BRD–NUT oncoproteins: a family of closely related nuclear proteins that
block epithelial differentiation and maintain the growth of carcinoma cells
CA French1, CL Ramirez2, J Kolmakova3, TT Hickman1, MJ Cameron1, ME Thyne1, JL Kutok1,
JA Toretsky4, AK Tadavarthy5, UR Kees6, JA Fletcher1and JC Aster1
1Department of Pathology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA;2Departments of Biological
and Biomedical Sciences, Harvard Medical School, Boston, MA, USA;3Boston University School of Medicine, Boston, MA, USA;
4Department of Oncology, Lombardi Comprehensive Cancer Center and Pediatrics, Georgetown University, Washington, DC, USA;
5Allina Hospitals and Clinics, Minneapolis, MN, USA and6Telethon Institute for Child Health Research, and Centre for Child Health
Research, The University of Western Australia, Perth, Australia
An unusual group of carcinomas, here termed nuclear protein
in testis (NUT) midline carcinomas (NMC), are characterized
by translocations that involve NUT, a novel gene on
chromosome 15. In about 2/3rds of cases, NUT is fused to
BRD4 on chromosome 19. Using a candidate gene approach,
we identified two NMCs harboring novel rearrange-
ments that result in the fusion of NUT to BRD3 on
chromosome 9. The BRD3–NUT fusion gene encodes a
protein composed of two tandem chromatin-binding bromo-
domains, an extra-terminal domain, a bipartite nuclear
localization sequence, and almost the entirety of NUT that
is highly homologous to BRD4–NUT. The function of NUT
is unknown, but here we show that NUT contains nuclear
localization and export sequences that promote nuclear-
cytoplasmic shuttling via a leptomycin-sensitive pathway. In
contrast, BRD3–NUT and BRD4–NUT are strictly nuclear,
implying that the BRD moiety retains NUT in the nucleus via
interactions with chromatin. Consistent with this idea, FRAP
studies show that BRD4, BRD4–NUT and BRD3–NUT have
significantly slower rates of lateral nuclear diffusion than that
of NUT. To investigate the functional role of BRD–NUT
fusion proteins in NMCs, we investigated the effects of
siRNA-induced BRD3–NUT and BRD4–NUT withdrawal.
Silencing of these proteins in NMC cell lines resulted in
squamous differentiation and cell cycle arrest. Together, these
data suggest that BRD–NUT fusion proteins contribute to
carcinogenesis by associating with chromatin and interfering
with epithelial differentiation.
Oncogene (2008) 27, 2237–2242; doi:10.1038/sj.onc.1210852;
published online 15 October 2007
Keywords: BRD3–NUT; t(15;19); t(9;15); NUT midline
Nuclear protein in testis (NUT) midline carcinoma
(NMC) is defined by chromosomal rearrangements
involving NUT a novel gene on chromosome 15 (French
et al., 2001, 2004). NMCs are undifferentiated or poorly
differentiated squamous cell carcinomas that arise in the
mid-line structures of children or adults. With a single
exception (Mertens et al., 2006), NMCs have proven fatal
within 6 months of diagnosis, even in patients treated
with multimodality therapy. In 2/3rds of NMCs,
essentially the entire coding region of NUT is inserted
into the 30end of BRD4, creating a BRD4–NUT fusion
gene (French et al., 2003). In the remaining cases, NUT is
rearranged, but the fusion partners are unknown (French
et al., 2004).
NUT midline carcinomas typically have simple karyo-
types; often the only aberration is the rearrangement
involving NUT. Such simple karyotypes are unusual in
carcinomas, which typically have complex chromosomal
abnormalities, but common in certain leukemias,
lymphomas and sarcomas. Based on precedents in these
other tumor types, NUT-fusion proteins may initiate
malignant transformation within epithelial cell precur-
sors and require relatively few collaborative mutations
or epigenetic changes to produce an NMC.
