The kinesin KIF1B? acts downstream
from EglN3 to induce apoptosis
and is a potential 1p36 tumor suppressor
Susanne Schlisio,1Rajappa S. Kenchappa,2Liesbeth C.W. Vredeveld,3Rani E. George,4
Rodney Stewart,4Heidi Greulich,5,6Kristina Shahriari,1Nguyen V. Nguyen,7Pascal Pigny,8
Patricia L. Dahia,7Scott L. Pomeroy,9John M. Maris,10A. Thomas Look,4Matthew Meyerson,1,5
Daniel S. Peeper,3Bruce D. Carter,2and William G. Kaelin Jr.1,11,12
1Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts 02115,
USA;2Department of Biochemistry and Center for Molecular Neuroscience, Vanderbilt University Medical School,
Nashville, Tennessee 37232, USA;3Division of Molecular Genetics, The Netherlands Cancer Institute, 1066 BE Amsterdam,
The Netherlands;4Department of Pediatric Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston,
Massachusetts 02115, USA;5Center for Cancer Genome Discovery, Dana-Farber Cancer Institute, Harvard Medical School,
Boston, Massachusetts 02115, USA;6The Broad Institute of Harvard and Massachusetts Institute of Technology, Cambridge,
Massachusetts 02141, USA;7Department of Medicine and Cellular and Structural Biology, San Antonio Cancer Institute,
University of Texas Health Science Center, San Antonio, Texas 78229, USA;8Laboratoire de Biochimie and Hormonologie,
Centre de Biologie et Pathologie, CHRU de Lille 59037, Lille cedex, France;9Department of Neurology, Children’s Hospital,
Harvard Medical School, Boston, Massachusetts 02115, USA;10Division of Oncology, Children’s Hospital of Philadelphia,
Abramson Family Cancer Research Institute, University of Pennsylvania, Philadelphia, Pennsylvania 19104, USA;11Howard
Hughes Medical Institute, Chevy Chase, Maryland 20815, USA
VHL, NF-1, c-Ret, and Succinate Dehydrogenase Subunits B and D act on a developmental apoptotic pathway
that is activated when nerve growth factor (NGF) becomes limiting for neuronal progenitor cells and requires
the EglN3 prolyl hydroxylase as a downstream effector. Germline mutations of these genes cause familial
pheochromocytoma and other neural crest-derived tumors. Using an unbiased shRNA screen we found that
the kinesin KIF1B? acts downstream from EglN3 and is both necessary and sufficient for neuronal apoptosis
when NGF becomes limiting. KIF1B? maps to chromosome 1p36.2, which is frequently deleted in neural
crest-derived tumors including neuroblastomas. We identified inherited loss-of-function KIF1B? missense
mutations in neuroblastomas and pheochromocytomas and an acquired loss-of-function mutation in a
medulloblastoma, arguing that KIF1B? is a pathogenic target of these deletions.
[Keywords: Apoptosis; kinesin; neuroblastoma; pheochromocytoma; prolyl hydroxylase]
Supplemental material is available at http://www.genesdev.org.
Received January 7, 2008; revised version accepted February 14, 2008.
Developmental apoptosis of neuronal precursors is cru-
cially important for determining the final number of ter-
minally differentiated cells (Sommer and Rao 2002). De-
regulation of this process can cause disease (Zhu and
Parada 2002). Paragangliomas and neuroblastomas are
tumors of the sympathetic nervous system (paraganglio-
mas arising in the adrenal medulla are called pheochro-
mocytomas). During normal embryological develop-
ment, neuronal progenitor cells, including cells capable
of forming the sympathetic nervous system, compete
with one another for growth factors such as nerve growth
factor (NGF), with the losers undergoing apoptosis.
Germline VHL, NF-1, c-Ret, and Succinate Dehydroge-
nase Subunits B and D (SDHB and SDHD) mutations
have been linked to the development of paragangliomas
(Nakamura and Kaelin 2006). We reported recently that
the products of these genes define a pathway that is ac-
tivated upon NGF withdrawal, leading to apoptosis me-
diated by the EglN3 prolyl hydroxylase (Lee et al. 2005).
This suggests that at least some paragangliomas arise
due to failure to properly cull neuronal progenitor cells
To investigate the requirement of the proapoptotic pro-
lyl hydroxylase EglN3 in neuronal apoptosis in a geneti-
cally defined system, we isolated primary sympathetic
neurons from EglN3+/+, EglN3+/−, and EglN3−/−mice. As
E-MAIL firstname.lastname@example.org; FAX (617) 632-4760
Article published online ahead of print. Article and publication date are
online at http://www.genesdev.org/cgi/doi/10.1101/gad.1648608.
884 GENES & DEVELOPMENT 22:884–893 © 2008 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/08; www.genesdev.org
expected, wild-type and heterozygous sympathetic neu-
rons died after NGF withdrawal (Fig. 1A; Supplemental
Fig. 1). In contrast, EglN3−/−sympathetic neurons were
remarkably resistant to apoptosis in this setting (Fig.
1A). Furthermore, EglN3−/−cerebellar granular neurons
(CGNs) were resistant to apoptosis induced by the neu-
rotoxin cytosine arabinoside (Ara-C) compared with
strengthen the earlier conclusion, reached with siRNAs
and pharmacological agents, that EglN3 plays an impor-
tant role in neuronal apoptosis (Lee et al. 2005).
Next, we infected different cell types with an adeno-
virus encoding EglN3. EglN3 killed SK-N-DZ and
SK-N-SH neuroblastoma cells, PC12 rat pheochromocy-
toma cells, and SK-Mel28 melanoma cells, all of which
are derived from neural crest progenitor cells (Fig. 1C,D).
This effect was specific as EglN3 did not kill 786-O renal
carcinoma cells, HK2 immortalized renal epithelial
cells, or primary mouse embryo fibroblasts (Fig. 1C,D).
EglN3 catalytic activity was required for killing since it
could be partially rescued with the hydroxylase inhibitor
dimethyl oxalylglycine (DMOG) (Fig. 1E). EglN3-in-
duced apoptosis was not restricted to neural crest deriva-
tives, however, because EglN3 also killed U2OS osteo-
sarcoma cells, prostatic carcinoma cells (DU145 and
PC3), H1299 lung carcinoma cells, and HCT116 colorec-
tal carcinoma cells (Fig. 1C; Supplemental Fig. 2A).
p53 is a critical regulator of apoptosis and has been
implicated in developmental cell death of sympathetic
neurons (Aloyz et al. 1998). Isogenic HCT116 cells that
are p53+/+or p53−/−were both killed by EglN3 (Supple-
mental Fig. 2A). Likewise, EglN3-induced cell death in
neural crest-derived cells (SK-Mel28 melanoma cells)
was not prevented by an effective p53 shRNA or by SV40
(Fig. 1B). Theseresults
T antigen, which blocks p53 function (Supplemental Fig.
