular significance for
neuroblastoma and othertumors that display 1p36 LOH,
as deletion of the active allele may result in a nearly
complete loss of p73 activity. Finally, p73? protein ex-
pression was easily detected in a majority of nonneuro-
blastoma cell lines. Incontrast, only two of eight neuro-
blastoma cell lines yielded p73? signal on Western
blots, and these exclusively express the A/T transcript
Figure 3. Analysis of p73 Expression in Cell Lines
(A) RT-PCR products corresponding to the coding regions of p53
and p73? from various cell lines used for sequence analysis (see
(B) Northern blot analysis of p73 and p53 transcripts inselected cell
lines. The 4.4 and 2.9 kb p73 transcripts result from the use of
distinct polyadenylation sites. All cell lines show a similar 2.5 kb
p53 transcript except SK-N-MC, which contains a previously char-
acterized deletion (Davidoff et al., 1992).
(C) Western blot analysis of total cell lysates cell lines using poly-
clonalantibody directed to C terminus of p73?(top) and monoclonal
antibody to p53 (bottom).
(D) Immunolocalization of endogenous and transfected p73? in cell
lines. (Top left panel) Immunolocalization of endogenous p73? in
U251 human glioblastoma cell line using anti-p73? antibodies, re-
vealing numerous discrete foci with the nucleus of some, but not
all cells in field. (Top rightpanel) Hoechstdye staining of field corre-
sponding to left panel, revealing nuclei of cells. (Bottom left panel)
Immunolocalization of myc epitope±tagged p73? in baby hamster
kidney cells transfected with a p73? expression vector. (Bottom
right panel) Hoechst dye stained nuclei corresponding to field pre-
sented at left.
p21wafInduction and Growth Suppression by p73
The SK-N-AS cell line expresses no detectable p73 pro-
tein and negligible levels of p73 transcript and therefore
represented an ideal model fortesting the effect of rein-
troducing p73 (Baker et al., 1990). SK-N-AS cells were
transfected with a plasmid expressing both the se-
lectable marker for neomycin resistance (Neor) and ei-
ther wild-type p73? or p73?(R292H), a mutant version
homologous with p53(R273H) that is defective for DNA
binding and transcriptional activation (Lin et al., 1994a;
Ko and Prives, 1996). Wild-type p53 or p53(V143A)
(Bakeretal., 1990)were transfectedseparately andused
for comparison with p73 samples. Cells were grown in
the presence of G418, a neomycin analog, and one set
of plates harvested after 48 hrs for Western blotting
analysis. Lysates fromthe transfected cells were probed
with antibodies to detect the expression of p73 or p53
as well as p21waf, a known p53 target gene (El-Deiry et
al., 1993). Significantly, cells expressing wild-type p73
showed elevated levels of p21 protein, comparable to
those seen in wild-type p53 transfectants, whereas mu-
tant p73- and mutant p53-expressing cells both failed
to show a similar p21wafinduction (Figure 5a).
We furthertestedthe effectof exogenously expressed
p73 on SK-N-AS cells using a standard colony assay
(Baker et al., 1990). Identical sets of plates to those
above were maintained under G418 selection for 3
weeks and assayed for colony production. No colonies
were obtained from cells expressing wild-type p73 or
p53 (Figure 5b), while the DNA-binding domain mutants
of each yielded high numbers of colonies. Although the
physiological significance of colony assays in general
is unclear, the obvious distinction between the wild-
type versions of p73 and p53 and their counterpart mu-
tants that fail to bind DNA suggests that the observed
growth suppression and endogenous p21 activation in
SK-N-AS cells are a function of the transcriptionactivity
of p73 and p53.
of the transcript that could theoretically form a stem-
loop structure, possibly anindicationof regulatory func-
tion. Interestingly, sequencing of p73 transcripts from
a wide variety of nonneuroblastoma cell lines, listed in
Table 1, revealed only the G/C allele, while three of eight
neuroblastoma lines possessed an A/T allele. We used
PCR-restrictionlength analysis (PCR-RL) to screenrap-
idly forallele types at boththe DNA andtranscript levels,
taking advantage of the additional Sty1 site resulting
from the double substitution (G→A and C→T) in exon 2
(Figures 4b and 4c). Notably, the single remaining allele
of both IMR-32 and CHP-212 bears the A/T polymor-
phism (Table 1). We were especially intrigued by the
analysis of the SK-N-SH cell line, which shows no p73
LOH and has both G/C and A/T alleles at the genomic
level. At the transcript level, however, using RT-PCR-
RL, we could only detect expression of the A/T allele
(Figure 4c). This result suggested that, at least in SK-
N-SH cells, p73 protein is a consequence of monoallelic
expression. Another possibility that might explain this
result is that our RT-PCR-RL assay may favor one allele
Figure 4. Monoalleleic Expression of p73
(A) Sequence of exon 2 alleleic variants of
p73? transcripts deduced from RT-PCR
product sequencing. The two alleles, which
represent dinucleotide substitutions at posi-
tion 4 and 14 of exon 2, are G4/C14 and A4/
T14 and appear at a region of the 5?-untrans-
lated sequence that might form a stem-loop
(B) Genotypic analysis of p73 G/C and A/T
alleles intennormalblood donors using PCR-
RL. Genomic PCR products that includeexon
2 yield a 482 bp fragment. Sty1 digestion of
PCR products yields two smaller size frag-
ments from AT allele-derived amplicons of
376 bp and 106 bp, respectively, whereas the
GC allele-derived amplicons remain uncut
(lanes 1, 2, 3, 4, 6, 7, and 10). The lanes 5, 8,
and 9 show heterozygotes GC/AT donors.
