Download Self-association of the SET domains of human ALL-1 and of

Oncogene (2000) 19, 351 ± 357
ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00
www.nature.com/onc
Self-association of the SET domains of human ALL-1 and of Drosophila
TRITHORAX and ASH1 proteins
T Rozovskaia1, O Rozenblatt-Rosen1, Y Sedkov2, D Burakov1, T Yano2, T Nakamura2,
S Petruck2, L Ben-Simchon1, CM Croce2, A Mazo2 and E Canaani*,1
1
Department of Molecular Cell Biology, Weizmann Institute of Science, Rehovot 76100, Israel; 2Kimmel Cancer Institute, Je€erson
Medical College, Philadelphia, Pennsylvania, PA 19107, USA
The human ALL-1 gene is involved in acute leukemia
through gene fusions, partial tandem duplications or a
speci®c deletion. Several sequence motifs within the ALL1 protein, such as the SET domain, PHD ®ngers and the
region with homology to DNA methyl transferase are
shared with other proteins involved in transcription
regulation through chromatin alterations. However, the
function of these motifs is still not clear. Studying ALL-1
presents an additional challenge because the gene is the
human homologue of Drosophila trithorax. The latter is a
member of the trithorax-Polycomb gene family which acts
to determine the body pattern of Drosophila by
maintaining expression or repression of the Antennapedia-bithorax homeotic gene complex. Here we apply yeast
two hybrid methodology, in vivo immunoprecipitation and
in vitro `pull down' techniques to show self association of
the SET motifs of ALL-1, TRITHORAX and ASH1
proteins (Drosophila ASH1 is encoded by a trithoraxgroup gene). Point mutations in evolutionary conserved
residues of TRITHORAX SET, abolish the interaction.
SET ± SET interactions might act in integrating the
activity of ALL-1 (TRX and ASH1) protein molecules,
simultaneously positioned at di€erent maintenance elements and directing expression of the same or di€erent
target genes. Oncogene (2000) 19, 351 ± 357.
Keywords: SET domain; oligomerization; chromatin
alterations
Introduction
The ALL-1 gene (also termed HRX and MLL) was
cloned by virtue of its involvement in 11q23
chromosome translocations (Gu et al., 1992; Tkachuk
et al., 1992) associated with acute leukemias, in
particular infant and secondary leukemias (Raimondi,
1993; Heim and Mitelman, 1995). The ALL-1 gene is
composed of 38 exons (Rasio et al., 1996) spanning
90 kb and encoding a protein of 3969 residues.
ALL-1 is the human homologue of Drosophila
trithorax. That gene is a member of the trithoraxPolycomb gene family acting during embryogenesis to
maintain the expression pattern of the homeotic gene
complexes Antennapedia and bithorax which determine
the body pattern along the anterior-posterior axis
*Correspondence: E Canaani
Received 27 July 1999; revised 6 October 1999; accepted 14 October
1999
(Lewis, 1978). The pattern is initially established by
the gap and pair-rule genes and is later maintained by
the trithorax and polycomb group genes, which
function as transcriptional activators and repressors,
respectively. ALL-1 contains several motifs shared with
other proteins. These include: (a) three AT hook motifs
which are known to bind to AT-rich sequences at the
minor groove of DNA, and consequently bend the
DNA to stabilize DNA-protein associations (Thanos
and Maniatis, 1995); (b) a Cysteine-rich segment which
shares homology with mammalian DNA methyl
transferases (MTase) as well as with the transcription
repressor MeCP1 (Cross et al., 1997). This ALL-1
element shows a transcriptional repression activity in
heterologous assays (Prasad et al., 1995); (c) a domain
of 55 residues identi®ed as a strong transcriptional
activator, containing a core composed of hydrophobic
residues and aspartic acid residues (ibid).