2B–D). Therefore, EglN3-induced apoptosis does not re-
To begin to understand how EglN3 induces apoptosis,
we screened for shRNAs that can prevent EglN3-induced
death. In pilot experiments, we determined the Ad-
EglN3 titer required to kill all of the SK-Mel28 mela-
noma cells in subconfluent cultures (Fig. 1C; data not
shown). Next, SK-Mel28 cells were infected with a pre-
viously described retroviral shRNA library (Berns et al.
2004) (or with empty retrovirus) prior to Ad-EglN3 in-
fection (Fig. 2A). No survivors emerged among the con-
trol cells pretreated with the empty virus. In cells pre-
treated with the shRNA library, however, 12 surviving
colonies emerged and were expanded for further analysis.
Three died when rechallenged with Ad-EglN3 and were
therefore considered false-positives, while nine re-
mained resistant (Fig. 2B). The shRNA inserts from these
latter colonies were isolated, sequenced, and retested for
their ability to protect naive SK-Mel28 cells from EglN3-
induced apoptosis. Sequence analysis of one of the two
shRNAs that scored positively in this assay predicted
that it targeted the ? splice variant of KIF1B, a member of
the kinesin 3 family (Nagai et al. 2000; Yang et al. 2001;
Zhao et al. 2001). Down-regulation of endogenous
KIF1B?, but not the alternative splice variant KIF1B?,
was confirmed by Western blot analysis (Fig. 2C). KIF1B?
and KIF1B? share an N-terminal motor domain but con-
tain different C-terminal cargo domains. Protection
against EglN3-induced cell death was conferred by two
additional, independent, KIF1B? shRNAs, arguing that
modulation of KIF1B?, rather than an off-target effect,
was responsible for this protection (data not shown).
Since KIF1B maps to 1p36 (Nagai et al. 2000; Yang et al.
sympathetic neurons of the indicated genotypes in the presence of NGF (+NGF) or 48 h after (−NGF) NGF withdrawal. (WT) EglN3+/+;
(Het) EglN3+/−; (KO) EglN3−/−. (B) Immunoblot analysis of primary CGNs of the indicated genotypes 24 h after treatment with 100 µM
Ara-C. Crystal violet (C) and immunoblot analysis (D) of indicated cell lines after infection with adenovirus encoding EglN3 (Ad-
EglN3) or Cre recombinase (Ad-Control). (E) Photomicrographs of SK-Mel-28 cells infected with the indicated amount of Ad-EglN3
(multiplicity of infection, MOI) in the presence or absence of 1 mM DMOG.
Regulation of apoptosis by EglN3. (A) Percentage of DAPI-stained nuclei exhibiting apoptotic changes in primary mouse
KIF1B? acts downstream from EglN3
GENES & DEVELOPMENT885
2001; Zhao et al. 2001), which is frequently deleted in
multiple tumor types including nervous system tumors
(Schwab et al. 1996), we hypothesized that it might func-
tion as a tumor suppressor gene through regulation of
apoptosis in neuronal, and perhaps other, tissues.
To ask whether EglN3 and KIF1B? belong to the same
pathway, PC12 cells were infected with Ad-EglN3. In-
duction of apoptosis, as determined by Caspase 3 cleav-
age, was accompanied by the induction of KIF1B?, but
not its splice variant KIF1B? (Fig. 3A). Conversely,
knockdown of human EglN3 in HeLa cervical carcinoma
cells with two independent siRNAs decreased KIF1B?
levels (Fig. 3B; Supplemental Fig. 3). Notably, an siRNA
against EglN1, which regulates the HIF? transcription
factor (Berra et al. 2003), did not affect KIF1B?, consis-
tent with the earlier conclusion that regulation of apo-
ptosis by EglN3 is HIF-independent (Lee et al. 2005).
zKIF1B levels were also markedly attenuated in zebrafish
treated with a morpholino oligonucleotide directed
against zEglN3 (Fig. 3C).
During normal neuronal development, many cells un-
dergo apoptosis as they compete for growth factors such
as NGF (Sommer and Rao 2002). Pheochromocytomas
are derived from sympathetic neuronal progenitor cells,
and PC12 cells have been used extensively as a model to
study the effects of NGF on neuronal differentiation and
survival. Consistent with previous reports (Lipscomb et
al. 1999; Lee et al. 2005), NGF withdrawal from PC12
cells caused the accumulation of EglN3, which coin-
cided with the onset of apoptosis (Fig. 3D; data not
shown). Importantly, KIF1B? was induced with similar
kinetics. Similarly, NGF withdrawal induced both
EglN3 and KIF1B? in primary rat sympathetic neurons
(Supplemental Fig. 4). Moreover, the induction of KIF1B?
by NGF withdrawal was prevented by DMOG in primary
rat sympathetic neurons and did not occur in EglN3−/−
primary mouse sympathetic neurons (Fig. 3E,F). Like-
wise, KIFB? was not induced by Ara-C in EglN3−/−CGN
(Fig. 3G). Collectively, these results indicate that EglN3
hydroxylase activity is necessary and sufficient for
KIF1B? induction. Regulation of KIF1B? by EglN3 ap-
pears to be post-transcriptional (Supplemental Fig. 5).
Whether KIF1B? is hydroxylated by EglN3 remains to be
Introduction of KIF1B? into PC12 cells was, like
EglN3 itself, sufficient to induce apoptosis, although
apoptosis occurred more rapidly with KIF1B? (1–2 d vs.
3 d) (Fig. 4A; data not shown), which is consistent with
KIF1B? acting downstream from EglN3. The percentage
of apoptotic cells at any time point did not exceed 20%,
however, because KIF1B?, like NGF withdrawal itself,
killed asynchronously (Lee et al. 2005). KIF1B? also
induced apoptosis in primary rat sympathetic neurons
(Fig. 5B). Conversely, multiple KIF1B? shRNAs pre-
vented apoptosis of primary rat sympathetic neurons
following NGF withdrawal (Fig. 4B). Therefore, KIF1B?
is both necessary and sufficient for apoptosis in this
Neuroblastomas, like pheochromocytomas, are neural
crest-derived tumors and frequently harbor deletions of
chromosome 1p encompassing the KIF1B locus (Benn et
al. 2000; White et al. 2005). KIF1B? mRNA levels are
decreased in neuroblastoma tumors with 1p36 deletions
and are also decreased in advanced-stage disease (Caren
et al. 2005; Wang et al. 2006; A. Nakagawara, pers.
comm.). In contrast, KIF1B? protein levels appear to be
highly variable across neuroblastoma lines and do not
strictly correlate with the presence or absence of 1p de-
fying shRNAs that protect against EglN3-induced apoptosis. (B)
Crystal violet staining of 12 isolated colonies that were retested
for resistance of EglN3-induced death. (Neg. control) Naive SK-
Mel28 cells. Clone found to contain KIF1B? shRNA is indi-
cated. (C) Immunoblot and crystal violet staining of SK-Mel28
cells infected to produce recovered KIF1B? shRNA and subse-
quently infected with Ad-EglN3.
