Oncogene (1998) 16, 1813 ± 1823
 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00
http://www.stockton-press.co.uk/onc
Murine NIMA-related kinases are expressed in patterns suggesting distinct
functions in gametogenesis and a role in the nervous system
Eli Arama1, Amiel Yanai1, Gillar Kil®n1, Alan Bernstein2 and Benny Motro1
1
Department of Life Sciences, Bar-Ilan University, Ramat-Gan 52900, Israel and 2Program in Molecular Biology and Cancer,
Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Ontario, Canada
NIMA protein kinase is a major regulator of progression
into mitosis in Aspergillus nidulans. Dominant negative
forms of NIMA protein prevent entrance into mitosis in
HeLa cells, suggesting that mammals have a similar
pathway. We have reported previously the isolation of a
murine NIMA-related kinase, designated Nek1, and
more recently several additional NIMA-related human
kinases have been cloned. The existence of several
mammalian NIMA-related genes raises the questions of
whether the di€erent mammalian members have redundant, overlapping or distinct functions, and whether
these functions are related to the role of NIMA in
controlling mitosis. To address these questions we have
studied the expression patterns of the di€erent murine
nek genes. To this end, we isolated a murine nek2 cDNA
and compared its patterns of expression, during both
gametogenesis and embryogenesis, to those of nek1. Both
genes were highly expressed in developing germ cells,
albeit in distinct patterns. In both females and males,
nek1 is expressed much earlier than nek2, suggesting
only limited ability for functional redundancy. Surprisingly, a striking speci®city of nek1 expression was found:
high levels of nek1 RNA were observed in distinct
regions of the nervous system, most notably in neurons of
the peripheral ganglia. These patterns suggest that the
di€erent mammalian NIMA-related kinases participate
in di€erent phases of the meiotic process and may also
have functions other than cell cycle control.
Keywords: NIMA; protein kinase; mitosis; meiosis;
nervous system
Introduction
The progression through critical checkpoints in the cell
cycle is controlled in eukaryotes by the activity of
cyclin-dependent kinases (CDKs) (reviewed by Pines,
1995; Elledge, 1996). The activity of the CDKs is
regulated at the di€erent stages of the cell-cycle
through reversible phosphorylation, and by binding
of the CDKs to di€erent cyclins, activators and
inhibitors (Tsai et al., 1994; Lew et al., 1994;
Morgan, 1995; Pines, 1995; Sherr and Roberts, 1995).
The onset of mitosis is controlled by p34cdc2 (CDK1),
which form with its regulator, a B-type cyclin, a
complex known as maturation, or mitosis promoting
Correspondence: B Motro
Received 21 April 1997; revised 5 November 1997; accepted 5
November 1997
factor (MPF) (Nurse and Bissett, 1981; Lohka et al.,
1988; Gautier et al., 1988; Arion et al., 1988; Simanis
and Nurse, 1986; Minshull et al., 1989). However,
although the pivotal role of p34cdc2 during G2/M
transition in eukaryotic cells is well established, several
studies demonstrate that entry into mitosis is also
regulated by additional mechanisms that may involve
additional pathway(s) (Amon et al., 1992; Sorger et al.,
1992; Stueland et al.,1993).
In the ®lamentous fungus, Aspergillus nidulans,
initiation of mitosis requires not only p34cdc2 activation, but also activation of an unrelated serine/
threonine protein kinase termed NIMA, for Never In
Mitosis A (Osmani et al., 1988, 1991). Inducible
overexpression of nima triggers premature entrance
into mitosis from any point in the cell cycle (Osmani et
al., 1991; Lu and Means, 1994). A temperature
sensitive nima allele arrests cells in G2 at the restrictive
temperature, although p34cdc2 is fully active, and upon
shifting to the permissive temperature the block is
released, and the cells enter mitosis rapidly and
synchronously (Osmani et al., 1988, 1991). The p34cdc2
and NIMA pathways are not totally independent of
each other as it has been demonstrated that while
inactivation of p34cdc2 causes cells to arrest in G2,
NIMA is in a phosphorylated and partially activated
form (Ye et al., 1995). Activation of p34cdc2 releases the
block and leads to hyperphosphorylation of NIMA,
boosting its activity (Ye et al., 1995). These results
suggest a model for NIMA activation which includes
two steps of phosphorylation with only the second step
of activation lies downstream of p34cdc2. Thus, it has
been suggested that activation of both kinases, p34cdc2
and NIMA, is required for cells to initiate mitosis in
Aspergillus nidulans (for recent review see Fry and
Nigg, 1995; Osmani and Ye, 1996).