Proteins of the BRD family contain two bromo-
domains (BD) that bind transcriptionally active chroma-
tin through associations with acetylated histones H3 and
H4 (Dey et al., 2003). The BRD4 BD and the flanking
extra-terminal (ET) domain also participate in interac-
tions with other proteins, including P-TEF-b (Jang et al.,
2005; Yang et al., 2005), through which BRD4 stimulates
RNA-polymerase II transcriptional activity; and LANA-
1 proteins (Ottinger et al., 2006), which may facilitate the
attachment of human herpesvirus 8 to chromosomally
associated BRD4 during mitotic segregation. The ET
domain of BRD3, which is closely related to BRD4 in
structure, also binds LANA-1 (Ottinger et al., 2006).
Overexpression of BRD4 perturbs cell cycle progression,
a phenotype that may be related to additional inter-
actions with SPA-1, a Rap GTPase-activating protein
(Farina et al., 2004), and replication factor C (Maruyama
et al., 2002).
Received 28 March 2007; revised 10 August 2007; accepted 17 September
2007; published online 15 October 2007
Correspondence: Dr CA French, Department of Pathology, Brigham
and Women’s Hospital, 75 Francis St, Boston, MA 02115, USA.
Oncogene (2008) 27, 2237–2242
& 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00
Detection and characterization of a BRD3–NUT
To better understand the structural requirements for epithe-
lial transformation by NUT oncoproteins, we screened
NMCs for rearrangements involving other members of the
bromodomain gene family. Studies were initiated with three
NUT-variant NMCs with sufficient paraffin-embedded
archival tissue (French et al., 2004). Dual-color ‘bring-
together’ FISH (fluorescent in situ hybridization) assays
were developed for 11 genes encoding bromodomain
proteins (Supplementary Data, Supplementary Table 1).
In 1 of 3 cases (case 1, a well-differentiated squamous cell
carcinoma; Figure 1a), nuclear co-localization of NUT and
BRD3 hybridization signals was observed (Figure 1c),
suggesting the presence of a BRD3–NUT fusion gene. In the
other 2 cases, none of the 11 genes encoding bromodomain
proteins were rearranged (data not shown). Subsequently,
fresh frozen material and viable cells were obtained from a
poorly differentiated carcinoma with a three-way 2;9;15
chromosomal translocation with breakpoints near the
locations of BRD3 (9q34.2) and NUT (15q14) (case 2;
Figure 1b). FISH performed on metaphase and interphase
cells from this tumor showed the presence of an apparent
BRD3–NUT fusion gene (Figure 1d). In line with prior
experience, the two patients with BRD3–NUT carcinoma
lived 148 weeks (case 1) and 8 weeks (case 2) following
Reverse transcriptase–PCR amplification of RNA
obtained from case 2 using 50BRD3- and 30NUT-
specific primers gave a single product, which proved to
be a fusion cDNA created by splicing of BRD3 exon 9 to
NUT exon 2 (Figure 1e). The BRD3–NUT mRNA is
predicted to encode a polypeptide comprised of the two
bromodomains and the ET domain of BRD3 fused to all
but the first 6 amino acids of NUT (Figure 1f). This
polypeptide is highly homologous to BRD4–NUT,
which in previously analysed tumors (French et al.,
2003; Haruki et al., 2005) was the product of a fusion
transcript created by splicing of BRD4 exon 10b to NUT
exon 2 (Figure 1f). In the region of structural overlap,
BRD3 and BRD4 are 54% identical and 64% similar,
carcinoma. Case 2 (b), poorly differentiated carcinoma. (c and d) Fluorescent in situ hybridization (FISH) demonstrating fusion of
BRD3 and NUT. Fusion signals (arrowheads) are seen in the interphase nuclei of the tumor cells from case 1 (c) and a metaphase
chromosome preparation of tumor cells from case 2 (d). Green signal, probe lying telomeric (30) of NUT; red signal, probe lying
telomeric (50) of BRD3. (e) Reverse transcriptase–PCR amplification of a B750bp BRD3–NUT fusion transcript. Case 2, input cDNA
from case 2; dH2O, no input cDNA; HeLa and TC-797, negative control input cDNAs from cell lines lacking BRD3–NUT fusion
genes. Sequencing of case 2 product revealed an in-frame fusion of BRD3 exon 9 to NUT exon 2. (f) Comparison of the predicted
structures of BRD3–NUT and BRD4–NUT.