Recovery of KIF1B? shRNA. (A) Schema for identi-
Schlisio et al.
886 GENES & DEVELOPMENT
letions, possibly due to adaptation in culture as well as
alternative mechanisms of KIF1B? down-regulation
(data not shown). The neuroblastoma cell line NB-1 has
an ∼500-kb homozygous deletion at 1p36 that spans
KIF1B and five other known genes (Ohira et al. 2000;
Yang et al. 2001; Krona et al. 2003). In contrast to
SK-N-SH and CHP212 cells, NB-1 cells were resistant to
EglN3-induced apoptosis but were, like other neuroblas-
toma lines, killed by wild-type KIF1B? (Fig. 4C–E). No-
tably, restoring the function of the other five 1p genes
deleted in NB-1 cells has no effect (A. Nakagawara, pers.
comm.). Similarly, LAN6 neuroblastoma cells and H460
lung carcinoma cells, both of which also produce low
levels of KIF1B?, were resistant to EglN3 (Fig. 4C,D).
Interestingly, H460 cells exhibit neuroendocrine fea-
tures (Lee et al. 1992). Killing by KIF1B? in neuroblas-
toma cells appears to be independent of its kinesin motor
function, since KIF1B?(600–1770), which lacks the KIF1B?
motor domain, still induced apoptosis (Figs. 4E, 5A,B).
Next, we sequenced the 46 coding KIF1B? exons in
111 neuroblastomas (including 44 with 1p loss of hetero-
zygosity), 52 pheochromocytomas, and 14 medulloblasto-
mas. The latter are derived from CGNs. We identified
(E646V, T827I, and P1217S), two pheochromocytomas
(S1481N and E1628K), and one medulloblastoma (S34L)
(Fig. 5A; Table 1; Supplemental Fig. 6). None of these
variants is a known polymorphism, nor were the vari-
ants detected among 270 controls of diverse ethnic back-
grounds (Thorisson et al. 2005; data not shown). In addi-
tion, we repeatedly identified common polymorphic al-
leles that lead to a Y1087C or V1554M substitution (data
The relevant exons of the corresponding germline
DNAs were, when available, also sequenced. Loss or re-
tention of the wild-type KIF1B allele within the tumors
was inferred by examination of DNA sequence tracing
and by 1p36 copy number information from high-density
single-nucleotide polymorphism (SNP) arrays or quanti-
tative real-time PCR. In ambiguous cases, allele frequen-
cies were determined by PCR amplification of the al-
tered exon from tumor DNA, followed by subcloning
and sequence analysis of multiple individual clones. The
medulloblastoma patient was germline wt/wt while the
tumor was wt/S34L (Table 1; Supplemental Fig. 6). The
neuroblastoma and pheochromocytoma variants were,
when evaluable, present in the germline (Table 1;
Supplemental Fig. 6). In three tumors there was loss of
the wild-type allele and in three tumors there was reten-
tion of the wild-type allele, including one in which there
was low-level amplification of the mutant allele (Table
1). Interestingly, the S1481N variant was present in a
28-yr-old female who at 17 mo of age presented with a
neuroblastoma and in adulthood developed a mature
ganglioneuroma and bilateral pheochromocytoma. Her
paternal grandfather harbored this allele and also devel-
oped bilateral pheochromocytoma (P.L. Dahia and P.
Pigny, in prep.).
Next, primary rat sympathetic neurons were electro-
porated with plasmids encoding wild-type KIF1B? or
these variants. The induction of apoptosis by all of the
putative disease-causing variants (S34L, E646V, T827I,
P1217S, S1481N, and E1628K) was clearly impaired rela-
tive to wild-type KIF1B? or the polymorphic variants
noblot analysis of PC12 cells infected with Ad-EglN3 or control
Ad-virus (A), HeLa cells transfected with the indicated siRNAs
(B), zebrafish embryos injected with the indicated morpholino
oligonucleotides (C), and differentiated PC12 cells subjected to
NGF withdrawal (D). (E) Percentage of DAPI-stained nuclei ex-
hibiting apoptotic changes in primary rat sympathetic neurons
subjected to NGF withdrawal in the presence or absence of
DMOG and immunoblotted for KIF1B? expression. (F) Immu-
nofluorescent detection of KIF1B? (green) and propidium iodide-
stained nuclei (red) of EglN3+/−and EglN3−/−primary mouse
sympathetic neurons before (+NGF) and after (−NGF) NGF
withdrawal. (G) Immunoblot analysis of KIF1B? induction in
EglN3+/−and EglN3−/−primary mouse CGNs after treatment
KIF1B? acts downstream from EglN3. (A–D) Immu-
KIF1B? acts downstream from EglN3
GENES & DEVELOPMENT887
Y1087C and V1554M (Fig. 5B,C; Supplemental Fig. 7).
Comparable levels of protein production were confirmed
by immunofluorescence and immunoblot analysis (Fig.
5C; Supplemental Fig. 7). These data argue that putative
disease-causing variants are pathogenic rather than the
result of benign polymorphisms or passenger mutations.
The existence of one or more human tumor suppressor
gene on chromosome 1p has been suspected for decades
(Brodeur et al. 1977; Haag et al. 1981; Stoler and Bouck
1985). Our data suggest that one such tumor suppressor
gene is KIF1B?, and that this gene is relevant to certain
tumors of neuronal origin. Nonetheless, we and others
observed that the remaining KIF1B? allele in 1p deleted
tumors and cell lines is often wild-type, contrary to the
Knudson Two-Hit scenario (Ohira et al. 2000; Yang et al.
2001; A. Nakagawara, pers. comm.; data not shown).
Moreover, two of the variants we identified (S34L and
S1481N) were not associated with the loss of the remain-
ing wild-type allele. Perhaps KIF1B? haploinsufficiency
is adequate for tumorigenesis in some contexts, espe-
cially when combined with the loss of other contiguous
1p genes such as CHD5 (Bagchi et al. 2007). In this re-
gard, we noted substantial protection against apoptosis
with an shRNA that decreased KIF1B? levels by ∼50%,
and the existence of multiple neuroblastoma and pheo-
chromocytoma suppressor genes on 1p has been sug-
gested before (Takeda et al. 1994; Cheng et al. 1995; Ichi-
miya et al. 1999; Benn et al. 2000; Opocher et al. 2003;
Wang et al. 2006). At the same time, complete loss of
KIF1B? promotes (rather than inhibits) neuronal apopto-
sis (Zhao et al. 2001), as does nearly complete elimina-
tion of KIF1B? using shRNAs (Supplemental Fig. 8).
