ã
Oncogene (2001) 20, 7234 ± 7242
2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00
www.nature.com/onc
Cellular proteins localized at and interacting within ND10/PML
nuclear bodies/PODs suggest functions of a nuclear depot
Dmitri Negorev1 and Gerd G Maul*,1
1
The Wistar Institute, 3601 Spruce Street, Philadelphia, Pennsylvania, PA 19104, USA
ND10, PML bodies or PODs have become the de®ning
nuclear structure for a highly complex protein complement involved in cell activities such as aging, apoptosis,
the cell cycle, stress response, hormone signaling,
transcriptional regulation and development. ND10 are
present in many but not all cell types and are not essential
for cell survival. Here, we review the cellular proteins
found in ND10, their few known interactions and their
contribution to the ND10 structure per se and to functions
elsewhere in the nucleus. The discrepancy between the
functions of the ND10 proteins and the nonessential
nature of the structure in which they are aggregated at
their highest concentrations leads to the conclusion that
the proteins function elsewhere. The regulated recruitment of speci®c proteins into ND10 as well as their
controlled release upon external induced stress points to a
regulated nuclear depot function for ND10. These nuclear
depot functions seem important as nuclear defense against
viral attack and other external insults. Oncogene (2001)
20, 7234 ± 7242.
Keywords: ND10; PML; Daxx; Sp100; nuclear bodies;
protein interaction
Introduction
Although this series of reviews surveys the molecular
analysis of acute promyelocytic leukemia in the last
decade, it should not be forgotten that the nuclear
structure that so prominently features as the location
of one of the fusion protein partners of the primary
molecular lesion, is not an oncogenic domain. Rather,
it is a cellular domain that had long been neglected for
lack of interest, support and ultimately probes for
investigation. The connection to an aspect of a rare
cancer has brought this nuclear structure over the
horizon of many with the bene®cial e€ect for the cell
biological elucidation of one of the last major nuclear
domains shrouded in complete darkness. That it was
the light microscope that started to shed the ®rst
illuminating rays on this domain is only ®tting and the
number of accidental discoveries and accumulation of
various probes began to drive the initial phenomenological characterization. What function ND10 (nu*Correspondence: GG Maul; E-mail: [email protected]
clear bodies, PML bodies or PODs) have, has been
asked for the last 40 years. But only during the last
decade has the basis for postulating potential functions
been laid by the often accidental discovery of various
proteins that are accumulated in these domains. The
presence of ND10 in most but not all cell types, the
survival of mice without these domains, and the
regulated recruitment and release of speci®c protein
are ®ndings that must be considered in assigning
function to the structure relative to the function of
the individual component proteins. We use the term
nuclear domain 10, ND10, instead of PML bodies
since most aggregates of PML, Sp100, Daxx and other
associated proteins are not present as a body (exclude
other nuclear components) (Maul et al., 1995) and
avoid the term PODs since the protein aggregates as
such are not oncogenic. Many authors consider ND10
a ubiquitous nuclear organelle, and we will here review
the sparse literature that indicates otherwise, followed
by a survey of the proteins that are aggregated in
ND10 or beside it. The aggregation of so many
proteins presumably depends on speci®c interactions
between them, and several interactions have been
identi®ed in the recent past. Based on these interactions
and the absence of many proteins essential for most
nuclear functions, and the apparently regulated recruitment/release of proteins upon some signaling pathway
induced by external insults, we postulate a regulated
depot function for ND10.
Presence of ND10 in various cell types
Various nuclear bodies have been discovered and
described using electron microscopic techniques. Those
eventually shown to contain the de®ning Sp100 and
PML proteins were described as `dense granular bodies
and pale round structures, at times showing a complex
lamellar pattern' (de The et al., 1960; Hinglais-Guillaud
and Bernhard, 1961). They were found in cancer cells
and in cells responsive to hormones (Doucas et al.,
1999; Padykula et al., 1981). When described as `newly
rediscovered' by Brasch and Ochs (1992), they had
been seen in hormone treated rooster liver cycling
according to hormone level (Ochs et al., 1995). ND10
are increasingly designated as `doughnut shaped' by
observers using light microscopy (Lam et al., 1995;
Yeager et al., 1999; Wu et al., 2000). This designation
Evidence of nuclear depot function for ND10
D Negorev and GG Maul
has entered review articles (Kass-Eisler and Greider,
2000; Zhong et al., 2000b) and is overtly wrong as the
topology of a doughnut has its center as part of its
outside, whereas a hollow sphere or oblate sphere, the
projection of which is seen by light microscopy, has its
center inside. The designation `doughnut shaped'
should be avoided, as the wrong mental image is
formed, which does not allow for certain material to be
surrounded by ND10-associated proteins. The shell
formation by PML and its concomitant segregation of
ND10-associated proteins into the interior of the
sphere have begun to be reported for Sp100 (Bell et
al., 2000; Lallemand-Breitenbach et al., 2001).
Most current analytical work on ND10 centers on
tissue culture cells. Information on the presence and
distribution of nuclear bodies and ND10 in vivo, i.e.,
di€erent cell types and tissues, is less well developed in
purely descriptive terms and even less from an
experimental approach. Several surveys have been
conducted, one in rats using a monoclonal antibody
reacting with PML (Lam et al., 1995), others of human
organs (Flenghi et al., 1995; Gambacorta et al., 1996;
Koken et al., 1995) and one with antibodies that react
with Sp100 and with an ill-de®ned monoclonal antibody that identi®es a protein or protein modi®cation
called NDP55 (Ascoli and Maul, 1991; Cho et al.,
1998). The latter study expanded the ®nding that sexhormone dependent cells have increased ND10 size,
with breast ducal epithelium, endometrium, myometrial
cells and testicular Leydig cells containing more than
10 visible dots per nucleus compared with the much
lower number in other cell types. It seems signi®cant
that these cells cycle and vary substantially in their
level of protein synthesis. That cycling cells vary in
ND10 frequency has been reported (Koken et al.,
1995). Additional evidence of hormone dependence
came from the correlation between the level of ND10
in the endometrium and serum estrogen levels (Koken
et al., 1995). Moreover the number and size of ND10
were shown to vary with the estrogen and progesterone
receptor status of ®broadenomas. Tumor cells that are
strongly positive for estrogen receptors had many large
ND10 (Cho et al., 1998). At the opposite end of ND10
frequency are cells without ND10. No ND10 were
found in the germinal center of spleen, lymph nodes
and the Peyer's patch (Daniel et al., 1993; Flenghi et
al., 1995; Gambacorta et al., 1996; Lam et al., 1995).