The structural and functional conservation of many
cell cycle-regulatory pathways among all eukaryotes
studied (Pines, 1995), raises the possibility that a
NIMA-like pathway may function in vertebrates as
well. Indeed, overexpression of NIMA promotes
premature meiotic and mitotic events in Xenopus
oocytes and HeLa cells, respectively (Lu and Hunter,
1995). In addition, expression of dominant negative
forms of nima induce a G2 arrest in HeLa cells (Lu and
Hunter, 1995). These results provide strong evidence
for the existence of NIMA-like kinases in vertebrate
cells. Consistent with this prediction, we have reported
previously the cloning of a murine NIMA-related
kinase designated Nek1 (Letwin et al., 1992), and
more recently cloning of additional members of the
Nek (NIMA-related kinase) family, have been reported, including the human genes Nek2, Nek3 and
STK2 (Schultz and Nigg, 1993; Cance et al., 1993;
Expression of murine NIMA-related kinases
E Arama et al
1814
Levedakou et al., 1994; Schultz et al., 1994). The
members of the mammalian Nek family share unique
motifs in their kinase domain, as well as general
structure. All have the catalytic kinase domain at the
N-terminus, and contain a long C-terminal domain.
Although there are no de®nite biological function
ascribed to any of the mammalian Nek kinases, our
previous work has demonstrated high levels of nek1
transcripts in male and female germ cells, suggesting a
role in meiosis (Letwin et al., 1992), while human Nek2
kinase activity in tissue culture cells ¯uctuates during
the cell cycle, suggesting a role in mitosis (Schultz et
al., 1994).
In the present study, we describe the isolation of a
cDNA encoding a novel murine NIMA-like protein
kinase which corresponds to a mouse homologue of
human Nek2. Expression analysis shows that, similar
to nek1, nek2 is preferentially expressed in the gonads.
Extensive comparison of the gonadal patterns of
expression of nek2 to those of nek1 revealed that
nek2 transcripts appears later than nek1 during
gametogenesis and are expressed in a narrower
spectrum of cells. Surprisingly, we found that in
addition to the expression in meiotic cells, nek1 is
very highly and speci®cally expressed in postmitotic
peripheral and motor neurons. Taken together, these
results suggest that di€erent Nek kinases regulate
di€erent phases of the meiotic process and have
speci®c functions in certain tissues.
1a), yielding a 2869 bp sequence. The combined
sequence begins with a long open reading frame. The
®rst in-frame ATG codon (beginning at nucleotide 58,
a
b
Results
Isolation and characterization of mouse nek2 cDNA.
To identify novel murine nima-related genes, we used a
reverse transcriptase polymerase chain reaction (RT ±
PCR) cloning strategy based on conserved sequences
amongst the NIMA family. Within the amino acid
sequence of the known members of the NIMA family
(Letwin et al., 1992; Schultz et al., 1994; Osmani et al.,
1988), there are two stretches of highly conserved
amino acids characteristic of NIMA-related catalytic
domains. At the N-terminaus is a highly conserved
sequence motif (L/M)(R/K)HPNI (residing in domain
IV), while the C-terminus includes the motif GTP(F/
Y)Y(M/L) (domain VIII). We designed degenerate
primers based on these sequences (see Materials and
methods, and Figure 1a), and performed RT ± PCR on
poly(A)+ RNAs from di€erent sources. Sequencing of
cloned fragments revealed several novel cDNA
sequences, corresponding to NIMA-related and unrelated murine protein kinase catalytic domains (to be
described elsewhere). One of the putative NIMArelated protein kinases, derived from adult testis
RNA, will be described here.
The original RT ± PCR fragment was 365 bp long.