Characterization of BRD3–NUT. (a and b) Histology BRD3–NUT carcinomas. Case 1 (a), well-differentiated squamous cell
CA French et al
with the highest degree of homology occurring in the
BRD and ET domains. Exon 10b of BRD4 is present
normally in an alternatively spliced mRNA transcript
that encodes a large BRD4 isoform (French et al., 2003).
In the BRD4–NUT transcript, exon 10b contributes the
coding sequence for a short low complexity sequence
that is rich in serine residues; based on its absence from
BRD3–NUT, this sequence is probably not required for
oncogenesis. In contrast, attempted RT–PCR with
multiple different 50NUT and 30BRD3 primer pairs
failed to yield products (data not shown). Thus, as in
translocations involving BRD4 and NUT (French et al.,
2003; Haruki et al., 2005), it appears that the reciprocal
NUT-BRD3 fusion gene is transcriptionally silent.
Identification of BRD3–NUT and BRD4–NUT fusion
proteins in NUT carcinoma cells
We developed NUT antibodies to verify expression of
BRD3–NUT and BRD4–NUT polypeptides in NMCs.
Polyclonal antibodies raised against NUT recognized a
single polypeptide of B215kDa in lysates prepared from
case 2, about the expected size of BRD3–NUT (191kDa);
and a single polypeptide of B230kDa in lysates prepared
from endogenous BRD4–NUT-expressing TC-797 cells,
about the expected size of BRD4–NUT (200kDa)
(Figure 2a). Upon immunohistochemical (IHC) staining,
all NMCs (including 2 of 2 tumors with BRD3–NUT
rearrangements; 11 of 11 tumors with BRD4–NUT
rearrangements; and 3 of 3 evaluable tumors with
NUT-variant rearrangements) displayed finely speckled
nuclear staining with NUT antiserum (Figures 2b–g, and
insets Figures 2b and d). In contrast, poorly differentiated
non-small cell carcinomas lacking NUT rearrangements
(based on FISH analysis showing that NUT was intact)
demonstrated no staining or weak staining in o10% of
tumor cells (N¼10) (Figure 2h). Although this focal
reactivity could indicate variable expression of NUT in
typical non-small cell carcinomas, attempts to detect
NUT mRNA in such tumors has been unsuccessful (data
not shown), suggesting that it more likely stems from
cross-reactivity with another antigen.
Nuclear retention of BRD–NUT
We sought to determine whether the subcellular localiza-
tion of NUT differs from that of BRD–NUT fusion
proteins. Nuclear localization sequence (NLS), BLASTX,
Swiss-Prot, PROSITE and SWISS-MODEL programs
identified two predicted nuclear localization signal se-
quences (NLSs) at the C-terminal end of NUT. However,
when expressed transiently, GFP–NUT showed either
cytoplasmic or nuclear localization, suggesting that it is
subject to nuclear/cytoplasmic shuttling (Figure 3a). Con-
sistent with this possibility, treatment with leptomycin B,
an inhibitor of CRM1-dependent nuclear export (Ossareh-
Nazari et al., 1997) resulted in re-distribution of GFP–
NUT to the nucleus (Figure 3b). Inspection of NUT
from a BRD4–NUT cell line (Toretsky et al., 2003); case 2, lysate from a BRD3–NUT tumor. Lysates from HeLa cells and 293T cells
transfected with pcDNA3 FLAG-BRD4–NUT or empty pcDNA3 vector are included controls. Staining for GAPDH (Ambion,
Austin, TX, USA) serves as a loading control. (b–h) Immunohistochemistry. Representative results of staining with anti-NUT are
shown for: (b) a NMC with a BRD4–NUT fusion gene; (c and d) two NMCs with BRD3–NUT fusion genes (case 1, c; and case 2, d);
and (e–g) three NMCs with uncharacterized NUT-variant fusion genes. (h) Staining of a representative squamous cell carcinoma of the
lung without a NUT rearrangement. (b and d insets) A speckled nuclear pattern of NUT antibody staining is characteristic of NMCs.