They are also absent from the neuronal lineage. Large
neurons seem to have no ND10 although only Sp100
has been shown to be missing from neuron-derived
cells such as NT2 cells and neuroblastoma cell lines.
Sp100 in these cell lines cannot be upregulated by
interferon (Hsu and Everett, 2001; Ishov et al., 1999;
Negorev et al., unpublished results). The other cell
types where ND10 cannot be recognized are both male
and female germ cells as determined with PML- and
Sp100-reactive antibodies (Cho et al., 1998; Gambacorta et al., 1996; Lam et al., 1995). Loss of some
ND10-associated antigens is speci®c for certain cell
types. For example, Sertoli cells, while having high
PML content lack Sp100, while hepatocytes lack
reactivity with NDP55 while having PML and Sp100
(unpublished). In most cases, it remains to be seen
whether lack of expression of certain ND10 proteins is
maintained when cells are placed in culture and begin
cycling or are interferon induced. Cell cycle variation
in the frequency and size of ND10 has been reported
(Koken et al., 1995) but that ND10 are `regulated' has
not been adequately established. Cell lines that
maintain a silenced phenotype for Sp100 such as
neuroblastoma cells could be used like a knock out
cell line. Their consistent absence of Sp100 might be
useful to investigate the physiological changes that
become apparent upon conditional reconstitution.
From the various staining intensities in di€erent
tissues it is dicult to draw testable conclusions. Most
prevalent and speci®c seems the sex hormone dependent increase of ND10. This said, it is not clear
whether the increase in ND10 frequency re¯ects
transcriptional induction or the consequence of increased or decreased cell proliferation, or whether
induction of a new set of proteins has a secondary
e€ect on ND10. The ®nding that MCF7 cells exposed
to 1075 to 1077 M 17-estradiol show a signi®cant
increase in ND10 (Cho et al., 1998) points to a
promising experimental system that might be exploited.
The position of ND10 has ®rst been de®ned as
interchromosomal by electron microscopy, locating
these domains in the interchromatinic granule clusters
(Puvion-Dutilleul et al., 1995). By immuno¯uorescence
microscopy, ND10 are often found associated with
splicing domains or speckles corresponding to the
interchromatinic granule clusters and by a method
called stress induced chromosome condensation
(SICC), they were found in the interchromosomal
spaces (Plehn-Dujowich et al., 2000). The position of
ND10 may be important for the success of DNA
viruses in associating with the structure since viruses
appear to traverse the nucleus along the intrachromosomal spaces (Bell et al., 2001; Tang et al., submitted).
Viruses might move by guided di€usion along the
interchromosomal space and thus, through the exclusion of much of the nuclear space by chromosomal
territories, be more likely to lodge at ND10 (Maul,
1998).
7235
Proteins interacting at ND10
By de®nition, proteins interacting at ND10 constitute
this domain. A higher concentration of these proteins
at speci®c nuclear loci provides for their visualization
over the lesser concentration in the rest of the nucleus.
That most ND10-associated proteins are present
throughout the nucleus perhaps with the exception of
the nucleolus has been documented (Muller et al.,
1998; Lehembre et al., 2001; Kamitani et al., 1998a).
By immunohistochemistry such nuclear staining is routinely seen although whether all of the nuclear staining
is due to the presence of speci®c proteins or to a
certain amount of background staining or cross
contamination in Western blots is dicult to ascertain.
Oncogene
Evidence of nuclear depot function for ND10
D Negorev and GG Maul
7236
Oncogene
Suppression of signal can come by the formaldehyde
®xation and is most pronounced by heat treatment
during in situ hybridization experiments (unpublished
observations). The proportion of PML, Sp100, Daxx
and others on chromatin, extrachromatinic spaces or
ND10 has not been well de®ned but is quite important
as those components may be the target of ND10associated proteins.
Within the structurally diverse nuclear space, whereever a protein reaches higher concentration one
reasonably expects a higher activity of that protein.
ND10 may be an exception from such expectation.
Polymerase II is active throughout the nucleus. Its
increased concentration at ND10 (Bregman et al.,
1995; von Mikecz et al., 2000) does not increase the
amounts of RNA found in ND10. In fact less is found
than in the surrounding space (Boisvert et al., 2000;
Ishov and Maul, 1996). Although some transcription
factors have been identi®ed at ND10, not all those
necessary for transcription are present (unpublished).
The enhanced accumulation of transcriptional activators such as p53 at ND10 is only due to speci®c
manipulations such as 95 overexpression of speci®c
PML isoforms (Fogal et al., 2000) or in a low
percentage of cells after g-radiation (Guo et al.,
2000). The involvement of transcription-enhancing or
modifying proteins in ND10 cannot be equated to
transcription or repression of transcription at that site
from their mere presence at this location.
Identi®cation of proteins that constitute the matrix
of ND10 on which other proteins may be deposited
should help to generate rational hypothesis for ND10
function. Analysis of several cell lines lacking certain
ND10-associated proteins, i.e., Sp100, Daxx and BLM
proved that they are not essential for the existence of
ND10. Only the absence of PML results in a di€use
distribution of all other ND10-associated proteins
although this does not prove that PML alone forms
the matrix onto which all the other proteins are
deposited. In addition, PML needs to be SUMOmodi®ed to recruit Daxx and other proteins (Ishov et
al., 1999). Several of these central ®ndings were
con®rmed in short order, such as the location of Daxx
at ND10 (Everett et al., 1999a; Li et al., 2000; Torii et
al., 1999; Zhong et al., 2000c), despite its reported
interaction with the death domain of the Fas receptor
(Yang et al., 1997; but see Torii et al., 1999) and the
need for SUMO-modi®cation of PML to bind Daxx
(Zhong et al., 2000a; Lehembre et al., 2001; Li et al.,
2000). Of the three reported sumo®cation sites
(Kamitani et al., 1998) only one (PML K160) seems
to be essential for the recruitment of other proteins to
ND10 (Lallemand-Breitenbach et al., 2001). The
functions of ND10 are then expected to be increased
by the interferon induced transcriptional upregulation
of PML, since such upregulation increases the number
and size of this nuclear domain (Maul et al., 1995; see
review by Chelbi-Alix in this issue, for details on
interferon induction of PML and Sp100). Interferon is
generally assumed to protect against viral infection.