ALL-1 shares homology with its Drosophila homologue TRITHORAX (TRX) (Mazo et al., 1990) in two
major and at least two minor domains. The latter were
implicated in the distribution of the ALL-1 protein in
nuclear speckles (Yano et al., 1997). In their central
portion, the two proteins contain three zinc-®nger-like
motifs with a unique Cys4 ± His ± Cys3 pattern, which
were termed PHD ®ngers. A related imperfect PHD
®nger is also shared by the two proteins. PHD ®ngers
were identi®ed in dozens of proteins, are present in all
eukaryotes, and are thought to be involved in
chromatin-mediated transcriptional regulation. A second major motif found in ALL-1 and TRX is the
*130 aa terminal SET domain. This highly conserved
sequence is present in proteins implicated in transcriptional activation as well as in proteins associated with
silencing (Jenuwein et al., 1998). A substantial e€ort
has been recently invested by us (Rozenblatt-Rosen et
al., 1998) and others (Cui et al., 1998; Cardoso et al.,
1998) in attempts to ®nd protein(s) interacting with
metazoan SET domains. In parallel, studies in yeast
demonstrated that the S. cerevisiae SET 1 protein
interacts through the SET motif with the checkpoint
MEC3 protein involved in DNA repair and telomere
function (Corda et al., 1999). Previously, the SET
domain within yeast SET 1 was shown to play a direct
role in telomeric silencing (Nislow et al., 1997), and the
SET protein CLR4 of S. pombe was shown to be a
modi®er of centromeric position e€ects (silencing)
(Ekwall et al., 1996).
Applying yeast two hybrid methodology as well as
other techniques, we have shown that the SET domains
of ALL-1 and its Drosophila homologue TRX interact
with human INI1 and its Drosophila homologue SNR1,
Self-association of SET
T Rozovskaia et al
352
respectively (Rozenblatt-Rosen et al., 1998). SNR1 is
encoded by a trithorax-group gene. Both SNR1 and
INI1 are components of the SWI/SNF multiprotein
complex involved in chromatin remodeling. Applying
similar approaches, others have shown that ALL-1
SET interacts with the dual-speci®city phosphatase
myelotubularin and its homologue Sbf1 which lacks a
catalytic domain (Cui et al., 1998). Others yet (Cardoso
et al., 1998), have found physical interaction between
the SET domain of the mammalian protein EZH2 and
the XNP/ATR-X protein which is structurally related
to the SNF2/SWI DNA helicase involved in chromatin
remodeling.
In the present study we provide evidence for another
interaction of the ALL-1 SET domain ± self association. The same interaction is shown here for the SET
domains of TRX and of ASH1 proteins, the latter
encoded by the trithorax-group gene ash1 (Tripoulas et
al., 1996) which, like trithorax, acts as an activator.
The ASH1 protein is associated with chromatin at
target sites, many of which are shared with target sites
of TRX (Rozovskaia et al., 1999).
Results and discussion
The C-terminal region of ALL-1 and TRX includes the
SET domain, which is highly conserved between the
two proteins and is also present within the trx-G/Pc-G
proteins ASH1 (Tripoulas et al., 1996) and E(Z) (Jones
and Gelbart, 1993), in the functionally related
Drosophila protein SU(VAR) 3 ± 9 (Tschiersch et al.,
1994), in the ALL-1 related protein ALR (Prasad et al.,
1997), and in a series of other proteins from yeast,
fungi, C. elegans, plants and humans (Jenuwein et al.,
1998; Nislow et al., 1997; Prasad et al., 1997). A TRX
C-terminal segment containing amino acid residues
3375 ± 3759, which span the SET motif and some
upstream sequences, was inserted into the pGBT9
vector and used in a two hybrid assay to screen a
Drosophila third instar larvae cDNA library in pACT.
Three positive clones were identi®ed from 26106
transformants. The interaction of one of these clones
could be con®rmed by other methodologies and
therefore was selected for additional characterization.