Tumor development was, however, associated with
loss of the remaining wild-type allele for the two germ-
line neuroblastoma mutations E646V and P1217S and
the pheochromocytoma mutation E1628K. We note that
these mutations are not completely null, based on our
apoptotic assay, and therefore loss of the remaining wild-
type allele might still confer a survival advantage. The
NB-1 line appears to be unusual insofar as it has a ho-
mozygous, rather than heterozygous, KIF1B? deletion.
(A) Percentage of PC12 cells exhibiting apo-
ptotic changes, visualized with a GFP-histone
marker, after cotranfection to produce either
KIF1B?, wild-type EglN3, or catalytic-dead
EglN3 H196A. (B) Primary rat sympathetic
neurons were electroporated with the indi-
cated pSuper-shRNA plasmids, subjected to
NGF withdrawal, and analyzed for apoptotic
DAPI-stained nuclei. (C,D) Crystal violet
staining (C) and immunoblot analysis (D) of
cell lines infected with Ad-EglN3. (E) Crystal
violet staining of neuroblastoma cell lines
tranfected to produce the indicated KIF1B?
Induction of apoptosis by KIF1B?.
Schlisio et al.
888GENES & DEVELOPMENT
Among several possibilities, these cells might harbor ad-
ditional mutations that allow them to tolerate total loss
of KIF1B? function.
Our findings have potential implications with respect
to the pathogenesis of certain neural crest-derived tu-
mors such as pheochromocytomas and neuroblastomas.
Many cases of pheochromocytoma without a positive
family history are nonetheless due to previously unsus-
pected germline mutations involving VHL, c-Ret, NF1,
SDHB, or SDHD. We reported earlier that these genes,
together with EglN3, define a pathway (Fig. 4I) that is
responsible for the elimination of excess neuroblasts
during normal embryological development when growth
factors such as NGF become limiting. It is noteworthy
that neuroblastomas frequently express an NGF receptor
(Brodeur 1994), and both NF1 and VHL mutations have
been linked to this form of cancer also (Johnson et al.
1993; The et al. 1993; H. Greulich and M. Meyerson,
unpubl.). In this report, we placed KIF1B? downstream
from EglN3 and identified loss-of-function germline
KIF1B? mutations in some pheochromocytomas and
neuroblastomas. Moreover, we obtained functional data
consistent with the idea that partial loss of KIF1B?, such
as might occur with the loss of one KIF1B allele, would
protect neuroblasts from apoptosis in response to stimuli
such as NGF withdrawal. We therefore suggest that some
neuroblastomas, like pheochromocytomas, result from
KIF1B? function and allow certain neuronal progenitor
cells to escape developmental culling. This model is con-
sistent with the prediction, based on epidemiological stud-
ies, that at least ∼20% of pheochromocytomas and neuro-
blastomas involve a hereditary component (Knudson and
Strong 1972; Knudson and Meadows 1976) and could ac-
count for previously described patients (Fairchild et al.
1979; Tatekawa et al. 2006) who, like our patient with
the S1481N variant, developed both of these otherwise
rare tumors as children or young adults.
Forkhead-associated domain; (PH) Pleckstrin homology domain. (B) Percentage of DAPI-stained, GFP-positive, primary rat sympa-
thetic neurons exhibiting apoptotic changes after electroporation to produce GFP along with the indicated KIF1B? variants. The values
are mean ± SD from at least three individual experiments, and their statistical significance in comparison with the controls in a
two-sample t-test with a pooled estimate of variance is indicated by asterisks. (*) P < 0.0007. (C) Immunofluorescent detection of
Flag-KIF1B? variants (red) and GFP (green) in the cells shown in B. (D) Model linking familial pheochromocytoma genes to apoptosis
when NGF becomes limiting during neuronal development.
Characterization of tumor-associated KIF1B? mutants. (A) KIF1B? schematic with the locations of mutations. (FHA)
KIF1B? acts downstream from EglN3
GENES & DEVELOPMENT889
KIF1B? and KIF1B? are motor proteins implicated in
anterograde transport of mitochondria and synaptic
vesicle precursors, respectively (Nangaku et al. 1994;
Zhao et al. 2001). The S34L mutation maps to the
KIF1B? motor domain, although the motor domain is
dispensible for KIF1B?-induced apoptosis (Figs. 4E, 5B).
Conceivably, the S34L mutation affects the folding of
KIF1B? or eliminates an important phosphorylation site.
Clearly, additional studies are now needed to address
how, mechanistically, KIF1B? regulates apoptosis and to
determine how often it is deregulated, epigenetically or
genetically, in cancer.
Materials and methods
Primary sympathetic neurons and cerebellar neurons
EglN3−/−mice were a gift of Regeneron Pharmaceuticals. Sym-
pathetic neurons from P4 rats or mice were isolated from the
superior cervical ganglia (SCG) and were cultured in 20 ng/mL
NGF (Harlan) as described previously (Palmada et al. 2002). Cer-
ebellar granule neurons were isolated from 4-d-old mouse pups.
Cerebella were removed and dissociated by trypsinization and
plated on poly-L-ornithine-coated plates. Cells were cultured in
neurobasal media (Gibco) supplemented with B27, 0.6% dex-
trose, 2 mM glutamine, 25 mM KCl, and penicillin/streptomy-
Cell extracts were prepared in EBC buffer (50 mM Tris at pH 8.0,
120 mM NaCl, 0.5% NP-40) containing protease inhibitors, un-
less otherwise noted. Primary neurons were lysed in NP-40 lysis
buffer (10% glycerol, 50 mM Tris-HCl at pH 7.5, 150 mM NaCl,
1% NP-40, 1 mM phenylmethylsulfonyl fluoride [PMSF], 2 µg/
mL leupeptin and aprotinin). For PARP assays, cells were lysed
in 62.5 mM Tris HCl (pH 6.8), 6 M urea, 10% glycerol, 2% SDS,
5% ?-mercaptoethanol, 0.000125% bromophenol blue, and pro-
tease inhibitors. Equal amounts of protein, as measured by the
Bradford assay, were immunoblotted as described previously
(Lee et al. 2005). Rabbit polyclonal anti-EglN3 sera were gener-
ously provided by Dr. Robert Freeman (specific to mouse, rat,
and human) (Straub et al. 2003), and rabbit polyclonal antibody
against HIF? has been described recently (Berra et al. 2003).