Interferon can induce PML, Sp100, p28 and thus
recruit other ND10-associated proteins. These proteins
may be segregated/recruited into the nuclear depot,
until deployed/released upon viral attack. That PML is
functioning outside of ND10 in the repression of the
human foamy virus by binding to Tas has recently
been reported (Regad et al., 2001).
We recently found that a viral tegument protein
from human cytomegalovirus pp71 is sequestered at
ND10 through interaction with Daxx. Signi®cantly, the
molecular interaction site is situated in the middle of
the molecule instead of the C-terminal end that usually
interacts with PML (Fas receptor, Pax3 and the
centromeric CENP-C and several others). Daxx has
therefore a second interaction domain (Ishov and
Maul, submitted). Such an interaction mechanism
suggests that Daxx functions as an adapter molecule
in the recruitment of proteins to ND10. Whether such
recruitment and possible neutralization of a viral
transactivator has protective value, as suggested by
the interferon response of an increase in such function,
is not clear. In addition, such adapter properties of
Daxx predict that cellular proteins can be recruited, a
mechanism that is either usurped by the virus for its
own advantage or is part of the interferon-dependent
nuclear defense of ND10 (see also the review by
Everett). The ®rst such cell protein binding to the Nterminal section of Daxx is a homeodomain protein
FIST/HIPK3 that can inhibit FAS mediated JUN
NH2-terminal kinase activation. Upon overexpression
it is found associated with ND10 (Rochat-Steiner et al.,
2000). That there are free centromeric Daxx binding
sites in the presence of ND10 has most clearly been
shown by using PML7/7 cells where, speci®cally
towards the end of the cell cycle, Daxx binds at
highest concentration in the heterochromatin (Ishov et
al., 1999).
In addition to the de®ned PML-Daxx interaction,
another important interaction appears to take place
between the other prototypic ND10 protein, Sp100,
and the heterochromatin-binding protein HP1 (Lehming et al., 1998; Seeler et al., 1998). Since there is very
little HP1 at ND10 and a much larger amount
elsewhere in the nucleus, speci®cally the heterochromatin, such interactions at ND10 are dicult to
interpret (there is no recognizable DNA or heterochromatin in ND10 (Boisvert et al., 2000). Sp100 also
interacts with the HMG2 proteins (Lehming et al.,
1998), raising the possibility that HP1 or HMG2
mediate the interaction of Sp100 with DNA, acting as
a transcriptional repressor. In such a scenario, it may
be the availability of Sp100 that is controlled through
recruitment to ND10.
Sp100, like PML, can be covalently modi®ed by
SUMO, but does not require SUMO for the
colocalization with PML or HP1 binding (Sternsdorf
et al., 1997, 1999), although such modi®cation appears
to marginally increase the observable e€ect (Seeler et
al., 2001, and our unpublished results). The likelihood
that this is a turnover e€ect, is supported by the ®nding
that Sp100 increases substantially in ND10 when the
proteosomal pathway is blocked. Under such blocked
Evidence of nuclear depot function for ND10
D Negorev and GG Maul
conditions, PML and Sp100 appear to segregate within
individual ND10, such that both components occupy
their own space (Everett et al., 1999a, and our
unpublished results). Both PML and Sp100 are present
in that structure but their segregation within ND10
indicates that they do not form the matrix of ND10
dependent on proportional amounts of these two
proteins. In this context, it is interesting to note that
upon overexpression of Sp100, new Sp100 sites emerge
that have no association with PML. Thus the depot
capacity for Sp100 is quite limited (Negorev et al.,
2001), which is surprising as Sp100 can self-assemble.
This led to the hypothesis that an as yet unknown
adapter protein brings Sp100 to ND10 and that this
adapter binds to the self-assembly domain, or that a
change in three-dimensional structure, somewhat
ampli®ed by SUMO modi®cation, changes the binding
avidity (Negorev et al., 2001). In general, Sp100, like
Daxx, appears to be recruited to ND10 in order to
reduce its availability in the nuclear space rather than
acting like PML in the structural maintenance of
ND10. At present, there seem to be at least two
recruiting mechanisms working at ND10, one that is
regulated by the UBC9 SUMO-modi®cation of PML,
resulting in Daxx acquisition, and a second independent mechanism that recruits Sp100.
Externally induced release of proteins from ND10
The notion that ND10 represent a nuclear depot for
various proteins implies the regulated release of these
proteins as otherwise ND10 would represent only a
nuclear dump for the disposal of excess or detrimental
proteins (Maul, 1998). Such a nuclear dump function
could be postulated based on the induced ubiquitination of proteins by certain viruses (see Everett's review,
pp 7266 ± 7273), on the accumulation of immunoproteosomes during interferon upregulation (Fabunmi et
al., 2001) and presence of proteosomes after As2O3
exposure (Lallemand-Breitenbach et al., 2001). Since
the amount of proteins in ND10 is rather small and the
total protein content barely enough to register in
Western blot analysis, in situ visualization in the form
of immunohistochemistry has become an important
tool. Using such methodology, the ®rst release inducer
found was heat shock (Maul et al., 1995). Heat shock
leads to a sequential release of ND10-associated
proteins, with most of the PML content retained in
aggregated form. Such PML is desumo®ed, indicating
a reversal of the recruitment mechanism. Induction of
the release is rapid and occurs at temperatures
equivalent to high fever in humans. A SUMO
isopeptidase, sentrin-1 (SENP-1), has been identi®ed
(Gong et al., 2000) that removes ND10-associated
proteins but leaves most PML aggregated. This enzyme
appears to be speci®c, since it does not desumofy
RanGAP. The release of ND10-associated protein by a
desumo®cation mechanism represents the reversal of
the recruitment process. It also indicates that PML
does not require SUMO modi®cation to remain at
ND10. This has been demonstrated by inducible
expression of the PML-3K mutant (lacking all three
suggested SUMO-modi®cation sites) (Lallemand-Breitenbach et al., 2001). PML-PML interaction appears
sucient to retain the multimeric PML complex.
SUMO modi®cation of PML had been postulated to
be essential for PML deposition at ND10 based on the
®nding that the soluble nuclear fraction of PML is not
SUMO-modi®ed, while the insoluble fraction of PML
remains modi®ed (Muller et al., 1998). The question
when PML is SUMO modi®ed, in solution, so as to
bring Daxx to ND10, or at ND10 to act as a sink for
Daxx, is unsolved and quite interesting from a
mechanistic point of view. E€ects of speci®c PML site
modi®cation by SUMO-1 have recently been reported
which will change our interpretation on several fronts
(Lallemand-Breitenbach et al., 2001). Quite unexpectedly modi®cation of the lysine at aa160, as judged from
its elimination, has the e€ect of allowing proteosome
dependent turnover. As2O3 induced SUMO-1 modi®cation at aa160 resulted in the loss of PML which could
be blocked by lactacystin. The ND10 accumulation of
overexpressed PML lacking the SUMO-1 160 modi®cation site shows that such modi®cation is not
necessary for ND10 accumulation but it is essential
for the recruitment of other ND10-associated proteins.