Detection of BRD4–NUT and BRD3–NUT polypeptides. (a) western blot analysis with NUT antibody. TC797, cell lysate
CA French et al
to either the nucleus or the cytoplasm. (b) GFP–NUT localizes only to the nucleus of cells treated with leptomycin B. (c) GFP–NUT-
S1026A/S1029A/S1031A localizes to the nucleus. (d) GFP–BRD4–NUT and (e) GFP–BRD3–NUT localize to the nucleus in a
distinctively speckled pattern. In (a–e) representative images of transiently transfected U2OS cells are shown: GFP, top; DAPI nuclear
counterstain, bottom. (f) Sequence of the putative NUT nuclear export signal sequence. S residues affecting export are underlined. (g)
Schematic of NUT. (h) Effect of BD moieties on the distribution and nuclear mobility of NUT. The recovery time after photobleaching
is depicted for representative cells. FRAP was performed in triplicate, and findings were similar in all cells tested.
Subcellular localization and nuclear mobility of GFP–NUT and GFP–BRD–NUT fusion proteins. (a) GFP–NUT localizes
CA French et al
revealed a C-terminal sequence (amino acids 1017–1042)
similar to known nuclear export sequences (NES)
(Figure 3f), which are often regulated by phosphorylation
(Fornerod et al., 1997). Consistent with the presence of a
functional NES, forms of GFP–NUT with deletions that
remove this sequence (not shown) or alanine substitutions
at residues S1026, S1029 and S1031 (Figure 3f) localized
only to the nucleus (Figure 3c). GFP–NUT polypeptides
bearing single or double S to A substitutions of the same
three residues demonstrated both nuclear and cytoplasmic
staining (data not shown), suggesting that these three S
residues influence shuttling of NUT coordinately. We also
performed fluorescence recovery after photobleaching
(FRAP) on rare cells where GFP–NUT was localized to
both the cytoplasm and the nucleus. Fluorescence rapidly
recovered in the bleached nucleus and was depleted in the
cytoplasm, indicating that in these cells GFP–NUT is
being actively transported into the nucleus (Supplementary
Figures S1a and b). Conversely, washout of leptomycin
resulted in redistribution of GFP–NUT from the nucleus
to the cytoplasm (Supplementary Figures S1c and d).
In contrast, GFP–BRD4–NUT and GFP–BRD3–NUT
were found only in the nucleus (Figures 3d–e) in a speckled
pattern resembling that seen in NMC cells (Figures 2b and
d, insets). BRD4 is known to be constitutively nuclear and
to associate with chromatin (Dey et al., 2003; Haruki et al.,
2005). Because the BDs of BRD4 are homologous to those
of BRD3, it seemed likely that the nuclear retention of
BRD–NUT polypeptides was mediated through BD–
chromatin interactions. In support of this idea, FRAP
revealed that the intranuclear diffusion rates of GFP–
BRD3–NUT and GFP–BRD4–NUT were equivalent to
one another and significantly less than that of GFP–NUT
(Figure 3h; see Supplementary Figures S2a and b for
representative images). In addition, the mobilities of GFP–
BRD3–NUT and GFP–BRD4–NUT were approximately
twofold less than that of GFP–BRD4, suggesting that
additional protein–protein interactions involving NUT
TC797, PER-403 (both BRD4–NUT-expressing) and 10326 (BRD3–NUT-expressing) cell lines (hematoxylin and eosin stain, ?400,
96h post-transfection). PER–403 has been described (Kees et al., 1991). (d–f) Morphologic changes induced by siRNA knockdown of
endogenous BRD4–NUT in TC797 and PER-403 cells, and BRD3–NUT in 10326 cells (hematoxylin and eosin stain, ?400, 96h post-
transfection). The pink amorphous extracellular material (b and e) is fibrin used to produce cell blocks. (a–f, insets) BRD3/4–NUT
knockdown induces keratin expression, as evidenced by increased staining with the pan-keratin monoclonal antibody (clone MNF116,
DAKO). (g) Knockdown of BRD3/4–NUT decreases proliferation. Staining for Ki-67þ (a marker of cell cycle progression) was
performed on sections of cell blocks prepared 96h post-siRNA transfection. Bars¼50mm, (a–f). (h) NUT-specific siRNA reduces
BRD–NUT protein levels in the cell lines TC797, PER-403 and 10326. Protein extracts obtained 24h after siRNA transfection were
analysed on a western blot stained with a polyclonal NUT antibody.