To clone the entire coding region, we screened adult
mouse testis cDNA libraries by DNA hybridization
using this fragment as a probe. Four clones, containing
inserts from 1.1 to 2.64 kb, were isolated (Figure 1a),
subcloned, and sequenced (see Materials and methods).
All the mouse cDNAs isolated in this screen encoded
overlapping polypeptides with the hallmarks of protein
kinases. A composite sequence of the cDNA was
constructed from two clones, mt29 and mt210 (Figure
Figure 1 (a) Alignment of murine nek2 cDNA clones used to
construct the composite cDNA sequence. The structure of the
composite nek2 cDNA is shown above the cDNA clones, with an
open reading frame (black bar), and 5' and 3' untranslated
sequences (solid lines) indicated. Also indicated are the relative
position of the PCR product (PCRp) used to screen the mouse
cDNA testis library, and the relative positions of the probes used
for in situ localization of nek2 transcripts (ISp1, ISp2). Restriction
endonuclease sites relevant for subclonings are indicated (S, SacI;
H, HindII; E, EcoRI; B, BglII). (b) Composite cDNA sequence
and the expected translation product of nek2. The single letter
amino acid code is used. The numbers on the right represent the
nucleotide number in the cDNA (top line) and the amino acid
position in the coding sequence (bottom line). Highly conserved
residues among protein kinases catalytic domain are underlined.
The initiation and termination codons are also underlined
Expression of murine NIMA-related kinases
E Arama et al
1815
nek1 and nek2 di€er in their patterns of expression
during gametogenesis
30 dpn
21 dpn
17 dpn
14 dpn
— nek2
12 dpn
28S —
10 dpn
To determine the tissue pattern of nek2 expression, we
®rst performed Northern blot analyses using a DNA
probe generated from the 3' untranslated region
(nucleotides 1505-2869, Figure 1a). Under stringent
hybridization conditions, only one RNA species of
4.2 kb was observed in all tissues examined. The levels
of nek2 RNAs were much higher in adult testis than in
any other tissues examined (Figure 2). nek2 transcripts
were also detected in total RNA isolated from early
and late developing embryos, and from adult intestine,
spleen and ovary but not in adult brain and kidney
(Figure 2).
The Northern RNA blot analysis identi®ed the testis
as the major site of nek2 expression. As our previous
work showed that nek1 was also expressed at high
levels in the testis (Letwin et al., 1992), we compared
the spatial distribution of the expressing cells in this
organ. This comparison took advantage of the well
characterized development of the prepuberal murine
testis (Bellve, 1977). At birth, the testis contains only
nonproliferating gonocytes, and shortly thereafter,
spermatogenesis is initiated. Spermatogenesis occurs
in three phases: mitotic, meiotic and spermiogenesis.
The mitotic phase starts by postnatal day 6 with the
appearance of the ®rst primitive type A spermatogonia,
followed by type B spermatogonia on day 8. The
meiotic phase starts at around day 10 with the
appearance of the ®rst preleptotene/leptotene spermatocytes, which develop into zygotene (by day 12) and
then pachytene spermatocytes (by day 14). The ®rst,
and immediately after, the second meiotic division,
starts at day 18, and subsequently postmeiotic round
spermatid enter the spermiogenesis phase in which the
spermatids condense and di€erentiate into spermatozoa. Using this information we followed the expression
of nek1 and nek2 in testes of di€erent developmental
ages by Northern and RNA in situ hybridization in
order to de®ne the stages in which the germ cells
express these genes.
Northern blot analysis with nek2 and nek1 probes
on RNA extracted from testes of successive postnatal
7 dpn
Kidney
Testis
Ovary
Brain
Placenta
Spleen
Intestine
13.5 dpc
8.5 dpc
Figure 1b) was preceded by a sequence conforming to
the Kozak consensus sequence (Kozak, 1984), suggesting that this ATG serves as an initiator codon. The
large open reading frame (nucleotides 58-1386) encodes
a predicted 443 a.a. protein (Figure 1b) with a
calculated molecular mass of 51.3 kDa, and a
calculated isoelectric point of 9.2. The cDNA contains
1483 bp of 3' untranslated sequences and lacks a
polyadenylation signal and poly(A) tail.