Further PML accumulation at ND10 then seems to be
dependent on a dephosphorylation step consistent with
the ®nding that phosphorylation during mitosis leads
to PML dispersal (Everett et al., 1999a). The
observation that some PML aggregates after mitosis
in the cytoplasm but that no other ND10-associated
proteins are recruited (Maul and Everett, 1994) may be
a consequence of dephosphorylation without SUMO
modi®cation which has been shown to take place only
in the nucleus (Ishov et al., 1999).
Not all forms of stress appear to induce desumo®cation of PML and thus release of ND10-associated
proteins. When cells are exposed to Cd2+ all of the
ND10-associated proteins are released, including PML
(Nefkens et al., submitted). The release might be
associated simply with loss of cohesiveness of PML
molecules. Such release can be inhibited by two kinase
inhibitors, indicating the involvement of p38 MAPK
and the ERK1/2 pathways (Nefkins et al., submitted).
That both inhibitors prevented the Cd2+ induced
release independently suggesting that they need to
act on the same substrate as shown for Mnk1/2 after
TPA treatment (Fukunaga and Hunter, 1997; Waskiewicz et al., 1997). Most importantly Cd2+ induced
release of proteins from ND10 represents a second
regulated release mechanism. Whether other external
factors, such as UV light, have their own regulated
mechanism to partially or completely release accumulated proteins from ND10 needs to be investigated. In
the two examples cited, released proteins are not
necessarily hydrolyzed but are found at many other
nuclear sites. It is possible that this distribution
re¯ects or even causes the substantial repression of
transcription and splicing after those external insults
(Bond and James, 2000). In this context, the
7237
Oncogene
Evidence of nuclear depot function for ND10
D Negorev and GG Maul
7238
Table 1
Proteins present at ND10 at endogenous levels of expression
PML
essential in ND10 formation; in SUMO-modified form recruits Daxx; interacts with CBP, pRb, p53; considered a
co-repressor;1 ± 4.
Sp100
interacts with HP1; ND10 are limited in their capacity to recruit Sp1005,6.
Daxx
recruited by PMLSUMO, 1; binds to an increasing number of proteins.
BLM
low relative amount of BLM is stationed in ND101,7.
SUMO
modifies PML and Sp100 covalently changing their capacity to bind other proteins1,8 ± 10.
NDP55
uncharacterized protein(s); potentially a protein modification recognized by Mab 138; increases after heat shock; present in all
vertebrate species tested11,12.
|‚‚‚‚‚‚‚‚‚‚‚‚‚‚‚{z‚‚‚‚‚‚‚‚‚‚‚‚‚‚‚}
Proteins conditionally
TRF1/2
NBS1
P95/nibrin
Mre11
RAD51
RAD52
Repl. factor A
WRN
SENP-1
HP1
Sp100C
HMG2
CBP
pRb
USP7
PA28
proteosome
RFP
GAPDH
eIF-4
PLZF
HSF2
Tax
p53
ISG-20
Mutant ataxin
PKM
Int6
PAX3
Sp1
FIST/HIPK3
hGCNP
Sp140
BRCA1
present in or at ND10
present in a subset of ND10 containing telomere repeats in cells depending on an alternative telomere maintenance
mechanism. These proteins colocalize with PML in telomerase-negative immortalized cells during the late S/G2-phase
of the cell cycle39 ± 42
removes SUMO from PML14.
interacts with Sp10 but is present in ND10 in variable amounts depending on the cell cycle phase5,6,15.
present in some ND1016.
interacts with Sp100B6.
interacts with PML17; found in some cell types but not others18.
phosphorylated form found by some groups and not by others; possibly cell cycle related colocalization43
at or beside some ND10; probably cell cycle dependent19,20.
proteosomal activator; increased recruitment to ND10 by interferon g exposure21.
only after application of proteosome inhibitors22 or after interferon g as immunoproteosomes21.
interacts with PML and associates with a subset of ND1023.
interaction with PML is RNA dependent24.
interaction with PML25.
no interaction with PML but side by side localization in nucleus26.
heat shock transcription factor 2 after overexpression in low number of cells, mostly cytoplasmic27.
recruited to ND10 when PML is overexpressed28.
in specific cells, after overexpression or proteosome inhibition4 or beside ND10 with T-ag down regulation29.
after overexpression of a fragment30.
PML but not the other ND10 proteins are located in ataxin-1 aggregates after polyglutamine tract extension31,32.
Mx interaction kinase after transient expression33.
34
in some cells possibly by binding to overexpressed Daxx35.
interacts with PML36.
after overexpression interacts with Daxx37
besides ND10 upon transient overexpression38.
after overexpression with some ND1044.
breast cancer protein 1 beside ND10 upon transient overexpression20.
1
Ishov et al., 1999; 2Mu et al., 1994; 3Alcalay et al., 1998; 4Guo et al., 2000; 5Seeler et al., 1998; 6Lehming et al., 1998; 7Yankiwski et al., 2000;
Boddy et al., 1996; 9Sternsdorf et al., 1997; 10Kamitani et al., 1998b; 11Ascoli and Maul, 1991; 12Maul et al., 1995; 13Gong et al., 1997; 14Gong
et al., 2000; 15Everett et al., 1999b; 16Seeler et al., 2001; 17Doucas et al., 1999; 18Boisvert et al., 2001; 19Everett et al., 1999b; 20Maul et al., 1998;
21
Fabunmi et al., 2001; 22Anton et al., 1999; 23Cao et al., 1998; 24Carlile et al., 1998; 25Lai and Borden, 2000; 26Ruthardt et al., 1998; 27Goodson
et al., 2001; 28Doucas and Evans, 1999; 29Jiang et al., 1996; 30Gongora et al., 1997; 31Skinner et al., 1997; 32Klement et al., 1998; 33Trost et al.,
2000; 34Desbois et al., 1996; 35Lehembre et al., 2001; 36Vallian et al., 1998; 37Rochat-Steiner et al., 2000; 38Maul, 1998; 39Yeager et al., 1999;
40
Wu et al., 2000; 41Lombard and Guarente, 2000; 42Johnson et al., 2001; 43Alcalay et al., 1998; 44Bloch et al., 1999
8
interaction of histone deacetylase with PML (Wu et
al., 2001) may result in deacetylation and silencing of
many genes to which PML is released after heat shock
or Cd2+ exposure (Maul et al., 1995). A third
mechanism of ND10-associated protein release can
be postulated after As2O3 exposure in which PML is
SUMO-1 modi®ed and then hydrolyzed through a
proteosome dependent pathway (Lallemand-Breitenbach et al., 2001). Those proteins bound to the larger
ND10-based PML isoforms should be released. Since
the common moiety of PML is not e€ected we may
see the retention of PML bodies. Thus, various
protein release mechanisms seem to exist responding
Oncogene
to di€erent external insults, like stress, poisons or viral
infections. They may provide the experimental settings
to test for physiological e€ects of protein release from
ND10.