This clone contained a TRX segment encompassing
residues 3389 ± 3759 and spanning the SET domain (aa
3610 ± 3759). We next applied the two hybrid yeast
interaction methodology to inquire whether the SET
domain of TRX human homologue, ALL-1, and trx-G
protein ASH1, undergo self-association too. Indeed, in
the case of ALL-1 protein, interaction was observed
between the segments containing amino acids 3617 ±
3969 and 3745 ± 3969 and spanning the SET domain
(residues 3818 ± 3969). ASH1 interacting region included SET (residues 1318 ± 1448) together with an
upstream sequence (aa 1160 ± 1317). An alternative
interaction region of ASH1 contained residues 1245 ±
1525 extending over the SET domain. To test the
speci®city of these interactions, TRX SET, ALL-1 SET
and ASH1 SET interacting regions were cotransfected
into SFY526 with the unrelated baits p53, LAMIN,
ANKYRIN, DAP KINASE, a partial FAS protein,
ICE, MORT-1 and TRAF-2. None of the unrelated
baits interacted with the tested SET domains. We note
that self association of ASH1 and TRX SET was
stronger than that of ALL-1 SET (faster staining
reaction).
To precisely delineate the domains participating in
self-association, we generated a series of deletions
encoding the TRX SET and ALL-1 SET polypeptides
and tested them in the yeast interaction assay. The
shortest active segment of TRX (aa 3540 ± 3759)
included the entire SET motif together with an
upstream stretch of 70 residues. Deletion of TRX
SET amino acid residues 3652 ± 3759 or 3704 ± 3759
resulted in the abolishment of self-association of TRX
SET. The shortest active segment of ALL-1 SET
contains aa 3745 ± 3969; deletion of amino acid residues
3853 ± 3969 or 3899 ± 3969 eliminated interaction (data
not shown).
To con®rm these results we applied GST pull down
methodology as well as co-immunoprecipitation
approaches. To demonstrate in vitro binding, TRX,
ALL-1 and ASH1 polypeptides spanning the SET
domains were linked to GST, expressed in bacteria,
Figure 1 Self-association of TRX, ASH1 and ALL-1 SET polypeptides as analysed in vitro by GST pull down assays. The
radiolabeled 35S SET fragments were generated by coupled transcription/translation. GST-TRX, ASH1 and ALL-1 SET
polypeptides and GST polypeptide were produced in bacteria. Amount in input is 35% of the amount applied to beads
Oncogene
Self-association of SET
T Rozovskaia et al
and bound to glutathione-Sepharose beads. TRX,
ALL-1 and ASH1 SET fragments were synthesized
and radiolabeled in a coupled transcription /translation
system and examined for binding to the GST chimera
polypeptides and to GST alone (Figure 1). Between
10% and 30% of the input radiolabeled polypeptides
were bound to the beads containing the GST fusions,
while only marginal amounts were bound to the beads
carrying the GST protein.
Another methodology utilized was to co-immunoprecipitate the radiolabeled TRX, ALL-1 and ASH1
SET polypeptides together with the epitope-tagged (T7)
corresponding partners, also made in vitro. Indeed, the
radiolabeled TRX SET, ALL-1 SET and ASH1 SET
polypeptides co-immunoprecipitated with the T7tagged relevant SET fragments (Figure 2). Coimmunoprecipitation with two unrelated polypeptides
was considerably less ecient (Figure 2).
To demonstrate in vivo binding in transfected cells,
TRX, ALL-1 and ASH1 SET polypeptides were tagged
with the T7 and hemagglutinin (HA) epitopes, and
plasmids encoding appropriate partners were transiently cotransfected into COS cells. The proteins
produced were examined for co-precipitation by
Western blotting. In each case, signi®cant co-precipitation was observed (Figure 3).
To further test the ability of the SET domain to
self associate in vivo, we generated transgenic ¯y lines
which carry an HA-tagged version of the SET domain
of TRX under the control of a heat shock promoter.
Extracts from transgenic adult ¯ies kept for 1 h at
378C, 30 min prior to collection, were incubated with
either anti-TRX antibodies or with preimmune serum,
and antibody-protein complexes were isolated using
Protein A-coated beads. Proteins in the pellet were
analysed by Western blotting with a mouse monoclonal antibody directed against the HA epitope. Only
low levels of TRX SET protein were found in the
pellet when preimmune serum was used for immunoprecipitation (Figure 4). In contrast, TRX SET was
readily detected in the pellet when either of two
di€erent anti-TRX antibodies directed against the Nterminal (N1) or the central (N4) regions of the TRX
molecule (Kuzin et al., 1994) were used for
immunoprecipitation (Figure 4). Essentially the same
results were obtained when extracts were prepared
from embryos following heat induction (not shown),
indicating that TRX and overexpressed SET are
physically associated in both embryonic and adult ¯y
extracts. The results of the described experiments
strongly suggest that the SET domain of TRX is
capable of homo-oligomerization both in vitro and in
vivo.