A search of the GenBank and EMBL databases
with the composite sequence revealed that the
predicted kinase was novel and most similar to
human Nek2 serine/threonine kinase (Schultz et al.,
1994), sharing 93% amino acid identity over the
catalytic domain, and 80.8% identity over the Cterminal noncatalytic domain. This high degree of
similarity suggests that the murine gene is a
homologue of human nek2, and we therefore
designated it mouse nek2.
The mammalian Nek2 catalytic domain contained
12 subdomains and several amino acid residues
(underlined in Figure 1b) recognized as being nearly
invariant throughout the protein kinase superfamily
(Hanks and Hunter, 1995). The mammalian Nek2
catalytic domain includes motifs unique to the
NIMA-related family, including the motifs (R/
K)HPNIV (sd-IV), LY(L/I)(Y/V)M(E/D)YCXGGDL
(sd-V), GTP(F/Y)YX(S/P)PE (sd-VIII) and CX(M/
L)YELC (sd-IX). Although the murine and human
Nek2 proteins are most closely related to NIMA
amongst the known mammalian kinases (46.3%
identical over the catalytic domain), these two
proteins display limited sequence similarity over the
C-terminal noncatalytic domain (27.9% identity), and
this domain is much shorter in Nek2 compared to
NIMA, Nek1, Nek3 and STK2.
nek2
18S —
nek1
— GAPDH
Figure 2 Northern blot analysis of nek2 expression in embryonic
and adult tissues. 12 mg of total RNA extracted from the
indicated murine tissues were blotted on a Magna nylon
membrane (MSI), and hybridized to the nek2 probe indicated in
Figure 1a (see Materials and methods). The placental RNA was
extracted from 12.5 d.p.c. embryos. The migration of the 28S and
18S ribosomal RNAs is indicated at the left. The same blot was
rehybridized to GAPDH probe for a loading control
28S rRNA
Figure 3 Northern blot comparison of nek2 and nek1 levels of
expression in the developing testis. 10 mg of total RNA were
extracted from testes at the indicated days after birth, blotted on
a Magna nylon membrane (MSI) and hybridized to nek2 or nek1
DNA probes. For nek1 only the 4.4 kb band is shown as the
6.5 kb band show the same pattern. Note that the rise in nek2
levels is delayed until 17 days after birth while nek1 levels already
rise at 10 days after birth. The ethidium bromide staining is
shown as a loading control
Expression of murine NIMA-related kinases
E Arama et al
1816
ages showed an obvious di€erence in the temporal
patterns of expression of the two kinases. nek2 levels
remained low throughout the mitotic and early meiotic
steps and the levels of the RNA peaked up sharply
toward the late pachytene stage (by day 17) (Figure 3).
In contrast, nek1 levels were low at the mitotic phase,
raised at the beginning of the meiotic phase (by day
10), and remained constant to adulthood (Figure 3).
To identify the testicular cells expressing nek1 and
nek2 we carried out in situ RNA hybridization to
sections of testes at di€erent postnatal ages. To allow
direct comparison of the levels of expression, all the
sections from the di€erent ages were embedded in a
single cryomold and processed on a single slide.
Detection of nek2 transcripts was done using two
di€erent 35S-labeled antisense RNA probes spanning
either the C-terminal noncatalytic domain and
untranslated region, or only the 3'-untranslated region
(probes ISp1 and ISp2, respectively; Figure 1a).