(C) Analysis of G/C A/T p73 allele expression using Sty1 and Nar1 digestions of transcript RT-PCR products from cell lines and five p73
heterozygote GC/AT individuals (1±5). As with the genomic analysis, the double digestion of RT-PCR products yields specific fragments
identitying each type of allele transcript-derived amplicons (284 bp for GC and 234 bp and 50 bp for AT transcripts). IMR-32, SK-N-SH, as
well as blood cells of individuals 1, 3, 4, and 5 predominantly express the A/T allele at the transcript level, while HT-29 and individual 2 express
the G/C allele in transcripts.
(D) A/T G/C titration assay to determine sensitivity and bias potential of RT-PCR-RL analysis for alleleic expression. G/C and A/T transcripts
from total mRNA of HT-29 and IMR-32 respectively were quantified by Northern blots and mixed at the ratios indicated and analyzed by the
RT-PCR-RL assay described for (C). A/T and G/C alleles can be detected even when mixed with a 20-fold excess of the other allele.
p73 Is Not Activated by Actinomycin D
or UV Irradiation
In light of the structural similarities between p73 and
p53, we asked whether p73, like p53, is induced by
agents that activate the DNA damage checkpoint (Ko
and Prives, 1996). To do this, p53 and p73 protein levels
were assayed in IMR-32 cells following exposure to ei-
ther actinomycin D or ultraviolet radiation (Kessis et
al., 1993; Caelles et al., 1994). Actinomycin D at low
concentrations (1nM)activates theDNA damage check-
point through producing DNA strand breaks, while at
higher concentrations (1 to 2 ?M) inhibits transcription
(Kessis et al., 1993; Caelles et al., 1994). After 24 hrs of
treatment with1nM actinomycinD, p53and p21waflevels
in the cell are markedly elevated, while p73? levels ap-
pear unaffected (Figure 6a). At micromolar concentra-
tions of actinomycin D that inhibit transcription, p53
levels continue to rise above those of untreated cells,
presumably due to a stabilization of the p53 protein. In
contrast,p21andp73? proteinlevels were notenhanced
by actinomycin D over those of untreated cells (Figure
6a). Similar results were obtained with other cell lines.
To assess further the effect of DNA damage on p73?
levels, IMR-32 cells were exposed to ultraviolet (254nm)
radiation and subsequently probed with antibodies to
p73?, p53, and p21. Although p53 protein was markedly
increased in cells at 15 hrs after irradiation, p73? levels
fail to show a similar increase after such treatments
(Figure 6b). A similar failure to induce p73? was ob-
served incells exposedto gammaradiationorgenotoxic
agents such as doxorubicin (data not shown). Thus,
despite the structural similarities between p73 and p53
and theircommonability toinduce p21waf, these proteins
do not respond in a similar manner to DNA damaging
The identification of a novel gene located at chromo-
some 1p36.2-3 that encodes proteins with significant
homology to p53 may have implications for our under-
standing of the etiology of neuroblastoma and other
tumors as well as for p53 evolution and function. The
remarkable homology between the core domain of p73
and the DNA-binding domain of p53, togetherwithp73's
ability to induce the p21wafprotein, suggest that p73
acts, in part, as a transcription factor. While it is not
obvious that p73 and p53 share common functions, the
presenceofdistal1p alterations ina widearrayoftumors
in additionto neuroblastoma, including melanoma (Dra-
copoli et al., 1989), hepatocellularcarcinoma (Yeh et al.,
1994), and ductile breast carcinoma (Genuardi et al.,
1989), supports the notion that p73 operates in path-
ways that coordinate cell growth, death, and differenti-
Neuroblastoma: LOH and Monoallelelic
Expression of p73 Gene
Neuroblastoma is thought to arise from primitive neu-
roectodermal stem cells that fail, at various stages, to
differentiate into sympathetic neurons, Schwann cells,
or melanocytes (Knudson and Meadows, 1980; Ross et
al., 1995). Associated cytogenetic characteristics in-
clude discrete and gross deletions of the short arm of
chromosome 1, the amplification of N-Myc, and distur-
bances in cell ploidy (Takeda et al., 1994; Ambros et al.,
1995; Caron et al., 1995; Cheng et al., 1995). Genetic
analyses of neuroblastoma tumors harboring discrete
1p36.2-3 LOH have shown this deletion to be sustained
predominantly by the maternally derived chromosome,
thereby implicating one or more imprinted tumor sup-
pressor genes in this region (Versteeg et al., 1995). In
p73, a Gene at Chromosome 1p36, Is Related to p53
Figure 6. p73 Induces p21 but Is Not Responsive to DNA Damage
(A) Immunoblots of lysates prepared from IMR-32 cells treated with
either 1 nM actinomycin D for 24 hrs or 1 ?M actinomycin D for 2
hrs were probed with antibodies to p73?, p53, and p21waf.