Rabbit anti-KIF1B antibody specific for the ? isoform (cross-
reactive with mouse, rat, human, and zebrafish) (sc-28540) or
the ? form (sc-18739) were purchased from Santa Cruz Biotech-
nology. Antibody raised against cleaved Caspase 3 was pur-
chased from Cell Signaling (Asp715), and antibody raised
against PARP was from Biomol International (P9055a).
Neuronal apoptosis assays
Sympathetic neurons from P4 rats or mice were cultured in 20
ng/mL NGF (Harlan) for 2 d, then rinsed twice in Ultraculture
medium lacking NGF and once with Ultraculture medium con-
taining anti-NGF (0.1 µg/mL; Chemicon International), and
then maintained in NGF-free media for 48 h, at which point
cells were fixed in 4% paraformaldehyde (PFA) and stained with
DAPI (Vector Laboratories). Approximately 70–100 nuclei were
scored for apoptotic changes for each condition, as described
before (Kenchappa et al. 2006). In some experiments, cells were
pretreated with 1 mM DMOG for 6 h prior to and during NGF
Sympathetic neurons from P4 rat were isolated and trans-
fected with a plasmid encoding GFP alone or cotransfected with
pSuper plasmids using the Amaxa Nucleofactor device as de-
scribed previously (Kenchappa et al. 2006). Neurons were main-
tained in 20 ng/mL NGF for 4 d and then subjected to NGF
withdrawal as above. After fixation, GFP-positive neurons were
evaluated for apoptosis as above.
Analysis of undifferentiated PC12 cells treated with NGF,
followed by NGF withdrawal, was as described (Lee et al. 2005).
For transfection experiments, undifferentiated PC12 cells were
plated onto collagen-coated six-well plates 1 d before transfec-
tion with Lipofectamine 2000 (Invitrogen) according to the
manufacturer’s instructions. Transfection mixes contained 500
ng of a plasmid encoding GFP-histone (a gift of Geoffrey Wahl)
and 1–2 µg of the plasmid of interest. Seventy-two hours later,
∼400 GFP-positive cells were scored for the presence of apopto-
tic nuclei for each set of conditions.
Expression plasmids and siRNA
Adenovirus encoding EglN3 (Ad-EglN3) was a gift from Robert
Freeman. The HA-EglN3 and HA-EglN3-H196A expression
plasmids were described before (Lee et al. 2005). The NKI pRS
hairpin library was described previously (Berns et al. 2004). A
KIF1B? cDNA was PCR-amplified from a SK-Mel28 cDNA pool
and ligated into pcDNA3 via Kpn1 and Xba1 sites to make
pcDNA3-Flag-KIF1B?. Site-directed mutagenesis to generate
pcDNA3-Flag-KIF1B?-E646V, S34L, T827I, P1217S, S1481N,
QuikChange Site-Directed Mutagenesis kit (Stratagene) using
the primers 5?-GGAGATCTTATACAAAAAGGTGAAGGAA
KIF1B? variants identified in tumor samples
changeExonBase pair change1p36 status
2AAGGCAACT [C/T] GACCAGTAT
ACAAAAAGG [A/T] GAAGGAAGA
GCGAAACCA [C/T] TGTGACTGG
GAACTGGAG [C/T] CTACAGGAG
AATCCCTGA [G/A] CGACTCGTT
CCAGCTGTG [G/A] AAACACCAT
TCCCAGAGT [A/G] TGCAGATAT
TTCAGCCAG [G/A] TGCACGGCA
Gain of mutant allele (2×)
(N/A) Not applicable.
Schlisio et al.
890GENES & DEVELOPMENT
ATCTTCTGTCAGT-3?, and 5?-CAACAGAGAATTCAGCCA
Retroviruses encoding human KIF1B? shRNAs were made
using the pRetroSuper plasmid (Brummelkamp et al. 2002b) and
the following sequences: (sh) #1, 5?-GGAGCCTCTTTACAG
TAAC-3?; (sh) #3, 5?-GCAATGCCGTGTACCTAAA-3?; (sh)
#7:,5?-CGAGAGCAGTGGCTATGAT-3?. pRS-shp53 has been
described recently (Brummelkamp et al. 2002b).
Plasmids encoding shRNAs for rat EglN3 and rat KIF1B? were
made using pSuper plasmid (Brummelkamp et al. 2002a) and the
following sequences: EglN3, 5?-CAGGTTATGTTCGTCATGT-
3?; KIF1B? #1, 5?-AGAGCCACTCTCCAGTAAC-3?; KIF1B? #2,
5?-CAAGCTGGTTCGGGAGCTG-3?. siRNAs against human
EglN3 #2, 5?-CAGGUUAUGUUCGCCACGU-3?; and EglN3
Cell culture and retroviral transduction
Undifferentiated PC12 cells were maintained in DMEM con-
taining 5% fetal bovine serum (FBS) (Hyclone) and 10% horse
serum (Sigma) in 10% CO2at 37°C. Human tumor cell lines
were maintained in RPMI (neuroblastoma lines) or DMEM
(other lines) containing 10% FBS (Hyclone) in the presence of
10% CO2at 37°C. The production of retroviruses and adenovi-
ruses using Phoenix and 293A packaging cells, respectively, and
subsequent infections was carried out as described (Peeper et al.
2002; Lee et al. 2005).
Morpholino injection in zebrafish
Zebrafish were maintained and bred as described (Westerfield
1993), and were staged according to Kimmel et al. (1995). Mi-
croinjections were performed on one-cell-stage embryos accord-
ing to standard procedures (Westerfield 1993). Based on the pub-
lished GenBank sequence for zegln3 (gi:47086946), a transla-
tion-blocking morpholino was designed by GeneTools. Inc.:
protein lysates were prepared from 100 3-d-old embryos by ho-
mogenizing in lysis buffer (1% NP-40, 0.1% SDS, 100 mM
NaCl, 50 mM Tris at pH 7.5, 10 mM EDTA, 0.1% PMSF,
supplemented with Roche complete protein inhibitor) using a
micropestle, then centrifuged at 15,000 rpm in a microcentri-
fuge for 10 min at 4°C. The supernatent was transferred to a new
tube and stored at −80°C.