Essentially identical patterns were obtained by both
probes, con®rming the speci®city of the signals. The
nek1 probe included the end of the `tail' and the 3'
noncoding region as described previously (Fragment 3,
in Letwin et al., 1992). The RNA in situ analysis
con®rmed and extended the Northern data. For both
nek1 and nek2 the hybridization signal was con®ned to
germ cells undergoing mitosis or meiosis. At the
mitotic phase (day 7), low levels of nek1 and nek2
were found in spermatogonia of all tubules (not
shown). However, nek2 was expressed in much later
stages of spermatogenesis and in a narrower spectrum
of cells relative to nek1. The onset of high levels of
nek2 expression was delayed until 17 days postnataly,
when about 20% of the tubules were moderately
labeled (Figure 4a,b). By day 21, moderate levels of
nek2 were observed in most of the tubules (Figure
4c,d). In adult testis very high levels were con®ned to
about half of the tubules, while the others had much
fainter signal (Figure 4e,f). Higher magni®cation, and
staging of the seminiferous tubules, revealed that late
pachytene-diplotene spermatocytes expressed the highest levels of nek2 (stage Xl in Figure 4g,h). Moderate
levels were found in earlier pachytene spermatocytes
(stage IX, and stage V-VI tubules in Figure 4g,h). The
levels of nek2 expression diminished around the
formation of round spermatids (Figure 4g,h), and
nek2 transcripts were not observed in spermatozoa
stored in the epididymis (not shown). No signal was
found over Sertoli or interstitial Leydig cells. To
con®rm that the expression of nek2 is con®ned to
germ cells, we determined the expression of nek2 in
atrichosis mutant mice. Mice homozygous for the
atrichosis (at) locus lack germ cells, while having
Sertoli and Leydig cells (Hummel, 1964). No signi®cant
signal was found in testis from mice homozygous for
this mutation (Figure 5a,b), while their heterozygous
littermates showed the wildtype patterns of nek2
expression (not shown).
In contrast to nek2, by day 12 about 60% of the
tubules already expressed high levels of nek1, and by
17, 21 days, and in adult testis, all the tubules
expressed high levels of nek1 (Figure 5d,e, and not
shown). Higher magni®cation revealed that nek1 is
highly expressed in germ cells from the early
spermatocyte stage to round spermatids, and its levels
diminished only at the transition from round to
elongated spermatids (Letwin et al., 1992, and data
not shown).
In the female reproductive system, mitosis of the
germ cells is already completed by mid-embryogenesis
and meiosis follows immediately. The germ cells enter
the leptotene and subsequently zygotene stages of
meiosis by 14 days post coitum (d.p.c.), and by 16
d.p.c. most of the germ cells are in the pachytene stage
(Borum, 1961). In early postnatal females, oocytes are
arrested in the diplotene stage, and the growth and
maturation of the oocyte is dependent on the cyclic
release of gonadotropic hormones, starting at puberty.
We have described previously high levels of nek1 RNA
in 15.5 d.p.c. embryonic ovary (Letwin et al., 1992). In
contrast, nek2 transcripts were not detectable at this
stage (not shown). In the adult ovary, nek2 expression
was highly restricted to oocytes of medium to large
follicles (staging according to Peters, 1969), and a
much lower signal was detected over proliferating
granulosa cells (Figure 5f,g). A similar pattern was
observed previously for nek1 RNA expression in the
adult ovary (Letwin et al., 1992).
nek1 is highly and speci®cally expressed in motor and
peripheral neurons
To determine the distribution of nek1 and nek2
transcripts during embryogenesis RNA in situ hybridization was performed on sections of embryos at
di€erent developmental stages. During early embryogenesis (7.5 ± 9.5 d.p.c.), nek2 was widely expressed
throughout the embryo. At later stages of embryogenesis, nek2 was expressed at low levels in some
correlation with the proliferation state of the tissue
examined. For example, in the embryonic central
nervous system, nek2 expression was mainly con®ned
to ventricular zones within the spinal cord and brain,
consistent with expression in proliferating cells (not
shown). In agreement with the Northern analysis, nek2
expression was not detected by RNA in situ analysis in
adult brain (not shown). In contrast, signi®cant
expression was observed in the nasal epithelium
(Figure 5h,i).
Only faint expression of nek1 was observed in
homogenous pattern in early embryos and embryonic
diploid placenta aged 7.5 to 9.5 d.p.c. (not shown).
However, by 11.5 d.p.c. localized nek1 expression was
observed in the main sites of embryonic nek1
expression, including the cranial ganglia (trigeminal,
and the ninth/tenth inferior complex) (Figure 6a,b), the
ventral side of the spinal cord, and the di€erentiating
dorsal root ganglia (not shown). As further development of the central and peripheral nervous system
proceeds, nek1 expression at these sites became more
apparent.