(B) Immunoblots of lysates prepared from IMR-32 cells grown for
15 hrs after 254 nm UV-C irradiation (20 or 100 J oules/m2) probed
with antibodies to p73? and p53.
al., 1995; Versteeg et al., 1995). The disruption of p73,
then, may be one in a numberof events that contribute
to the onset and progression of neuroblastoma.
While the majority of neuroblastoma cell lines ana-
lyzed here lacked p73 expression, we were able to de-
tect protein in two exceptions: the SK-N-SH cell line,
which lacks detectable 1p or p73 LOH, and IMR-32,
which harbors a large 1p35-36 deletion characteristic
of N-myc-amplified neuroblastomas. Notably,sequence
analysis of p73 transcripts in both SK-N-SH and IMR-
32 revealed a dinucleotide A/T polymorphism in the 5?
untranslated region. Using this polymorphism, we dem-
onstrated monoallelic expression of the p73 gene not
only in cell lines but also in peripheral blood cells of
normal donors. To date, we have analyzed only one
informative family regarding the parental dominance in
expression of p73 alleles. Inthis case, a normal A/T;G/C
heterozygote donor was found to express the A/T allele
at the transcript level, while the maternal and paternal
genotypes were A/T;A/T and G/C;G/C, respectively
(data not shown). While obviously limited, this analysis
shows a case in which the p73 gene, like the putative
tumor suppressor at 1p36, is expressed predominantly
from the maternal allele. Further, is there any signifi-
cance to the observation that in the only two neuro-
blastoma lines examined with p73 protein expression,
namely IMR-32 and SK-N-SH, this protein was derived
from an A/T transcript? At present, we have no experi-
mental evidence fora functional difference between the
G/C and A/T alleles. We note, however, that the p73
transcripts contain an in-frame CUG codon 5? of this
polymorphic region, which, at least in several genes,
including basic fibroblastgrowthfactor andC-Myc, acts
as an alternative translation initiation codon (Prats et
al., 1989). Moreover, conceptual translation from this
CUG codonyields a coding sequencesomewhathomol-
ogous to that of the transcriptional activation domain
of p53. We are presently investigating the possibility
that this 5? CUG codon is in fact an alternative site for
the initiation of p73 translation.
In this study, we concentrated on neuroblastoma cell
lines showing 1p LOH and found that the remaining p73
Figure 5. Expression of Exogenous p73 and p53 in SK-N-AS cells
(A) SK-N-AS cells were transfected with vectors expressing Neor
alone or together with p73?, the p73? mutant (R292H), p53, or
p53(V143A), grown for 48 hr, and analyzed by Western blot using
antibodies to p53, p73, p21waf, and Bcl-2.
(B) Colony formation in SK-N-AS cells transfected as in (A) after
growth for 3 weeks under G418 selection.
contrast, the more aggressive stage 3 and 4 neuro-
blastomas appear to have amplified N-Myc and have
sustained largertelomeric deletions,including 1p36 and
1p35 from chromosome 1 of random parental origin
(Cheng et al., 1995). Whether 1p35 harbors additional
tumor suppressorgenes or one that affects N-Myc am-
plification or modifies expression of other genes on 1p
is unknown, but it is obvious that neuroblastoma is a
highly complex and heterogeneous disease involving
the disruption of activities at multiple loci.
The chromosomal localization and monoallelic ex-
pression of p73, its frequent LOH in neuroblastoma,
and its conspicuous lack of expression in a majority of
neuroblastoma cell lines are all consistent with the no-
tion that p73 is a candidate for the putative, imprinted
neuroblastoma suppressor gene at 1p36. Importantly,
p73 maps to the distal borderof the consensus deletion
found in neuroblastoma, as defined by a wide range of
tumor cell lines (Cheng et al., 1996). It is interesting to
note, however, that a constitutional 1p36 deletionin one
case reportedly retained the D1Z2 marker at 1p36.33
(Biegel et al., 1993). This potentially places p73 immedi-
atelydistalto the regionofoverlap betweentheconstitu-
tional case and deletions in neuroblastoma tumor cell
lines. Itremains possible,nonetheless, that theconstitu-
tional deletion encompasses a region extremely close
to p73 and could mediate its disruption without a strict
deletion of the gene. On the other hand, various studies
havereportedtumorcelllines and constitutionaltranslo-
cations at 1p36 whose break points do not map within
the deletion in SK-N-AS. This has led to the speculation
that multiple neuroblastoma suppressor genes exist at
1p36 (Takeda et al., 1994; Amler et al., 1995; Laureys et
allele lacked mutations similar to those that inactivate
p53. However, as only ?30% to 40% of sporadic neuro-
blastoma tumors display an obvious 1p LOH (Caron et
al., 1995), an extensive analysis of neuroblastoma tu-
mors will be required to determine the actual signifi-
cance of p73 lesions, monoallelic expression, and the
A/T polymorphisminthe developmentofneuroblastoma
and other diseases.
activities of p53 in cell growth regulation or promote
survival of a disregulated cell. Importantly, recent stud-
ies have begun to establish links between active cell cy-
cle progression and differentiation (Huttner and Brand,
1997), thereby suggesting a potential mechanism by
which p73 might be required for differentiation of neu-
roectodermal stem cells. Analysis of p73 activity during
differentiation and throughout the cell cycle should pro-
vide insight into these issues.