We next applied point mutagenesis methodology
to test whether modi®cation of conserved residues
within SET would a€ect the interaction in yeast.
Ten highly conserved amino acids within TRX SET
(Tripoulas et al., 1996; Jenuwein et al., 1998) were
353
Figure 2 In vitro co-immunoprecipitation of the TRX, ASH1 and ALL-1 SET polypeptides together with the corresponding SET
fragments. T7-tagged TRX, ASH1, and ALL-1 SET polypeptides, two unrelated proteins (T7-tagged p62 nucleoporin related
protein and T7-tagged Dsup35 protein cloned by us in the context of other experiments) and 35S-labeled ASH1, TRX and ALL-1
SET segments were synthesized in a coupled transcription/translation system. Immunoprecipitation was performed with 5 mg antiT7 mAb (Novagen), 35S-labeled co-precipitated proteins were analysed by autoradiography
Oncogene
Self-association of SET
T Rozovskaia et al
354
Figure 3 In vivo co-immunoprecipitation from transfected COS cells producing TRX, ASH1 and ALL-1 SET polypeptides.
Immunoprecipitation was performed with anti-T7 (TRX SET and ASH1 SET) or anti-HA (ALL-1 SET) mAb. The proteins in the
pellets were detected with anti-HA and anti-T7 mAb, respectively
Figure 4 Endogenous TRX co-immunoprecipitates from Drosophila extracts together with HA-tagged TRX SET induced by heat
shock. Following heat treatment, extracts from HA-TRX SET
transgenic ¯ies were incubated with preimmune serum (PRE), or
with either of anti-TRX Ab (N1 or N4), and precipitated with
protein A-sepharose beads. Ab N1 and N4 do not recognize any
portion of the TRX SET construct protein. After precipitation,
the presence of TRX SET was examined in the pelleted material
(PELLET) eluted from the beads, by immunoblotting with anti
HA mAb. SUP, supernatant (one tenth of the probe). 47 kD size
marker is indicated
mutated. These alterations resulted in loss of all or
most of all the capacity of the modi®ed polypeptides
Oncogene
to self-associate in yeast (Figure 5a). In contrast,
mutagenesis of three nonconserved residues located
within the SET domain or immediately upstream of
it did not a€ect the interaction. In all cases, the
expression level of the mutated polypeptides in yeast
was monitored by Western blot analysis (Figure 5b).
In parallel, all mutants were examined for interaction with SNR1, and the results obtained were
analogous to those observed for self association
(Figure 5a). Finally, limited mutagenesis of ASH1
SET conserved amino acids showed that conversion
of GWG (residues 1319 ± 1321) into VWV, PN
(1390, 1391) into AA, I (1413) into A, and DY
(1422, 1423) into AA, also resulted in loss of all or
most of all of the potential of ASH1 SET to selfassociate (not shown).
Self association of the SET domains of ALL-1,
TRX and ASH1 is evidenced by several methodologies
described above, including coimmunoprecipitation of
overexpressed TRX SET with endogenous TRX from
¯y extracts. Perhaps the best evidence for the
physiological relevance of this interaction comes from
the demonstration that mutations in evolutionary
conserved residues in TRX SET abolish self association. The interaction is not shared by all SET
domains. Thus, the SET domains (together with
upstream and/or downstream adjacent sequences) of
Drosophila E(Z) and SU(VAR)3-9, as well as of S.