Primary sympathetic neurons were subjected to NGF with-
drawal for 24 h as described above. The cells were then fixed in
4% PFA, permeabilized with 0.1% sodium citrate and 0.1%
Triton X-100, blocked with 10% goat serum in PBS, and incu-
bated with the KIF1B antibody (1:100 dilution) or EglN3 anti-
body (1:100) in PBS containing 0.1% Triton X-100. After incu-
bation with anti-rabbit Alexa 488 (Molecular Probes) and stain-
ing with DAPI, images were acquired using a confocal laser
imaging system (LSM 510; Carl Zeiss MicroImaging, Inc.) at
Tumor sample sequencing
Ninety-eight primary neuroblastoma tumor samples were iden-
tified from the Children’s Oncology Group (COG) Neuroblas-
toma Nucleic Acids Bank. Samples were collected after obtain-
ing parental informed consent, and institutional review board
(IRB) guidelines were followed for the procurement of each
sample. They were obtained at original diagnosis from patients
who had received no previous treatment and immediately snap-
frozen, and had a tumor cell content of >90% based on differ-
ential count, clonal hyperdiploid percentage in some tumors,
and direct examination of H&E-stained tumor slides. All 98
patients met the COG criteria for having high-risk disease
(Maris 2005). Patients were staged according to the Interna-
tional Neuroblastoma Staging System and histology was ana-
lyzed using the Shimada Pathology Classification (Shimada et
al. 1984; Brodeur et al. 1993). Loss-of-heterozygosity (LOH) sta-
tus was determined using conventional microsatellite markers
and high-resolution SNP array analysis as described previously
(George et al. 2007). DNA was extracted using conventional
methods (Qiagen kit) and was sequenced by Agencourt, Inc., by
automated sequencing. All 46 coding KIF1B? exons were PCR-
amplified and sequenced in a duplicate, bidirectional manner.
Sequence traces were analyzed to identify potential somatic
mutations using the Mutation Surveyor software package (Soft-
An additional 13 neuroblastoma tumors from the COG, as
described above, and 14 medulloblastoma samples obtained at
Children’s Hospital in Boston under IRB approval were se-
quenced for KIF1B? at the Broad Institute. Briefly, DNA was
extracted from the tumor and matched normal blood sample
(Qiagen DNeasy kit), quantified using picogreen (Molecular
Probes), and isothermally amplified using the Repli-g whole-
genome amplification kit (Amersham). Five nanograms of DNA
for each exon of KIF1B? were individually PCR-amplified
(primer sequences available upon request) with the HotStar En-
zyme (Qiagen) and the following cycling parameters: one cycle
of 15 min at 95°C; followed by 35 cycles of 20 sec at 95°C, 30 sec
at 60°C, and 1 min at 72°C; followed by a final extension of 3
min at 72°C. PCR products were sequenced in a duplicate, bi-
directional manner. Sequence traces were analyzed to identify
potential somatic mutations using an automated analysis pipe-
line comprised of the commercial software package Mutation
Surveyor (SoftGenetics), PolyPhred 3.5 (Nickerson et al. 1997),
and PolyDHAN (D. Richter, pers. comm.). The KIF1B? S34SL
variant was confirmed by Sequenom mass spectrometric geno-
Fifty-two pheochromocytoma or paraganglioma samples were
used to sequence the KIF1B? gene under an IRB-approved pro-
tocol. Fragments were obtained from the core of the tumor and
contained >70% tumor cells. Samples with a clear adjacent cor-
tical component were macrodissected. Specimens were snap-
frozen at the time of surgical resection and stored at −70°C or in
liquid nitrogen until processed. Diagnosis of pheochromocy-
toma and/or paraganglioma was confirmed by histology in every
case. Eleven of these tumors came from individuals with he-
reditary disease (two MEN2A, one MEN2B, one VHL, two fa-
milial paraganglioma syndromes type 4-PGL4/SDHB, and five
familial cases without an identifiable primary mutation in
pheochromocytoma susceptibility genes). The remaining 41 tu-
mors were sporadic or had an unknown familial history. Four
tumors were recurrent or malignant (the latter were defined by
the detection of metastasis at nonchromaffin sites), while the
others were considered benign or had short follow-up.
Two approaches were used for sequencing these samples. Ge-
nomic DNA was isolated from 36 tumors using standard meth-
ods (Qiagen). Ten nanograms were used to amplify 50 amplicons
spanning the 46 coding exons and exon–intron boundaries of the
KIF1B? gene (primer sequences available upon request). For 16
tumors, only cDNA (prepared using Applied Biosystems
KIF1B? acts downstream from EglN3
GENES & DEVELOPMENT891
Reverse Transcription kit) was available. Twenty primer pairs
spanning the entire coding region of the longest KIF1B? tran-
script were used for PCR and sequencing of these samples.
PCR was performed using HotMaster Enzyme (Eppendorf).
PCR conditions were as follows: one cycle of 5 min at 95°C;
followed by 35 cycles of 30 sec at 95°C, 30 sec at 59°C, and 45
sec at 72°C; followed by a final extension of 5 min at 72°C. PCR
products were purified and sequenced in both directions by
Agencourt Bioscience using dye terminator technology. Se-
quence traces were analyzed using the commercial software
Mutation Surveyor (SoftGenetics) and were manually verified.
Variants were confirmed by the sequence of an independent
The copy number of KIF1B? was determined by real-time
PCR. Pooled results from three reference housekeeping genes
with distinct genomic locations (?2 microglobulin, Albumin,
and TRIM43) were used to calculate the copy number using the
??Ctmethod as described previously. Primer sequences are
available upon request.
Frequency of KIF1B? S34L, E646V, T827I, P1217S, S1481N,
E1628K, Y1087C, andV1554M in 270 controls of diverse ethnic
backgrounds was determined by Sequenom mass spectrometric
genotyping of the HapMap collection of normal DNA (Thoris-
son et al. 2005).
We thank Robert Freeman, Jacques Pouyssegur, Peter Ratcliffe,
Bert Vogelstein, and Geoffrey Wahl for valuable reagents; Amit
Dutt for sharing unpublished data; Regeneron Pharmaceuticals
for EglN3−/−mice; Arthur Young for help with real-time PCR
assays; and members of the Kaelin Laboratory for useful discus-
sions. S.S. is supported by grants from Charles A. King Trust,
Charles H. Hood Foundation, and the VHLFA foundation;
P.L.D. is supported by grants from the Sidney Kimmel Cancer
Foundation and the San Antonio Cancer Institute; R.K., B.D.C.,
A.T.L., and W.G.K. are supported by grants from NIH. W.G.K. is
also supported by the Doris Duke Foundation and is an HHMI
Aloyz, R.S., Bamji, S.X., Pozniak, C.D., Toma, J.G., Atwal, J.,
Kaplan, D.R., and Miller, F.D. 1998. p53 is essential for de-
velopmental neuron death as regulated by the TrkA and p75
neurotrophin receptors. J. Cell Biol. 143: 1691–1703.
Bagchi, A., Papazoglu, C., Wu, Y., Capurso, D., Brodt, M., Fran-
cis, D., Bredel, M., Vogel, H., and Mills, A.A. 2007. CHD5 is
a tumor suppressor at human 1p36. Cell 128: 459–475.