It is unclear at present how the identification of this
novel gene will impact on our understanding of p53.
From anevolutionary standpoint, however, it is interest-
ing to note that p73 shares greater homology with p53-
like proteins found in mollusks than with p53 itself. It is
formallypossible thatp53evolved froma moreprimitive,
p73-likegene involvedinmany aspects ofdifferentiation
and growth control to assume more specific functions
in cell cycle control and tumor suppression. It is also
possible, given the general quadruplication of the Hox
complex and other essential genes at the chordate±
vertebrate transition (Holland, 1996), that other p53-like
genes exist and comprise a p53 regulatory network.
Finally, the discovery of a gene related to p53 will
likely contribute to our understanding of the complex
etiology of neuroblastoma and other diseases involving
1p LOH and may lead to new avenues in developmental
and cell cycle regulation, tumor biology, and alternate
strategies for cancer therapy.
The identification of p73, its homology with p53, and its
link to tumor suppressors at 1p36 raise fundamental
questions regarding p73 function in development and
cellcyclecontrol. Does p73, forinstance, actinamanner
similar to p53 to sense cellular stresses such as DNA
damage and hypoxia and integrate this information for
cell cycle and cell death regulation? Whereas p73 is
shownto be capable of enhancing levels of endogenous
p21wafprotein, it is not induced in cell lines by agents,
including UV radiation and actinomycin D, that result in
p53 stabilization and activation. Although experiments
are under way to determine what signals influence p73
expression, it is apparent that p73 and p53 may be
serving distinctfunctions inthe cell. An equally pressing
question is whether p73 and p53 interact to yield novel
activities not displayed by either molecule alone. Given
the dramatic consequences of heterodimerization by
members of the c-Myc family (c-Myc, Myb, Max, and
Mxi1) on target gene expression (Bernards, 1995), it will
be important to determine if such interactions occur
amongst p53-like proteins. Ourinitial efforts focused on
the yeast two-hybrid system for testing potential inter-
actions between p53 and p73. Interestingly, both p53
and p73? show strong homotypic interactions, while
p73? has a very low propensity for self interactions in
this assay.Moreover, p73? displays an ability to interact
withboth p73?and p53, despite its preferentialassocia-
tion with other p73? molecules. Although we are pres-
ently examining p73±p53interactions inavariety ofcells,
the data obtained from the yeast interaction assay sug-
gest that such interactions are possible and may be
likely. The analysis of p73±p53 interactions in cells may
be especially critical for understanding neuroblastoma,
as p53 is often wild type in this disease, and yet recent
studies indicate that it is aberrantly cytoplasmic (Moll
et al., 1995). An intriguing, though speculative, explana-
tion for this unusual behavior of p53 in neuroblastoma
is that p53 requires p73 for some activities that are lost
inthe absenceof p73.Definitive answers to whetherp73
has consequential interactions with p53 will obviously
require extensive genetic and biochemical studies.
Major questions remain for understanding possible
p73 functions incellcycleregulationand growthcontrol.
Curiously, the majority of nonneuroblastoma tumor cell
lines was found to express high levels of wild-type p73
transcript andprotein, suggesting a role forp73 inprolif-
eration. This observation is perplexing in that most nor-
mal tissues show low levels of p73transcriptand protein
(data not shown). One possible explanationfor elevated
p73 in tumorcell lines is that a disruption of normal p53
function, as seen in a majority of these cell lines, results
in compensatory or deleterious upregulation of p73 ex-
pression. In this scenario, p73 may then either assume
All cell lines were grown in Dulbecco's modified Eagle's medium
(DMEM) containing 4 g/l glucose, 10% fetal calf serum, 2 mM gluta-
mine, and 5000 ?g/ml penicillin and streptomycin.
The two-hybrid assay was performed essentially as described (Gy-
uris et al., 1993). Nucleotide sequences corresponding to amino
acids 72±393 of p53, 85±636 of p73?, and 85±499 of p73? were
introduced into the vectors pEG202 (resulting in a fusion protein
withthe LexA DNA-binding domain)and pJ G4-5(resulting inafusion
protein with a transcriptional activation domain). The yeast strain
EGY48 was transformed with a pEG202 plasmid, a pJ G4-5 plasmid,
and the pSH18.34 plasmid containing the lacZ gene under the con-
trol of eight lexA operators.
p21wafProtein Induction Assay
p73 cDNAs were cloned into a pCMV1 vector containing a 200 bp
lamin 5?-untranslated region and encoding an N-terminal myc tag
(Heald et al., 1993). pCMV-p53 and pCMV-p53V143A were gifts of
Dr.Bert Vogelstein(Bakeretal.,1990).IMR-32 cellswere transfected
on 100 mm plates with 20 ?g of the indicated plasmid and extracts
prepared by lysis in SDS sample buffer. Lysates were fractionated
on a 10% polyacrylamide gel and transferred to Immobilon mem-
branes by electrophoresis. Membranes were probed with a p21
polyclonal antibody (Calbiochem) and developed by chemilumines-
DNA Damage Response
Cell lines, including IMR-32, MCF-7, and SK-N-SH, were grown to
50% confluenceand incubated with either1nM or1?Mactinomycin
D (Calbiochem) for 24 hrs and 2 hrs, respectively. Cells were then
washed with PBS, lysed in 2? SDS sample buffer, fractionated
on 10% polyacrylamide gels, and electrophoretically transferred to
Immobilon (Millipore) membranes. Membranes were probed with a
polyclonal antibody to p73?, a monoclonal antibody to p53 (ATCC,
p73, a Gene at Chromosome 1p36, Is Related to p53
Bethesda), or a monoclonal antibody to p21 (Calbiochem). For ex-
amining responses to ultravioletradiation, media was removed from
plates and cells exposed to 20 and 100 J oules/m2254 nm UV-C.