cerevisae SET1, do not self-associate (Rozovskaia,
unpublished). Therefore, self-oligomerization does not
Self-association of SET
T Rozovskaia et al
355
Figure 5 E€ect of point mutations in TRX SET on self-association and on interaction of SET with SNR1, as determined in yeast
two hybrid assays (a) and expression of mutant TRX SET proteins in yeast (b). Association between the polypeptides was
determined by synthesis of b-galactosidase in the SFY526 reporter strain as assayed on ®lters. `Strong' and `very weak' interactions
indicate that color developed in less than 1 h or in more than 12 h, respectively
appear to be a general characteristic of all SET
domains. We also note that SNR1 interaction with
TRX SET is not shared with the SET domains of
ASH1, E(Z) and SU(VAR)3-9 (ibid). The human
homologue of SNR1, INI1, does interact with the
SET domain of ALL-1 (Rozenblatt-Rosen, 1998) and
of ALR (Prasad et al., 1997) protein (Ben-Simchon,
unpublished). Thus, currently it appears that in
addition to a general unidenti®ed function presumably shared by all SET domains, some individual
motifs might have acquired unique functions. Self
association might operate in linking ALL-1 (TRX and
ASH1) protein molecules residing simultaneously on
di€erent maintenance elements, so as to integrate their
activity in activation of a shared target gene(s).
Materials and methods
Yeast two hybrid assays
The HF7c and SFY526 yeast reporter strains as well as
pGBT9 vector were purchased from Clontech. Drosophila
third instar larvae cDNA library was obtained from Dr S
Elledge. 2.66106 independent clones were screened according
to the Matchmaker Two-Hybrid System Protocol (Clontech).
The analysis included growth of HF7c transformants on His7
plates containing 5 ± 20 mM 3-aminotriazol and synthesis of
b-galactosidase in SFY526 and HF7c transformants. Point
mutants with single amino acid changes were produced by a
PCR-based strategy using oligonucleotide-directed mutagenesis and cloning into the pGBT9-DBD vector. The binding
properties of the examined proteins were assessed by a ®lter
assay in a two hybrid b-galactosidase expression test in
Oncogene
Self-association of SET
T Rozovskaia et al
356
SFY526. To determine the synthesis of the mutated proteins
in yeast, the inserts were recloned into the yeast vector pAS1CYH2. Following transformation, the produced proteins
were detected by Western blotting.
In vitro binding
ASH1, TRX and ALL-1 SET polypeptides were expressed as
glutathione S-transferase (GST) fusion proteins in Escherichia
coli, puri®ed and immobilized on anity matrix beads
according to standard methodology. 35S-labeled ASH1,
TRX and ALL-1 SET fragments were prepared in vitro by
utilizing the TNT T7 coupled transcription/translation
reticulocyte lysate system (Promega), according to the
instructions of manufacturer. 35S-labeled ASH1, TRX, or
ALL-1 SET polypeptides were diluted in binding bu€er
(20 mM Tris [pH 8.0], 0.2% Triton X-100, 2 mM EDTA,
150 mM NaCl, 1 mM phenylmethylsulfonyl ¯uoride, 1 mg/ml
chymostatin, 2 mg/ml antipain, 2 mg/ml pepstatin A, 2 mg/ml
aprotinin, 5 mg/ml leupeptin) and incubated for 2 h at 48C
with the beads containing equal amounts of immobilized
GST fusion or GST protein. The beads were washed three
times with 1 ml of binding bu€er, boiled in 26sample bu€er
and the eluted proteins were resolved on 10% SDS ±
polyacrylamide gels and visualized by autoradiography.
In vitro immunoprecipitation
T7-tagged ASH1, TRX and ALL-1 SET polypeptides, two
unrelated proteins (T7-tagged p62 nucleoporin related protein
and T7-tagged Dsup35 protein, cloned by us in the context of
other experiments) and 35S-labeled ASH1, TRX and ALL-1
SET fragments were synthesized in a coupled transcription/
translation system (Promega). Immunoprecipitation was
performed by incubation (2 h at 48C) of equal amounts of
T7-tagged proteins with radiolabeled 35S SET polypeptides
followed by incubation (2 h at 48C) with 5 mg anti-T7 mAb
(Novagen) in 0.5 ml binding bu€er containing 50 mM Tris
[pH 7.4], 150 mM NaCl, 5 mM EDTA, 0.1% NP 40, 1 mM
phenylmethylsulfonyl ¯uoride, 1 mg/ml chymostatin, 2 mg/ml
antipain, 2 mg/ml pepstatin A, 2 mg/ml aprotinin, 5 mg/ml
leupeptin. Thirty microliters of protein G-Sepharose beads
were added and incubation was continued for 1 h at 48C.