Knockdown of BRD–NUT results in squamous differentiation and growth arrest. (a–c) Control siRNA has no effect on the
CA French et al
contribute to the relatively slow nuclear mobility of BRD–
NUT fusion proteins.
BRD–NUT knockdown causes NMC cell differentiation
To investigate the functional role of BRD3–NUT and
BRD4–NUT fusion proteins in NMCs, we used siRNA
to knock down BRD4–NUT in PER-403 and TC797
cells, and BRD3–NUT in 10326 cells (Figure 4h).
Starting on day two following knockdown, all three cell
lines exhibited profound morphologic changes consis-
tent with the induction of squamous differentiation,
including increased cellular cohesion, stratification,
flattening, and enlargement of the cells; and increased
keratin production(as indicated
changes and markedly increased immunohistochemical
staining for keratins, Figures 4a–f, and insets). These
changes were accompanied by decreases in cellular
proliferation, as indicated by reduced staining for Ki-
67 (Figure 4g). Overall, siRNA against NUT reduced
the Ki-67þ fraction by 9-, 17- and 1.6-fold in TC797,
PER-403 and 10326 cells, respectively (Po0.004).
BRD4 influences a number of cellular activities
through protein–protein interactions, including the tran-
scription of specific target genes. These activities appear
to be dependent on the binding of BRD4 to acetylated
chromatin through the dual bromodomains, as well as
additional protein–protein interactions mediated through
the BDs and the flanking ET domain (Dey et al., 2003;
Farina et al., 2004; Jang et al., 2005; Yang et al., 2005),
which are retained in BRD3– and BRD4–NUT fusion
proteins. In this paper, we provide the first evidence that
BRD–NUT fusion proteins contribute to the mainte-
nance of the transformed state, based on siRNA knock-
down studies showing that withdrawal of BRD3– and
BRD4–NUT results in squamous differentiation and
growth arrest. We propose that tethering of NUT to
transcriptionally active, acetylated chromatin through the
bromodomains of BRD3 and BRD4 perturbs gene
expression and maintains NMC cells in a relatively
undifferentiated state. These effects may be exacerbated
by the retention of BRD-fused NUT in the nucleus.
To the best of our knowledge, the phenotype observed
when NUT-fusion proteins are withdrawn from NMC cell
lines is unique among human carcinomas, and serves to
further highlight the unusual biology of NMCs. This
proposed oncogenic mechanism is a common theme in
other cancers, such as acute leukemias, where many fusion
oncoproteins act by interfering with transcriptional
programs that drive differentiation. In leukemias, many
of the involved genes, even those affected by rare
chromosomal aberrations, have proven to be of central
importance in understanding normal hematopoiesis.
Although NMC is a rare tumor, it seems likely that
BRD–NUT oncoproteins will provide new insights of
some general relevance to understanding normal and
pathophysiologic epithelial cell growth and differentiation.