Media was returned to the plates and cells grown for an additional
15 hrs prior to lysis and Western blotting, as above.
5?CACCTGCTCCAGGGATGC and 5?AAAATAGAAGCGTCAGTC de-
rived from intronic sequences were employed. For PCR-2, 2 ml of
purified PCR-1 reaction products were used with a more internal
set ofprimers (5?CAGGCCCACTTGCCTGCC and 5?CTGTCCCCAAG
CTGATGA). The resulting amplicons were Sty1 digested and ana-
lysed on 1.5% agarose gels.
Fluorescence in Situ Hybridization (FISH)
The p73 cosmid probe was labeled by nick translation using biotin-
16-dUTP (Boehringer Mannheim) according to a BRL (Bethesda
Research Laboratories) protocol. Twenty microliters of the hybrid-
ization solution containing 40 ng of biotinylated probe and 10 ?g
of human placental DNA was incubated at 80?C for 5 min. DNA was
allowed to reanneal at 37?C for 6 hrs before placing on the slides.
Chromosome spreads of peripheral blood lymphocytes and the SK-
N-AS, SK-N-MC, and IMR-32 neuroblastoma cells were prepared
from asynchronous populations grown in DMEM with 10% fetal
bovine serum and exposed to colcemid (500 nM) or nocodazole
(100 nM) for 2 hr. After hybridization in 50% formamide, 2? SSC,
and 10% dextransulfate for 12 hr at 37?C, biotinylated probes (p73,
D1Z5) were detected with an FITC-conjugated anti-biotin antibody
(GIBCO), and the digoxigenin-labeled probe (D1Z5) was revealed
by a rhodamine-conjugated anti-digoxigenin antibody (Boehringer
Mannheim). Chromosomes were counterstained with either DAPI
(4,5-diamino-2-phenylindole) or propidium iodide, and slides were
mounted in antifade solution (Vectastain).
We would like to thank members of the Caput and McKeon groups,
as well as Howard Green, Tim Bestor, Alain Bernheim, David Shire,
and J ean Benard for helpful discussions and Linda Buck for cell
lines. This work was supported in part by grants from the National
Institutes of Health, the Council for Tobacco Research, and the
American Cancer Society to F. M.
Received J une 6, 1997; revised J uly 22, 1997.
Ambros, P.F., Ambros, I.M., Strehl, S., Bauer, S., Luegmayr, A.,
Kovar, H., Ladenstein, R., Fink, F.-M., Horcher, E., Printz, G., et al.
(1995). Regression and progression in neuroblastoma. Does genet-
ics predict tumour behaviour? Eur. J . Cancer 31A, 510±515.
Amler, L.C., Corvi, R., Praml, C., Savelyeva, L., Le Paslier, D., and
Schwab, M. (1995). Reciprocal translocation at 1p36.2/D1S160 in a
neuroblastoma cell line: isolation of a YAC clone at the break. Eur.
J . Cancer 31A, 527±530.
Baker, S.J ., Markowitz, S., Fearon, E.R., Willson, J .K., and Vo-
gelstein, B. (1990). Suppression of human colorectal carcinoma cell
growth by wild-type p53. Science 249, 912±915.
Balaban, G.B., Herlyn, M., Clark, W.H., J r., and Nowell, P.C. (1986).
Karyotypic evolution in human malignant melanoma. Can. Genet.
Cytogenet. 19, 113±122.
Ballester, R., Marchuk, D., Boguski, M., Saulino, A., Letcher, R.,
Wigeler, M., and Collins, F. (1990). The NF1 locus encodes a protein
functionally related to mammalianGAP and yeast IRA proteins. Cell
Barlow, D.P. (1995). Gametic imprinting in mammals. Science 270,
Benedict, W.F., Murphree, A.L., Banerjee, A., Spina, C.A., Sparkes,
M.C., and Sparkes, R.S. (1983). Patient with 13 chromosome dele-
tion: evidence that the retinoblastoma gene is a recessive cancer
gene. Science 219, 973±975.
Bernards, R. (1995). Flipping the Myc switch. Curr. Biol. 5, 859±861.
Biegel, J .A., White, P.S., Marshall, H.N., Fujimori, M., Zackai, E.H.,
Scher, C.D., and Brodeur, G.M. (1993). Constitutional 1p36 deletion
in a child with neuroblastoma. Am. J . Hum. Genet. 52, 176±182.
Brodeur, G.M., Sekhon, G., and Goldstein, M.N. (1977). Chromo-
somal aberrations inhumanneuroblastomas.Cancer40, 2256±2263.