Potential function of ND10 as a regulated depot
A schematic outline of the information leading to the
nuclear depot hypothesis is presented in Figure 1 for
the PML-Daxx-pp71 recruitment and release mechanism including the physical inducers and enzymatic
e€ectors of the depot functions.
Evidence of nuclear depot function for ND10
D Negorev and GG Maul
In our scheme, the four dimensional interactivity of
available or sequestered proteins might represent a
mechanism to extend the range within which the cell
can function without the need for apoptotic selfdestruction. In such an interpretation, involvement of
PML or Daxx in apoptosis or simply cell death is only
by default. An elaborate enzymatic and signaling
system appear to ensure release of ND10-associated
proteins to which protective functions may be assigned.
Whenever ND10 cannot maintain the nuclear balance
as evidenced in excessive heat shock or heavy metal
exposure, the cell succumbs when no further proteins
can be released.
In a previous review (Seeler and Dejean, 1999), the
authors concluded that the presence of various proteins
involved `in transcription, chromatin structure, cellular
growth control and apoptosis, coupled with their
SUMO-regulated dynamic structure makes it clear that
a simple, single function for NBs is unlikely to be found'.
This is true if one considers any of the resident proteins
functions as indicative of the functions of ND10.
However, if ND10 regulates the balance of available
protein for sites other than ND10, a uni®ed function for
ND10 becomes apparent, independent of the function of
individual proteins in the multitude of de®ned cellular
synthetic activities or metabolic responses. The function
of ND10 then is that of a nuclear depot, speci®cally the
regulated recruitment and release of certain proteins that
change the nuclear synthetic activities in response to
external stimuli. The various stress responses are a good
example, since they result in release that is di€erentially
induced by regulatory loops such as the MAPK pathway
or by activation of an isopeptidase that changes the
covalent modi®cation of the matrix-forming PML.
Evidence that such recruitment and release has physiological consequences is emerging from studies in which
individual proteins are recruited into ND10, e.g., Daxx
removal from Pax3 (Lehembre et al., 2001), HDAC (Li et
al., 2000) or Tax (Doucas and Evans, 1999). The release
of PA28 after heat shock may activate speci®c proteosomes (Fabunmi et al., 2001). The potential to increase
the capacity of the nuclear depot system by interferon
will provide new leads in how cell protection may be
enhanced. The necessity for viruses to circumvent any
interferon-induced nuclear defenses concentrated at
ND10 for deployment upon viral attack will lead to
further de®nition of such nuclear depot functions.
Recognition of faulty or foreign DNA/protein complexes (Bell et al., 2000; Tang et al., 2000; Tsukamoto et
al., 2000) may play a substantial part in the functional
signi®cance of ND10. Overall, the high diversity of
ND10-associated proteins and their potential e€ects
after release or recruitment should not detract from the
simplicity inherent in the concept of ND10 as a regulated
nuclear depot system.
7239
Proteins associated with ND10
The list of proteins associated with ND10 is growing.
Recognition of such associations often comes from the
nuclear distribution similarities of the now familiar
dot-like images. Also ®nding physical interactions with
established ND10 proteins such as PML or Sp100 are
Figure 1
Oncogene
Evidence of nuclear depot function for ND10
D Negorev and GG Maul
7240
Oncogene
strong inducers to ®nd that such interacting proteins
are localized at that speci®c site. However, such
proteins may not necessarily be segregated in ND10
as is obvious for pRB or might be so segregated only
under certain conditions as the whole system of
proteins involved in alternative telomere maintenance
illustrates. Table 1 lists those proteins that have been
identi®ed at ND10 by immunohistochemistry, since
this is still the only useful technique until we develop a
method to isolate ND10. The proteins are grouped
according to their permanent, transitory or even
questionable association with ND10. The latter is
particularly applicable for those proteins that are
found only after transient overexpression in or beside
ND10. The potential diculty with interpretation is
best illustrated by the di€erent locations of cytomegalovirus immediate early 2 protein (IE2). If transient
expressed IE2 perfectly colocalizes with ND10; when
expressed in the context of infection, IE2 localizes
perfectly beside ND10, where it accumulates transcription factors (Ishov et al., 1997). Other than the crude
grouping in Table 1 no evaluation of the reproducibility or the physiological signi®cance of ND10association has been attempted.
Occasionally, proteins that can associate with ND10
are discovered by a similarity of distribution or shape
of PML distribution in a few large ND10. Such a case
is all the more remarkable when as few as 5% of the
cells show this distribution. In the case of TRF1 and
WRN (Johnson et al., 2001; Yeager et al., 1999) the
low 5% of cells showing TRF1/ND10 association
suggested that it was the G2 phase of the cell cycle
where it was found. Proteins associated with the TRF1,
such as NBS1 and Mre11 were also found at ND10
(Lombard and Guarente, 2000; Wu et al., 2000). As
these proteins belong to the telomeric complex in cells
with the alternate lengthening of telomeres (ALT)
system, such an association was suspected, if the
complex is either retained as such or functions at that
site. In situ hybridization with probes to detect
telomeres showed that such large aggregates containing
PML enclosed telomere repeats (Tokutake et al., 1998;
Yeager et al., 1999). BrdU incorporation was found in
or beside ND10 at the S/G2 interface, suggesting that
DNA synthesis takes place at ND10 in telomerasenegative cells and is involved in alternative telomere
length maintenance. The association of PML and
TRF1 occurs only in ALT-immortalized cells or in
5% of tumors (Bryan and Reddel, 1997; Colgin and
Reddel, 1999) that had lost telomerase activity and not
in normal or otherwise immortalized cells. Thus, it is a
speci®c case that might have come to light because of
the repetitive nature of the telomere DNA sequences
and thus an ampli®cation of protein binding sites to a
level recognizable with this technique. These ®ndings
suggest that DNA may be present in a select set of
ND10 under certain not quite normal circumstances
like ALT.