Protein G-Sepharose beads were spun down and washed
three times with 1 ml of binding bu€er. Proteins were
recovered by boiling in 26sample bu€er and bound
radiolabeled proteins were analysed by electrophoresis and
autoradiography. Amounts of T7-tagged proteins were
determined by Western blot analysis utilizing anti-T7 mAb
(Novagen) diluted 1 : 10 000.
Immunoprecipitation from transfected cells
For transfection experiments the SET polypeptides of TRX,
ASH1 and ALL-1 were linked by their N-termini to T7 or
HA epitopes, as well as to nuclear localization signal. COS
cells were transfected by the calcium phosphate technique
with 10 mg of plasmid (pSG5 vector, Stratagene). At 48 ± 72 h
post-transfection, the cells were washed with PBS, scraped
and homogenized in 500 ml of a lysis bu€er containing 50 mM
Tris [pH 7.4], 250 mM NaCl, 5 mM EDTA, 0.1% NP40,
1 mM phenylmethylsulfonyl ¯uoride, 1 mg/ml leupeptin, 1 mg/
ml aprotinin and 1 mg/ml pepstatin A and kept on ice for
30 min. After centrifugation at 14 000 r.p.m. for 2 min, the
supernatant was collected and diluted by adding 400 ml of
similar bu€er lacking NaCl. For immunoprecipitation the
extracts were mixed gently and subsequently incubated with
5 mg of anti-T7 (Novagen) or 2 mg anti-HA (Boehringer
Mannheim) mAb for 2 h at 48C, 30 ml of protein GSepharose beads were added and incubation was continued
for 1 h. Protein G-Sepharose beads were spun down and
washed three times with 1 ml of binding bu€er. Proteins were
recovered by adding 26sample bu€er and boiling. For
detection on blots, the anti-T7 and HA mAb were diluted
1 : 10 000 and 1 : 250, respectively.
Construction of transgenic ¯ies and immunoprecipitation from
Drosophila extracts
TRX SET was tagged at the N-terminus with the HA
epitope, inserted into the pCaSpeR-hs vector and injected
into y7 w7 embryos. The Drosophila line homozygous for
insertion of the transgene was used for experiments of coimmunoprecipitation with endogenous TRX. Adult ¯ies or
4 ± 18 h-old embryos were incubated for 1 h at 378C and
followed by 30 min at 258C and then collected. One hundred
¯ies or 150 ± 200 ml of dechorinated embryos were homogenized in 500 ml of lysis bu€er (Rozenblatt-Rosen et al.,
1998) and kept on ice for 30 min. After centrifugation at
14 000 r.p.m. for 2 min, the supernatant was collected and
diluted by adding 400 ml of similar bu€er lacking NaCl. The
solution was precleared by incubation with protein ASepharose beads for 40 min followed by centrifugation at
14 000 r.p.m. for 5 min. The supernatant was mixed gently
for 1 h and subsequently incubated with anity puri®ed antiTRX Ab (N1 or N4) or anity puri®ed preimmune Ab for
2 h at 48C. Thirty microliters of protein A-Sepharose beads
were added and incubation was continued for 1 h at 48C.
Protein A-Sepharose beads were spun down and washed
three times with 1 ml of binding bu€er. Proteins were
recovered by adding 26sample bu€er and boiling. HAtagged TRX SET polypeptide was detected by immunoblotting using mAb directed towards the tag (Babco).
Acknowledgements
This work was supported by grants from the German
Institute of Cancer Research (DKFZ), The Israeli Academy of Science, The Louis and Fannie Tolz Fund, Israel
Cancer Research Fund, The Israel Committee Against
Cancer and by grant CA 50507 from the National Cancer
Institute.