Brodeur, G.M., Seeger, R.C., Schwab,M., Varmus, H.E., and Bishop,
J .M. (1984).Amplificationof N-myc inuntreatedhuman neuroblasto-
mas correlates with advanced disease stage. Science 224, 1121±
Buchberg, A.M., Cleveland, L.S., J enkins, N.A., and Copeland, N.G.
(1990). Sequence homology shared by neurofibromatosis type-1
gene and IRA-1 and IRA-2 negative regulators of the RAS cyclic
AMP pathway. Nature 347, 291±294.
Caelles, C., Helmberg, A., and Karin, M. (1994). p53-dependent
apoptosis in the absence of transcriptional activation of p53-target
genes. Nature 370, 220±223.
Caron, H., Peter, M., van Sluis, P., Speleman, F., de Kraker, J .,
Laureys, G., Michon, J ., Brugieres, L., Voute, P.A., Westerveld, A., et
al. (1995). Evidence fortwo tumoursuppressorloci onchromosomal
bands 1p35±36 involved in neuroblastoma:one probably imprinted,
another associated with N-myc amplification. Hum. Mol. Genet. 4,
Cavenee, W.K., Dryja, T.P., Phillips, R.A., Benedict, W.F., Godbut,
R., Gallie, B.L., Murphee, A.L., Strong, L.C., and White, R.L. (1983).
Transfections and Immunofluorescence
Construction of the myc-tagged p73 vectors, transfection, and im-
munofluorescence were done essentially as described (Heald et al.,
1993).Cells were fixed with3% formaldehyde inphosphate-buffered
saline (PBS), blocked with 3% milk in PBS containing 0.1% Triton
X-100 (PBST), and sequentially incubated with primary and CY3-
conjugated secondary antibodies (J ackson ImmunoResearch) for
30 min each at room temperature. DNA was labeled using Hoechst
dye 33258 (Sigma)at1 ?g/mlfor 1 min. The9E10 mouse monoclonal
antibody against the c-Myc epitope was obtained from the Cell
Culture Facility at Harvard University. The rabbit polyclonal anti-
p73? antibody was generated against a C-terminal p73? (427±636)
glutathione S-transferase fusion protein.
SK-N-AS Colony Formation Assay
SK-N-AS cells on100 mmplates were transfected with the indicated
pCDNA3 vectors using the calcium phosphate method. Forty-eight
hours later, cellextracts were prepared, immunoblotted, and probed
with antibodies to p73, p53, p21, and Bcl-2. Identical sets of plates
were grown in 500 ?g/ml G418 (GIBCO) for 3 weeks, fixed, stained,
and counted (Baker et al., 1990).
Molecular Biology Methods
RNA preparation, Northern blot analysis, immunoblotting, genomic
and cDNA cloning, as well as screening and nucleotide sequencing,
were performed using standard protocols (Maniatis et al., 1992).
p73 Gene and Transcript Analysis
cDNA synthesis was performed using 5 ?g of total RNA incubated
in a 20 ml volume reaction containing 50 mM Tris-HCl [pH 8.3], 10
mM DTT, 10 mM KCl, 0.5 mM dNTP, 30 ?g RNAsin (Promega), 150
?g superscript II reverse transcriptase (GIBCO; BRL) for 1 hr at
37?C. PCR reactions were performed with either 2 ml of reverse
transcriptase products or 300 ng genomic DNA in 50 ml reaction
containing 50 mM Tris-HCl [pH 9.2], 16 mM (NH4)2SO4, 1.75 mM
MgCl2, 10% DMSO, 0.3 mM each primer, and 3.5 ?g of TAQ and
PWO DNA polymerases (Boehringer Mannheim). The amplification
sequence consisted of 30 cycles of 95?C/0.5 min, 58?C/1 min, 68?C/
2.5 min after starting with a denaturation step at 95?C for 1 min and
ending at 68?C/10 min. The amplicons were purified by spin dialysis
sequentially on S400 and P10 resins. For restriction length analysis
ontranscripts, amplicons were generated as described above using
primers (5?CGGGACGGACGCCGATG and 5?AGACCGTAGACCGTC
ATC derived from the p73 cDNA sequence and analysed on 1.5%
agarose gels after digestion with Nar1 and Sty1 restriction endonu-
cleases.Restrictionlength analysis forgenomic DNA was performed
using two sequential, nested PCR reactions. For PCR-1, primers
Expression of recessive alleles by chromosomal mechanisms in
retinoblastoma. Nature 305, 779±784.
Cheng, N.C., Van Roy, N., Chan, A., Beitsma, M., Westerveld, A.,
Speleman, F., and Versteeg, R. (1995). Deletion mapping in neuro-
blastoma cell lines suggests two distinct tumor suppressor genes
in the 1p35-36 region, only one of which is associated with N-myc
amplification. Oncogene 10, 291±297.
Cheng, N.C.,Chan,A.J .K.,Beitsma, M.M., Speleman, F.,Westerveld,
A., and Versteeg, R.A. (1996). Human modifier of methylation for
class IHlagenes (memo-1) maps to chromosomalbands 1p35±36.1.
Hum. Mol. Genet. 5, 309±317.
Davidoff, A.M., Pence, J .C., Shorter, N.A., Iglehart, J .D., and Marks,
J .R. (1992). Expression of p53 in human neuroblastoma- and neu-
roepithelioma-derived cell lines. Oncogene 7, 127±133.