Proteins enriched in ND10 may be involved in some
form of DNA synthesis, metabolism or repair, or in the
recognition of speci®c chromosomal sites. The Bloom
syndrome RecQ helicase, BLM, is a case in point
(Ishov et al., 1999; Yankiwski et al., 2000), although
we do not expect this protein to function in ND10. On
the other hand, the increased ®nding of ND10 in late
G2-phase close to or overlapping the centromeric
region (Everett et al., 1999a) shares with telomeres in
ALT the presence of repetitive DNA and late
replication. More examples of such sightings will help
to de®ne functions related to the accumulation of
ND10-associated proteins. The case of ALT associated
proteins with ND10 in some cell types and only in a
limited part of the cell population, may also suggest an
explanation for the lack of a generally observable
presence of pRb in ND10. This protein has been found
at ND10 by one group as an interaction partner with
PML and at some ND10 (Alcalay et al., 1998) but is
not universally recognized at these sites by others. It
may be that the presence of pRB in ND10 is dependent
on a speci®c part of the cell cycle or on speci®c cell
types with lesions in telomere maintenance.
An in¯uential paper by LaMorte et al. (1998) ®rst
suggested the presence of transcriptionally active DNA
in ND10. Using microinjection of ¯uoresceinated
uridine triphosphates, they observed colocalization of
new transcripts with a subset of ND10. A number of
controls suggested a very high rate of polymerase II
dependent transcription, although an aggregation of
new transcripts cannot be excluded. Other studies
could ®nd no transcription at ND10 using transcription run-on analysis (Grande et al., 1996; Ishov and
Maul, 1996; Jackson et al., 1993; Wansink et al., 1993;
Weis et al., 1994), although neither the rate of
transcription or the possible aggregation of transcripts
from elsewhere would be recognized with this run-on
technique. The apparent massive accumulation of
RNA in PML co-staining nuclear bodies in the report
by LaMorte et al. (1998), however should be
considered with caution as the regressive staining
procedure (Bernhard, 1969) is selective but not a
speci®c stain for RNA. Certain proteins, when highly
packed and as found in dense bodies may also result in
increased uranyl acid staining. In addition no RNA or
DNA was found in ND10 using electron spectroscopic
imaging (Boisvert et al., 2000).
The second signi®cant subject introduced by
LaMorte et al. (1998) was that high concentrations
of the CREB binding protein CBP, a histone
acetylase, were localized in all ND10 of HEp-2 cells,
supporting the notion that these nuclear domains are
regulated transcription sites. PML physically interacts
with the histone acetylase CBP (Doucas et al., 1999)
and the histone deacetylase (HDAC) (Wu et al.,
2001), where CBP is found colocalizing in ND10 in
some cell types (Boisvert et al., 2001; LaMorte et al.,
1998; von Mikecz et al., 2000) and HDAC is not (Li
et al., 2000). Since a large amount of DNA justifying
the high concentration of CBP cannot be found in
ND10 (Boisvert et al., 2000), this apparent symmetry
could be taken as segregation of CBP so as to keep it
from activating additional transcription units and
HDAC ± PML maintaining silencing out there in the
Evidence of nuclear depot function for ND10
D Negorev and GG Maul
nucleoplasm. In both cases PML would be involved in
transcriptional modi®cation but only in the case of
CBP would the interaction with PML lead to
localization at ND10 where it might remain to be
released at the appropriate signals for transcriptional
activation somewhere in the nucleus. The PML
binding of two proteins that can have opposite
modifying e€ects on histones is remarkable. PML
binds to the N-terminal domain of HDAC which
contains its catalytic site and as such might be
responsible for the PML-mediated repression of
transcription (Wu et al., 2001). Essential is that such
interaction occurs only with PML IV whereas the
short isoform PML VI did not interact. The control
of such interactions may de®ne certain transcriptional
regulation e€ected by externally applied conditions.
Those conditions that result in recruitment and release
of proteins from ND10 would be the most useful in
attempting the experimental veri®cation of the nuclear
depot concept.
The concept of a nuclear depot system provides a
rational basis for the ®ndings that certain proteins are
only temporarily present in ND10 or only in speci®c
cell types. Upon the requisite balance in certain cells no
di€erential segregation may be needed so even the
absence of observable ND10 in nerve or germ cells
should not surprise. Even the life of a mouse without
ND10 (PML knock-out) may not su€er if removed
from stresses like making a living under the gaze of a
cat, heat exposure from sun drenched rocks, heavy
metal contaminated and overcrowded underground
burrows and viral infections from fellow members of
the species. Tests of such conditions will be possible
with the availability of PML7/7 mice.
7241
References
Alcalay M, Tomassoni L, Colombo E, Stoldt S, Grignani F,
Fagioli M, Szekely L, Helin K and Pelicci PG. (1998). Mol.
Cell. Biol., 18, 1084 ± 1093.
Anton LC, Schubert U, Bacik I, Princiotta MF, Wearsch PA,
Gibbs J, Day PM, Realini C, Rechsteiner MC, Bennink JR
and Yewdell JW. (1999). J. Cell. Biol., 146, 113 ± 124.
Ascoli CA and Maul GG. (1991). J. Cell. Biol., 112, 785 ±
795.
Bell P, Brazas R, Ganem D and Maul GG. (2000). J. Virol.,
74, 5329 ± 5336.
Bell P, Montaner LJ and Maul GG. (2001). J. Virol., in press.
Bernhard W. (1969). J. Ultrastruct. Res., 27, 250 ± 265.
Bloch DB, Chiche JD, Orth D, de la Monte SM, Rosenzweig
A and Bloch KD. (1999). Mol. Cell. Biol., 19, 4423 ± 4430.
Boddy MN, Howe K, Etkin LD, Solomon E and Freemon
PS. (1996). Oncogene, 13, 971 ± 982.
Boisvert FM, Hendzel MJ and Bazett-Jones DP. (2000).
J. Cell. Biol., 148, 283 ± 292.
Boisvert FM, Kruhlak MJ, Box AK, Hendzel MJ and
Bazett-Jones DP. (2001). J. Cell. Biol., 152, 1099 ± 1106.
Bond U and James TC. (2000). Int. J. Biochem. Cell. Biol.,
32, 643 ± 656.