We thank Drs Andrew Weng and Stephen Blacklow for
helpful advice and discussion. CAF is supported by a grants
from the National Cancer Institute.
Dey A, Chitsaz F, Abbasi A, Misteli T, Ozato K. (2003). The double
bromodomain protein Brd4 binds to acetylated chromatin during
interphase and mitosis. Proc Natl Acad Sci USA 100: 8758–8763.
Farina A, Hattori M, Qin J, Nakatani Y, Minato N, Ozato K. (2004).
Bromodomain protein Brd4 binds to GTPase-activating SPA-1,
modulating its activity and subcellular localization. Mol Cell Biol
Fornerod M, Ohno M, Yoshida M, Mattaj IW. (1997). CRM1 is an export
receptor for leucine-rich nuclear export signals. Cell 90: 1051–1060.
French CA, Kutok JL, Faquin WC, Toretsky JA, Antonescu CR,
Griffin CA et al. (2004). Midline carcinoma of children and young
adults with NUT rearrangement. J Clin Oncol 22: 4135–4139.
French CA, Miyoshi I, Aster JC, Kubonishi I, Kroll TG, Dal Cin P
et al. (2001). BRD4 bromodomain gene rearrangement in aggressive
carcinoma with translocation t(15;19). Am J Pathol 159: 1987–1992.
French CA, Miyoshi I, Kubonishi I, Grier HE, Perez-Atayde AR,
Fletcher JA. (2003). BRD4–NUT fusion oncogene: a novel
mechanism in aggressive carcinoma. Cancer Res 63: 304–307.
Haruki N, Kawaguchi KS, Eichenberger S, Massion PP, Gonzalez A,
Gazdar AF et al. (2005). Cloned fusion product from a rare
t(15;19)(q13.2;p13.1) inhibit S phase in vitro. J Med Genet 42: 558–564.
Jang MK, Mochizuki K, Zhou M, Jeong HS, Brady JN, Ozato K.
(2005). The bromodomain protein Brd4 is a positive regulatory
component of P-TEFb and stimulates RNA polymerase II-
dependent transcription. Mol Cell 19: 523–534.
Kees UR, Mulcahy MT, Willoughby ML. (1991). Intrathoracic
carcinoma in an 11-year-old girl showing a translocation t(15;19).
Am J Pediatr Hematol Oncol 13: 459–464.
Maruyama T, Farina A, Dey A, Cheong J, Bermudez VP, Tamura T
et al. (2002). A mammalian bromodomain protein, brd4, interacts
with replication factor C and inhibits progression to S phase. Mol
Cell Biol 22: 6509–6520.
Mertens F, Wiebe T, Adlercreutz C, Mandahl N, French CA. (2006).
Successful treatment of a child with t(15;19)-positive tumor. Pediatr
Blood Cancer; e-pub ahead of print; 2 January 2007.
Ossareh-Nazari B, Bachelerie F, Dargemont C. (1997). Evidence for a
role of CRM1 in signal-mediated nuclear protein export. Science
Ottinger M, Christalla T, Nathan K, Brinkmann MM, Viejo-Borbolla A,
Schulz TF. (2006). The Kaposi’s Sarcoma-associated Herpesvirus
LANA-1 interacts with the short variant of BRD4 and releases cells
from a BRD4- and BRD2/RING3- induced G1 cell cycle arrest.
J Virol 80: 10772–10786.
Toretsky JA, Jenson J, Sun CC, Eskenazi AE, Campbell A, Hunger SP.
(2003). Translocation (11;15;19): a highly specific chromosome
rearrangement associated with poorly differentiated thymic carcino-
ma in young patients. Am J Clin Oncol 26: 300–306.
Yang Z, Yik JH, Chen R, He N, Jang MK, Ozato K et al. (2005).
Recruitment of P-TEFb for stimulation of transcriptional elonga-
tion by the bromodomain protein Brd4. Mol Cell 19: 535–545.
Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).
CA French et al