Donehower, L.A., Harvey, M., Slagle, B.L., McArthur, M.J ., Mont-
gomery, C.A., J r., Butel, J .S., and Bradley, A. (1992). Mice deficient
for p53 are developmentally normal butsusceptible to spontaneous
tumours. Nature 356, 215±221.
Dracopoli, N.C., Harnett, P., Bale, S.J ., Stanger, B.Z., Tucker, M.A.,
Housman, D.E., and Kefford, R.F. (1989). Loss of alleles from the
distal short arm of chromosome 1 occurs late in melanoma tumor
progression. Proc. Natl. Acad. Sci. USA 86, 4614±4618.
El-Deiry, W.S., Tokino, T., Velculescu, V.E., Levy, D.B., Parsons, R.,
Trent, J .M., Lin, D., Mercer, W.E., Kinzler, K.W., and Vogelstein, B.
(1993). WAF1, a potential mediator of p53 tumor suppression. Cell
Fearon, E.R., Hamilton, S.R., and Vogelstein, B. (1987). Clonal analy-
sis of human colorectal tumors. Science 238, 193±197.
Genuardi, M., Tsihira, H., Anderson, D.E., and Saunders, G.F. (1989).
Distal deletion of chromosome Ip in ductalcarcinoma of the breast.
Am. J . Hum. Genet. 45, 73±82.
Graeber, T.G., Osmanian, C., J acks, T., Housman, D.E., Koch, C.J .,
Lowe, S.W., and Giaccia, A.J . (1996). Hypoxia-mediated selection
of cells with diminished apoptotic potential insolid tumours. Nature
Gyuris, J ., Golemis, E., Chertov, H., and Brent, R. (1993). Cdi1, a
human G1 and S phase protein phosphatase that associates with
Cdk2. Cell 75, 791±803.
Hartwell, L.H., and Kastan, M.B. (1994). Cell cycle control and can-
cer. Science 266, 1821±1828.
Heald, R., McLoughlin, M., and McKeon, F. (1993). Human Wee1
maintains mitotic timing by protecting the nucleus from cytoplasmi-
cally activated cdc2 kinase. Cell 74, 463±474.
Holland, P.W.H. (1996). Hox genes and chordate evolution. Dev.
Biol. 173, 382±395.
Hollstein, M., Sidransky, D., Vogelstein, B., and Harris, C.C. (1991).
p53 mutations in human cancers. Science 253, 49±53.
Huttner, W.B., and Brand, M. (1997). Asymmetric divisionand polar-
ity of neuroepithelial cells. Curr. Opin. Neurobiol. 7, 29±39.
J acks, T., Remington, L., Williams, B.O., Schmitt, E.M., Halachmi,
S., Bronson, R.T.,and Weinberg, R.A. (1994).Tumorspectrumanaly-
sis in p53-mutant mice. Curr. Biol. 4, 1±7.
Kastan, M.B., Zhan, Q., El-Deiry, W.S., Carrier, F., J acks, T., Walsh,
W.V., Plunkett, B.S., Vogelstein, B., and Fornace, A.J ., J r. (1992). A
mammaliancellcyclecheckpointpathway utilizingp53 and GADD45
is defective in ataxia-telangiectasia. Cell 71, 587±597.
Kessis, T.D., Slebos, R.J ., Nelson, W.G., Kastan, M.B., Plunkett,
B.S., Han, S.M., Lorincz, A.T., Hedrick, L., and Cho, K.R. (1993).
Human papilloma virus 16 E6 expressiondisrupts the p53-mediated
cellular response to DNA damage. Proc. Natl. Acad. Sci. USA 90,
Kinzler, K.W., and Vogelstein, B. (1996). Life (and death) in a malig-
nant tumour. Nature 379, 19±20.
Knudson, A.G., J r., and Meadows, A.T. (1980). Regression of neuro-
blastoma IV-S: a genetic hypothesis. New Eng. J . Med. 302, 1254±
Ko, L.J ., and Prives, C. (1996). p53: puzzle and paradigm. Genes
Dev. 10, 1054±1072.
Lane, D.P., and Crawford, L.V. (1979). T antigen is bound at a host
protein in SV 40-transformed cells. Nature 278, 261±263.
Laureys, G., Versteeg, R., Speleman, F., van der Drift, P., Francke,
U., Opdenakker, G., and Van Roy, N. (1995). Characterisation of the
chromosome breakpoints in a patient with a constitutional translo-
cation t(1;17)(p36.31;q11.2-q12) and neuroblastoma. Eur. J . Cancer
Levine, A.J ., Wu, M.C., Chang, A., Silver, A., Attiyeh, E.F., Lin, J .,
and Epstein, C.B. (1995). The spectrum of mutations at the p53
locus.Evidence fortissue-specific mutagenesis,selectionof mutant
alleles, and a ªgain of functionº phenotype. Ann. NY Acad. Sci. 768,
Li, F.P., and Fraumeni, J .F., J r. (1969). Soft-tissue sarcomas, breast
cancer, and other neoplasms. A familial syndrome? Ann. Intern.
Med. 71, 747±752.