Brasch K and Ochs RL. (1992). Exp. Cell. Res., 202, 211 ±
223.
Bregman DB, Du L, van der Zee S and Warren SL. (1995).
J. Cell. Biol., 129, 287 ± 298.
Bryan TM and Reddel RR. (1997). Eur. J. Cancer, 33, 767 ±
773.
Cao T, Duprez E, Borden KL, Freemont PS and Etkin LD.
(1998). J. Cell. Sci., 111, 1319 ± 1329.
Carlile GW, Tatton WG and Borden KL. (1998). Biochem.
J., 335, 691 ± 696.
Cho Y, Lee I, Maul GG and Yu E. (1998). Int. J. Mol. Med.,
1, 717 ± 724.
Colgin LM and Reddel RR. (1999). Curr. Opin. Genet. Dev.,
9, 97 ± 103.
Daniel MT, Koken M, Romagne O, Barbey S, Bazarbachi A,
Stadler M, Guillemin MC, Degos L, Chomienne C and de
The H. (1993). Blood, 82, 1858 ± 1867.
de The G, Riviere M and Bernhard W. (1960). Bull. Cancer,
47, 570 ± 584.
Desbois C, Rousset R, Bantignies F and Jalinot P. (1996).
Science, 273, 951 ± 953.
Doucas V and Evans RM. (1999). Proc. Natl. Acad. Sci.
USA, 96, 2633 ± 2638.
Doucas V, Tini M, Egan DA and Evans RM. (1999). Proc.
Natl. Acad. Sci. USA, 96, 2627 ± 2632.
Everett RD, Earnshaw WC, Pluta AF, Sternsdorf T,
Ainsztein AM, Carmena M, Ruchaud S, Hsu W and Orr
A. (1999a). J. Cell. Sci., 112, 3443 ± 3454.
Everett RD, Meredith M and Orr A. (1999b). J. Virol., 73,
417 ± 426.
Fabunmi RP, Wigley WC, Thomas PJ and DeMartino GN.
(2001). J. Cell. Sci., 114, 29 ± 36.
Flenghi L, Fagioli M, Tomassoni L, Pileri S, Gambacorta M,
Pacini R, Grignani F, Casini T, Ferrucci PF, Martelli MF.
et al. (1995). Blood, 85, 1871 ± 1880.
Fogal V, Gostissa M, Sandy P, Zacchi P, Sternsdorf T,
Jensen K, Pandol® PP, Will H, Schneider C and Del Sal G.
(2000). EMBO J., 19, 6185 ± 6195.
Fukunaga R and Hunter T. (1997). EMBO J., 16, 1921 ±
1933.
Gambacorta M, Flenghi L, Fagioli M, Pileri S, Leoncini L,
Bigerna B, Pacini R, Tanci LN, Pasqualucci L, Ascani S,
Mencarelli A, Liso A, Pelicci PG and Falini B. (1996).
Am. J. Pathol., 149, 2023 ± 2035.
Gong L, Kamitani T, Fujise K, Caskey LS and Yeh ET.
(1997). J. Biol. Chem., 272, 28198 ± 28201.
Gong L, Millas S, Maul GG and Yeh ET. (2000). J. Biol.
Chem., 275, 3355 ± 3359.
Gongora C, David G, Pintard L, Tissot C, Hua TD, Dejean
A and Mechti N. (1997). J. Biol. Chem., 272, 19457 ±
19463.
Goodson ML, Hong Y, Rogers R, Matunis MJ, Park-Sarge
OK and Sarge KD. (2001). J. Biol. Chem., 15, 15.
Grande MA, van der Kraan I, van Steensel B, Schul W, de
The H, van der Voort HT, de Jong L and van Driel R.
(1996). J. Cell. Biochem., 63, 280 ± 291.
Guo A, Salomoni P, Luo J, Shih A, Zhong S, Gu W and
Pandol® PP. (2000). Nature Cell. Biol., 2, 730 ± 736.
Hinglais-Guillaud NMR and Bernhard W. (1961). Bull.
Cancer, 48, 282 ± 316.
Hsu WL and Everett RD. (2001). J. Virol., 75, 3819 ± 3831.
Ishov AM and Maul GG. (1996). J. Cell. Biol., 134, 815 ±
826.
Oncogene
Evidence of nuclear depot function for ND10
D Negorev and GG Maul
7242
Oncogene
Ishov AM, Sotnikov AG, Negorev D, Vladimirova OV, Ne€
N, Kamitani T, Yeh ET, Strauss III JF and Maul GG.
(1999). J. Cell. Biol., 147, 221 ± 234.
Ishov AM, Stenberg RM and Maul GG. (1997). J. Cell. Biol.,
138, 5 ± 16.
Jackson DA, Hassan AB, Errington RJ and Cook PR.
(1993). EMBO J., 12, 1059 ± 1065.
Jiang WQ, Szekely L, Klein G and Ringertz N. (1996). Exp.
Cell. Res., 229, 289 ± 300.
Johnson FB, Marciniak RA, McVey M, Stewart SA, Hahn
WC and Guarente L. (2001). EMBO J., 20, 905 ± 913.
Kamitani T, Kito K, Nguyen HP, Wada H, FukudaKamitani T and Yeh ET. (1998a). J. Biol. Chem., 273,
26675 ± 26682.
Kamitani T, Nguyen HP, Kito K, Fukuda-Kamitani T and
Yeh ET. (1998b). J. Biol. Chem., 273, 3117 ± 3120.
Kass-Eisler A and Greider CW. (2000). Trends Biochem. Sci.,
25, 200 ± 204.
Klement IA, Skinner PJ, Kaytor MD, Yi H, Hersch SM,
Clark HB, Zoghbi HY and Orr HT. (1998). Cell., 95, 41 ±
53.
Koken MH, Linares-Cruz G, Quignon F, Viron A, ChelbiAlix MK, Sobczak-Thepot J, Juhlin L, Degos L, Calvo F
and de The H. (1995). Oncogene, 10, 1315 ± 1324.
Lai HK and Borden KL. (2000). Oncogene, 19, 1623 ± 1634.
Lallemand-Breitenbach V, Zhu J, Puvion F, Koken M,
Honore N, Doubrekovsky A, Duprez E, Pandol® PP,
Puvion E, Freemont P and de The H. (2001). J. Exp. Med.,
193, 1361 ± 1372.
Lam YW, Ammerlaan WOWS, Kroese F and Opstelten D.
(1995). Exp. Cell Res., 221, 344 ± 356.