Benn, D.E., Dwight, T., Richardson, A.L., Delbridge, L., Bam-
bach, C.P., Stowasser, M., Gordon, R.D., Marsh, D.J., and
Robinson, B.G. 2000. Sporadic and familial pheochromocy-
tomas are associated with loss of at least two discrete inter-
vals on chromosome 1p. Cancer Res. 60: 7048–7051.
Berns, K., Hijmans, E.M., Mullenders, J., Brummelkamp, T.R.,
Velds, A., Heimerikx, M., Kerkhoven, R.M., Madiredjo, M.,
Nijkamp, W., Weigelt, B., et al. 2004. A large-scale RNAi
screen in human cells identifies new components of the p53
pathway. Nature 428: 431–437.
Berra, E., Benizri, E., Ginouves, A., Volmat, V., Roux, D., and
Pouyssegur, J. 2003. HIF prolyl-hydroxylase 2 is the key oxy-
gen sensor setting low steady-state levels of HIF-1? in nor-
moxia. EMBO J. 22: 4082–4090.
Brodeur, G.M. 1994. Molecular pathology of human neuroblas-
tomas. Semin. Diagn. Pathol. 11: 118–125.
Brodeur, G.M., Sekhon, G., and Goldstein, M.N. 1977. Chromo-
somal aberrations in human neuroblastomas. Cancer 40:
Brodeur, G.M., Pritchard, J., Berthold, F., Carlsen, N.L., Castel,
V., Castelberry, R.P., De Bernardi, B., Evans, A.E., Favrot, M.,
Hedborg, F., et al. 1993. Revisions of the international crite-
ria for neuroblastoma diagnosis, staging, and response to
treatment. J. Clin. Oncol. 11: 1466–1477.
Brummelkamp, T.R., Bernards, R., and Agami, R. 2002a. A sys-
tem for stable expression of short interfering RNAs in mam-
malian cells. Science 296: 550–553.
Brummelkamp, T.R., Bernards, R., and Agami, R. 2002b. Stable
suppression of tumorigenicity by virus-mediated RNA inter-
ference. Cancer Cell 2: 243–247.
Caren, H., Ejeskar, K., Fransson, S., Hesson, L., Latif, F., Sjoberg,
R.M., Krona, C., and Martinsson, T. 2005. A cluster of genes
located in 1p36 are down-regulated in neuroblastomas with
poor prognosis, but not due to CpG island methylation. Mol.
Cancer 4: 10. doi: 10.1186/1476-4598-4-10.
Cheng, N.C., Van Roy, N., Chan, A., Beitsma, M., Westerveld,
A., Speleman, F., and Versteeg, R. 1995. Deletion mapping in
neuroblastoma cell lines suggest two distinct tumor suppres-
sor genes in the 1p36 region, only one of which is associated
with N-myc amplification. Oncogene 10: 291–297.
Fairchild, R., Kyner, J., Hermreck, A., and Schimke, R. 1979.
Neuroblastoma, pheochromocytoma, and renal cell carci-
noma. Occurrence in a single patient. JAMA 242: 2210–2211.
George, R.E., Attiyeh, E.F., Li, S., Moreau, L.A., Neuberg, D., Li,
C., Fox, E.A., Meyerson, M., Diller, L., Fortina, P., et al. 2007.
Genome-wide analysis of neuroblastomas using high-den-
sity single nucleotide polymorphism arrays. PLoS ONE 2:
e255. doi: 10.1371/journal.pone.0000255.
Haag, M.M., Soukup, S.W., and Neely, J.E. 1981. Chromosome
analysis of a human neuroblastoma. Cancer Res. 41: 2995–
Ichimiya, S., Nimura, Y., Kageyama, H., Takada, N., Sunahara,
M., Shishikura, T., Nakamura, Y., Sakiyama, S., Seki, N.,
Ohira, M., et al. 1999. p73 at chromosome 1p36.3 is lost in
advanced stage neuroblastoma but its mutation is infre-
quent. Oncogene 18: 1061–1066.
Johnson, M.R., Look, A.T., DeClue, J.E., Valentine, M.B., and
Lowy, D.R. 1993. Inactivation of the NF1 gene in human
melanoma and neuroblastoma cell lines without impaired
regulation of GTP.Ras. Proc. Natl. Acad. Sci. 90: 5539–5543.
Kenchappa, R.S., Zampieri, N., Chao, M.V., Barker, P.A., Teng,
H.K., Hempstead, B.L., and Carter, B.D. 2006. Ligand-depen-
dent cleavage of the P75 neurotrophin receptor is necessary
for NRIF nuclear translocation and apoptosis in sympathetic
neurons. Neuron 50: 219–232.
Kimmel, C.B., Ballard, W.W., Kimmel, S.R., Ullmann, B., and
Schilling, T.F. 1995. Stages of embryonic development of the
zebrafish. Dev. Dyn. 203: 253–310.
Knudson Jr., A.G. and Meadows, A.T. 1976. Developmental ge-
netics of neuroblastoma. J. Natl. Cancer Inst. 57: 675–682.
Knudson Jr., A.G. and Strong, L.C. 1972. Mutation and cancer:
Neuroblastoma and pheochromocytoma. Am. J. Hum.
Genet. 24: 514–532.
Krona, C., Ejeskar, K., Abel, F., Kogner, P., Bjelke, J., Bjork, E.,
Sjoberg, R.M., and Martinsson, T. 2003. Screening for gene
mutations in a 500 kb neuroblastoma tumor suppressor can-
didate region in chromosome 1p; mutation and stage-spe-
cific expression in UBE4B/UFD2. Oncogene 22: 2343–2351.
Lee, M., Draoui, M., Zia, F., Gazdar, A., Oie, H., Bepler, G.,
Bellot, F., Tarr, C., Kris, R., and Moody, T. 1992. Epidermal
growth factor receptor monoclonal antibodies inhibit the
Schlisio et al.
892GENES & DEVELOPMENT
growth of lung cancer cell line. J. Natl. Cancer Inst. Monogr. Download full-text
Lee, S., Nakamura, E., Yang, H., Wei, W., Linggi, M.S., Sajan,
M.P., Farese, R.V., Freeman, R.S., Carter, B.D., Kaelin Jr.,
W.G., et al. 2005. Neuronal apoptosis linked to EglN3 prolyl
hydroxylase and familial pheochromocytoma genes: Devel-
opmental culling and cancer. Cancer Cell 8: 155–167.
Lipscomb, E., Sarmiere, P., Crowder, R., and Freeman, R. 1999.
Expression of the SM-20 gene promotes death in nerve
growth factor-dependent sympathetic neurons. J. Neuro-
chem. 73: 429–432.