Lin, J ., Chen, J ., Elenbaas, B., and Levine, A.J . (1994a). Several
hydrophobic amino acids in the p53 amino-terminal domain are
required for transcriptional activation, binding to mdm-2 and the
adenovirus 5 E1B-kD protein. Genes Dev. 8, 1235±1246.
Lin, J ., Wu, X., Chen, J ., Chang, A., and Levine, A.J . (1994b). Func-
tions of the p53 protein ingrowth regulationand tumorsuppression.
CSH Sym. Quant. Biol. 59, 215±223.
Linzer, D.I., and Levine, A.J . (1979). Characterization of a 54K dalton
cellular SV40 tumor antigen present in SV40-transformed cells and
uninfected embryonal carcinoma cells. Cell 17, 43±52.
Livingstone, L.R., White, A., Sprouse, J ., Livanos, E., J acks, T., and
Tisty, T.D. (1992). Altered cell cycle arrest and gene amplification
potential accompany loss of wild-type p53. Cell 70, 923±935.
Lowe, S.W., Schmitt, E.M., Smith, S.W., Osborne, B.A., and J acks,
T. (1993). p53 is required for radiation-induced apoptosis in mouse
thymocytes. Nature 362, 847±849.
Malkin, D., Li, F.P., Strong, L.C., Fraumeni, J .F., J r., Nelson, C.E.,
Kim, D.H., Kassel, J ., Gryka, M.A., Bischoff, F.Z., Tainsky, M.A., et
al. (1990). Germ line p53 mutations in a familial syndrome of breast
cancer, sarcomas, and other neoplasms. Science 250, 1233±1238.
Maniatis, T., Fritsch, E.F., and Sambrook, J . (1992). Molecular Clon-
ing: A Laboratory Manual (Cold Spring Harbor: Cold Spring Harbor
Moll, U.M., LaQuaglia, M., Benard, J ., and Riou, G. (1995). Wild-type
p53 protein undergoes cytoplasmic sequestration in undifferenti-
ated neuroblastomas but not in differentiated tumors. Proc. Natl.
Acad. Sci. USA 92, 4407±4411.
Nigro, J .M., Baker, S.J ., Preisinger, A.C., J essup, J .M., Hostetterm,
R., Cleary, K., Bigner, S.H., Davidson, N., Baylin, S., Devilee, P., et
al. (1989). Mutations in p53 gene occur in diverse human tumour
types. Nature 342, 705±708.
Prats, H., Kaghad, M., Prats, A.C., Klagsbrun, M., Lelias, J .M., Liau-
zun, P., Chalon, P., Tauber, J .P., Amalric, F., Smith, J .A., and Caput,
D. (1989). High molecular mass forms of basic fibroblast growth
factor are initiated by alternative CUG codons. Proc. Natl. Acad.
Sci. USA 86, 1836±1840.
Ross, R.A., Spengler, B.A., Domenech, C., Porubcin, M., Rettig,
W.J ., and Biedler, J .L. (1995).Human neuroblastoma I-type cells are
malignant neural crest stem cells. Cell Growth Differ. 6, 449±456.
Sozzi, G., Bertoglio, M.G., Pilotti, S., Rilke, F., Pierotti, M.A., and
Della Porta,G. (1988).Cytogenetic studies inprimary and metastatic
neuroendocrine Merkel cell carcinoma. Cancer Genet. Cytogenet.
Takeda, O., Homma, C., Maseki, N., Sakurai, M., Kanda, N., Schwab,
M., Nakamura, Y., and Kaneko, Y. (1994). There may be two tumor
suppressor genes on chromosome arm 1p closely associated with
biologicallydistinctsubtypes ofneuroblastoma.Genes Chrom. Can-
cer 10, 30±39.
Versteeg, R., Caron, H., Cheng, N.C., van der Drift, P., Slater, R.,
Westerveld, A., Voute, P.A., Delattre, O., Laureys, G., Van Roy, N.,
et al. (1995). 1p36: every subband a suppressor? Eur. J . Cancer
Weinberg, R. (1993). Tumor suppressor genes. Neuron 11, 191±196.
Werness, B.A., Levine, A.J ., and Howley, P.M. (1990). Association
of human papillomavirus types 16 and 18 E6 proteins with p53.
Science 248, 76±79.
White, P.S., Maris, J .M., Beltinger, C., Sulman, E., Marshall, H.N.,
p73, a Gene at Chromosome 1p36, Is Related to p53 Download full-text
Fujimori, M., Kaufman, B.A., Biegel, J .A., Allen, C., Hilliard, C., et al.
(1995).A regionof consistentdeletion inneuroblastoma maps within
human chromosome 1p36.2±36.3. Proc. Natl. Acad. Sci. USA 92,
Yeh, S.H., Chen, P.J ., Chen, H.L., Lai, M.Y., Wang, C.C., and Chen,
D.S. (1994). Frequent genetic alterations at the distal region ofchro-
mosome 1p in human hepatocellular carcinomas. Cancer Res. 54,
Zambetti, G.P., and Levine, A.J . (1993).A comparisonof the biologi-
cal activities of wild-type and mutant p53. FASEB J . 7, 855±865.
EMBL Accession Number
DNA sequences corresponding tothe humanp73? and p73?cDNAs
have been deposited in the EMBL database under the accession
number Y11416 EMBL.