LaMorte VJ, Dyck JA, Ochs RL and Evans RM. (1998).
Proc. Natl. Acad. Sci. USA, 95, 4991 ± 4996.
Lehembre F, Muller S, Pandol® PP and Dejean A. (2001).
Oncogene, 20, 1 ± 9.
Lehming N, Le Saux A, Schuller J and Ptashne M. (1998).
Proc. Natl. Acad. Sci. USA, 95, 7322 ± 7326.
Li H, Leo C, Zhu J, Wu X, O'Neil J, Park EJ and Chen JD.
(2000). Mol. Cell. Biol., 20, 1784 ± 1796.
Lombard DB and Guarente L. (2000). Cancer Res., 60,
2331 ± 2334.
Maul GG and Everett RD. (1994). J. Gen. Virol., 75, 1223 ±
1233.
Maul GG. (1998). Bioessays, 20, 660 ± 667.
Maul GG, Jensen DE, Ishov AM, Herlyn M and Rauscher
III FJ. (1998). Cell Growth Di€er., 9, 743 ± 755.
Maul GG, Yu E, Ishov AM and Epstein AL. (1995). J. Cell.
Biochem., 59, 498 ± 513.
Mu ZM, Chin KV, Liu JH, Lozano G and Chang KS. (1994).
Mol. Cell. Biol., 14, 6858 ± 6867.
Muller S, Matunis MJ and Dejean A. (1998). EMBO J., 17,
61 ± 70.
Negorev D, Ishov AM and Maul GG. (2001). J. Cell. Sci.,
114, 59 ± 68.
Ochs RL, Stein Jr TW, Andrade LE, Gallo D, Chan EK, Tan
EM and Brasch K. (1995). Mol. Biol. Cell., 6, 345 ± 356.
Padykula HA, Fitzgerald M, Clark JH and Hardin JW.
(1981). Anal. Rec., 201, 679 ± 696.
Plehn-Dujowich D, Bell P, Ishov AM, Baumann C and Maul
GG. (2000). Chromosoma., 109, 266 ± 279.
Puvion-Dutilleul F, Chelbi-Alix MK, Koken M, Quignon F,
Puvion E and de The H. (1995). Exp. Cell. Res., 218, 9 ± 16.
Regad T, Lallemand-Breitenbach V, Pandol® PP, de The H,
Mounira K and Chelbi-Alix MK. (2001). EMBO J., in
press.
Rochat-Steiner V, Becker K, Micheau O, Schneider P, Burns
K and Tschopp J. (2000). J. Exp. Med., 192, 1165 ± 1174.
Ruthardt M, Orleth A, Tomassoni L, Puccetti E, Riganelli
D, Alcalay M, Mannucci R, Nicoletti I, Grignani F,
Fagioli M and Pelicci PG. (1998). Oncogene, 16, 1945 ±
1953.
Seeler JS and Dejean A. (1999). Curr. Opin. Genet. Dev., 9,
362 ± 367.
Seeler JS, Marchio A, Losson R, Desterro JM, Hay RT,
Chambon P and Dejean A. (2001). Mol. Cell. Biol., 21,
3314 ± 3324.
Seeler JS, Marchio A, Sitterlin D, Transy C and Dejean A.
(1998). Proc. Natl. Acad. Sci. USA, 95, 7316 ± 7321.
Skinner PJ, Koshy BT, Cummings CJ, Klement IA, Helin K,
Servadio A, Zoghbi HY and Orr HT. (1997). Nature, 389,
971 ± 974.
Sternsdorf T, Jensen K, Reich B and Will H. (1999). J. Biol.
Chem., 274, 12555 ± 12566.
Sternsdorf T, Jensen K and Will H. (1997). J. Cell. Biol., 139,
1621 ± 1634.
Tang Q, Bell P, Tegtmeyer P and Maul GG. (2000). J. Virol.,
74, 9694 ± 9700.
Tokutake Y, Matsumoto T, Watanabe T, Maeda S, Tahara
H, Sakamoto S, Niida H, Sugimoto M, Ide T and Furuichi
Y. (1998). Biochem. Biophys. Res. Commun., 247, 765 ±
772.
Torii S, Egan DA, Evans RA and Reed JC. (1999). EMBO J.,
18, 6037 ± 6049.
Trost M, Kochs G and Haller O. (2000). J. Biol. Chem., 275,
7373 ± 7377.
Tsukamoto T, Hashiguchi N, Janicki SM, Tumbar T,
Belmont AS and Spector DL. (2000). Nat. Cell. Biol., 2,
871 ± 878.
Vallian S, Chin KV and Chang KS. (1998). Mol. Cell. Biol.,
18, 7147 ± 7156.
von Mikecz A, Zhang S, Montminy M, Tan EM and
Hemmerich P. (2000). J. Cell. Biol., 150, 265 ± 273.
Wansink DG, Schul W, van der Kraan I, van Steensel B, van
Driel R and de Jong L. (1993). J. Cell. Biol., 122, 283 ± 293.
Waskiewicz AJ, Flynn A, Proud CG and Cooper JA. (1997).
EMBO J., 16, 1909 ± 1920.
Weis K, Rambaud S, Lavau C, Jansen J, Carvalho T,
Carmo-Fonseca M, Lamond A and Dejean A. (1994).
Cell, 76, 345 ± 356.
Wu G, Lee WH and Chen PL. (2000). J. Biol. Chem., 275,
30618 ± 30622.
Wu WS, Vallian S, Seto E, Yang WM, Edmondson D, Roth
S and Chang KS. (2001). Mol. Cell. Biol., 21, 2259 ± 2268.
Yang X, Khosravi-Far R, Chang HY and Baltimore D.
(1997). Cell, 89, 1067 ± 1076.
Yankiwski V, Marciniak RA, Guarente L and Ne€ NF.
(2000). Proc. Natl. Acad. Sci. USA, 97, 5214 ± 5219.
Yeager TR, Neumann AA, Englezou A, Huschtscha LI,
Noble JR and Reddel RR. (1999). Cancer Res., 59, 4175 ±
4179.
Zhong S, Muller S, Ronchetti S, Freemont PS, Dejean A and
Pandol® PP. (2000a). Blood, 95, 2748 ± 2752.
Zhong S, Salomoni P and Pandol® PP. (2000b). Nat. Cell.
Biol., 2, E85 ± E90.
Zhong S, Salomoni P, Ronchetti S, Guo A, Ruggero D and
Pandol® PP. (2000c). J. Exp. Med., 191, 631 ± 640.