Maris, J.M. 2005. The biologic basis for neuroblastoma hetero-
geneity and risk stratification. Curr. Opin. Pediatr. 17: 7–13.
Nagai, M., Ichimiya, S., Ozaki, T., Seki, N., Mihara, M., Furuta,
S., Ohira, M., Tomioka, N., Nomura, N., Sakiyama, S., et al.
2000. Identification of the full-length KIAA0591 gene encod-
ing a novel kinesin-related protein which is mapped to the
neuroblastoma suppressor gene locus at 1p36.2. Int. J. On-
col. 16: 907–916.
Nakamura, E. and Kaelin Jr., W.G. 2006. Recent insights into
the molecular pathogenesis of pheochromocytoma and para-
ganglioma. Endocr. Pathol. 17: 97–106.
Nangaku, M., Sato-Yoshitake, R., Okada, Y., Noda, Y., Take-
mura, R., Yamazaki, H., and Hirokawa, N. 1994. KIF1B, a
novel microtubule plus end-directed monomeric motor pro-
tein for transport of mitochondria. Cell 79: 1209–1220.
Nickerson, D.A., Tobe, V.O., and Taylor, S.L. 1997. PolyPhred:
Automating the detection and genotyping of single nucleo-
tide substitutions using fluorescence-based resequencing.
Nucleic Acids Res. 25: 2745–2751.
Ohira, M., Kageyama, H., Mihara, M., Furuta, S., Machida, T.,
Shishikura, T., Takayasu, H., Islam, A., Nakamura, Y., Ta-
kahashi, M., et al. 2000. Identification and characterization
of a 500-kb homozygously deleted region at 1p36.2–p36.3 in
a neuroblastoma cell line. Oncogene 19: 4302–4307.
Opocher, G., Schiavi, F., Vettori, A., Pampinella, F., Vitiello, L.,
Calderan, A., Vianello, B., Murgia, A., Martella, M., Tac-
caliti, A., et al. 2003. Fine analysis of the short arm of chro-
mosome 1 in sporadic and familial pheochromocytoma.
Clin. Endocrinol. (Oxf.) 59: 707–715.
Palmada, M., Kanwal, S., Rutkoski, N.J., Gustafson-Brown, C.,
Johnson, R.S., Wisdom, R., and Carter, B.D. 2002. c-jun is
essential for sympathetic neuronal death induced by NGF
withdrawal but not by p75 activation. J. Cell Biol. 158: 453–
Peeper, D.S., Shvarts, A., Brummelkamp, T., Douma, S., Koh,
E.Y., Daley, G.Q., and Bernards, R. 2002. A functional screen
identifies hDRIL1 as an oncogene that rescues RAS-induced
senescence. Nat. Cell Biol. 4: 148–153.
Schwab, M., Praml, C., and Amler, L.C. 1996. Genomic insta-
bility in 1p and human malignancies. Genes Chromosomes
Cancer 16: 211–229.
Shimada, H., Chatten, J., Newton Jr., W.A., Sachs, N., Hamoudi,
A.B., Chiba, T., Marsden, H.B., and Misugi, K. 1984. Histo-
pathologic prognostic factors in neuroblastic tumors: Defi-
nition of subtypes of ganglioneuroblastoma and an age-
linked classification of neuroblastomas. J. Natl. Cancer Inst.
Sommer, L. and Rao, M. 2002. Neural stem cells and regulation
of cell number. Prog. Neurobiol. 66: 1–18.
Stoler, A. and Bouck, N. 1985. Identification of a single chro-
mosome in the normal human genome essential for suppres-
sion of hamster cell transformation. Proc. Natl. Acad. Sci.
Straub, J.A., Lipscomb, E.A., Yoshida, E.S., and Freeman, R.S.
2003. Induction of SM-20 in PC12 cells leads to increased
cytochrome c levels, accumulation of cytochrome c in the
cytosol, and caspase-dependent cell death. J. Neurochem. 85:
Takeda, O., Homma, C., Maseki, N., Sakurai, M., Kanda, N.,
Schwab, M., Nakamura, Y., and Kaneko, Y. 1994. There may
be two tumor suppressor genes on chromosome arm 1p
closely associated with biologically distinct subtypes of neu-
roblastoma. Genes Chrom. and Cancer 10: 30–39.
Tatekawa, Y., Muraji, T., Nishijima, E., Yoshida, M., and
Tsugawa, C. 2006. Composite pheochromocytoma associ-
ated with adrenal neuroblastoma in an infant: A case report.
J. Pediatr. Surg. 41: 443–445.
The, I., Murthy, A.E., Hannigan, G.E., Jacoby, L.B., Menon,
A.G., Gusella, J.F., and Bernards, A. 1993. Neurofibromato-
sis type 1 gene mutations in neuroblastoma. Nat. Genet. 3:
Thorisson, G.A., Smith, A.V., Krishnan, L., and Stein, L.D.
2005. The International HapMap Project Web site. Genome
Res. 15: 1592–1593.
Wang, Q., Diskin, S., Rappaport, E., Attiyeh, E., Mosse, Y., Shue,
D., Seiser, E., Jagannathan, J., Shusterman, S., Bansal, M., et
al. 2006. Integrative genomics identifies distinct molecular
classes of neuroblastoma and shows that multiple genes are
targeted by regional alterations in DNA copy number. Can-
cer Res. 66: 6050–6062.
Westerfield, M. 1993. The zebrafish book. University of Oregon
Press, Eugene, OR.
White, P.S., Thompson, P.M., Gotoh, T., Okawa, E.R., Igarashi,
J., Kok, M., Winter, C., Gregory, S.G., Hogarty, M.D., Maris,
J.M., et al. 2005. Definition and characterization of a region
of 1p36.3 consistently deleted in neuroblastoma. Oncogene
Yang, H.W., Chen, Y.Z., Takita, J., Soeda, E., Piao, H.Y., and
Hayashi, Y. 2001. Genomic structure and mutational analy-
sis of the human KIF1B gene which is homozygously deleted
in neuroblastoma at chromosome 1p36.2. Oncogene 20:
Zhao, C., Takita, J., Tanaka, Y., Setou, M., Nakagawa, T., Ta-
keda, S., Yang, H.W., Terada, S., Nakata, T., Takei, Y., et al.
2001. Charcot-Marie-Tooth disease type 2A caused by mu-
tation in a microtubule motor KIF1B?. Cell 105: 587–597.
Zhu, Y. and Parada, L.F. 2002. The molecular and genetic basis
of neurological tumours. Nat. Rev. Cancer 2: 616–626.
KIF1B? acts downstream from EglN3
GENES & DEVELOPMENT893