References
Cardoso C, Timsit S, Villard L, Khrestchatisky M, Fontes M
and Colleaux L. (1998). Hum. Mol. Genet., 7, 679 ± 684.
Corda Y, Schramke V, Longhese MP, Smokvina T, Paciotti
V, Brevet V, Gilson E and GeÂli V. (1999). Nature Genet.,
21, 204 ± 208.
Cross SH, Meehan RR, Nan X and Bird A. (1997). Nature
Genet., 16, 256 ± 259.
Cui X, De Vivo I, Slany R, Miyamoto A, Firestein R and
Cleary ML. (1998). Nature Genet., 18, 331 ± 337.
Oncogene
Ekwall K, Nimmo ER, Javerzat JP, Borgstrom B, Egel R,
Cranston G and Allshire R. (1996). J. Cell Sci., 109,
2637 ± 2648.
Gu Y, Nakamura T, Alder H, Prasad R, Canaani O, Cimino
G, Croce CM and Canaani E. (1992). Cell, 71, 701 ± 708.
Heim S and Mitelman F. (1995). Cancer Cytogenetics. WileyLiss, NY:NY, USA.
Jenuwein T, Labile G, Dorn R and Reuter G. (1998). Cell.
Mol. Life Sci., 54, 80 ± 93.
Self-association of SET
T Rozovskaia et al
Jones RS and Gelbart WM. (1993). Mol. Cell Biol., 13,
6357 ± 6366.
Kuzin B, Tillib S, Sedkov Y, Mizrokhi L and Mazo A.
(1994). Gen. Dev., 8, 2478 ± 2490.
Lewis EB. (1978). Nature, 276, 565 ± 570.
Mazo A, Huang DH, Mozer BA and Dawid IB. (1990). Proc.
Natl. Acad. Sci. USA, 87, 2112 ± 2116.
Nislow C, Ray E and Pillus L. (1997). Mol. Cell Biol., 8,
2421 ± 2436.
Prasad R, Yano T, Sorio C, Nakamura T, Rallapalli R, Gu
Y, Leshkowitz D, Croce CM and Canaani E. (1995). Proc.
Natl. Acad. Sci. USA, 92, 12160 ± 12164.
Prasad R, Zhadanov AB, Sedkov Y, Bullrich F, Druck T,
Rallapalli R, Yano T, Alder H, Croce CM, Huebner K,
Mazo A and Canaani E. (1997). Oncogene, 15, 549 ± 560.
Raimondi SC. (1993). Blood, 81, 2237 ± 2251.
Rasio D, Schichman SA, Negrini M, Canaani E and Croce
CM. (1996). Cancer Res., 56, 1766 ± 1769.
Rozenblatt-Rosen O, Rozovskaia T, Burakov D, Sedkov Y,
Tillib S, Blechman J, Croce C, Mazo A and Canaani E.
(1998). Proc. Natl. Acad. Sci. USA, 95, 4152 ± 4157.
Rozovskaia T, Tillib S, Smith S, Sedkov Y, RozenblattRozen O, Petruk S, Yano T, Nakamura T, Ben-Simchon
L, Gildea Y, Croce CM, Shearn A, Canaani E and Mazo
A. (1999). Mol. Cell Biol., 19, 6441 ± 6447.
Thanos D and Maniatis T. (1995). Cell, 89, 1091 ± 1100.
Tkachuk DC, Kohler S and Cleary ML. (1992). Cell, 71,
691 ± 700.
Tripoulas N, LaJeuness D, Gildea J and Shearn A. (1996).
Genetics, 143, 913 ± 928.
Tschiersch B, Hofmann A, Kraus V, Dorn R, Korge G and
Reuter G. (1994). EMBO J., 13, 3822 ± 3831.
Yano T, Nakamura T, Blechman J, Sorio C, Dang CV,
Geiger B and Canaani E. (1997). Proc. Natl. Acad. Sci.
USA, 94, 7286 ± 7291.
357
Oncogene