Coronal sections through 11.5 d.p.c. spinal cord
suggest that nek1 expression in the ventrolateral side
was localized over the putative motor neuroblasts (not
shown). Similar sections through 13.5 and 14.5 d.p.c.
embryos con®rmed that nek1 expression in the spinal
cord was con®ned to the motor column (Figure 6c,d
and not shown). Interestingly, a peak in localized nek1
expression at 13.5 to 14.5 d.p.c. embryos was followed
by less intense and less localized signal over the entire
ventral horn at 15.5 to 17.5 d.p.c. embryos, and in
newborns scattered neurons all over the spinal cord
Expression of murine NIMA-related kinases
E Arama et al
1817
Figure 4 nek2 expression in the developing testis. (a, c, e,) Bright- and (b, d, f) dark-®eld view of sections of 17 (a, b), 21 (c, d) days
postnatal (d.p.n.), and adult (e, f) testis hybridized to nek2 antisense RNA probe. Only few tubules hybridized strongly to nek2 at
17 d.p.n. (a, b), while higher proportion of the tubules express very high levels at 21 d.p.n. (c, d) and in adult (e, f) testis. The low
expressing tubules at 21 days testis (marked by arrows in c, d) are less mature tubules. (g, h) Higher magni®cation of the adult testis
section shown in e, f. Highest levels of nek2 mRNA levels are seen in diplotene spermatocytes in stage XI tubules. Lower levels are
seen in stage IX pachytene spermatocytes and lowest in earlier pachytene spermatocytes (in stage V ± VI tubules). No grains are seen
over round spermatids (rs), elongated spermatids (es), Sertoli (S) and interstitial Leydig (L) cells. The calibration bar in a indicates
100 mm, and a ± f are of the same magni®cation. In g the bar indicates 50 mm
Expression of murine NIMA-related kinases
E Arama et al
1818
Figure 5 nek2 expression in at/at mutant testis, adult ovary and embryogenesis, and nek1 expresson in adult testis. (a) Bright- and
(b) dark-®eld view of adult at/at testis hybridized to nek2 antisense probe. The tubules do not contain germ cells, and no speci®c
signal is seen. (c) Adult testis hybridized to nek2 sense RNA (control) probe. (d) Bright- and (e) dark-®eld view nek1 expression in
adult testis. Note that all tubules express quite similar, high levels of nek1. (f) Bright- and (g) dark-®eld view of adult ovary
hybridized to nek2 antisense probe. Silver grains are prominent over large oocytes (O). (h) Bright- and (i) dark-®eld view of a cross
section of the nasal system of a 14.5 d.p.c. embryo. Signi®cant levels of nek2 expression are con®ned to the nasal epithelium (n.e.).
(n.cp.) nasal capsule; (n.s.) nasal septum. The bars in a, d and h indicate 100 mm, and in f 50 mm
Expression of murine NIMA-related kinases
E Arama et al
(not shown). Neurons at the motor cranial ganglia also
expressed nek1. For example, nek1 transcripts were
detected at high levels in the facial motor nucleus of
16.5 d.p.c. embryo (Figure 6e,f).
Low levels of nek1 were observed in the entire fetal
brain at all stages of embryonic development, without
noticeable regional localization (excluding the motor
nuclei mentioned above) (Figure 6e ± h). However,
there was localized expression of nek1 in the adult
brain. For example, high levels of nek1 were observed
over the pyramidal cells of the hippocampus, Purkinje
neurons of the cerebellum, and tufted cells of the
olfactory bulb (Figure 7c ± f). In contrast, very low
levels were seen in the granular layers of the olfactory
bulb and the cerebellum, and in the striatum (Figure
7c ± f, and not shown). nek1 transcripts were not
observed in glial speci®c regions such as layer 1 of
the cerebral cortex (molecular layer) and ®mbria
hippocampus (Figure 7c,d, and not shown).
From 11.5 d.p.c. onward, all peripheral ganglia
surveyed expressed nek1, albeit at varying levels. High
levels of nek1 transcripts were observed in dorsal root,
cranial, sympathetic and parasympathetic ganglia,
while lower levels were observed in enteric ganglia.
As shown in Figure 6c,d, dorsal root ganglia, but not
their nerve bundles, expressed high levels of nek1.
Although the neurons of the dorsal root ganglia
withdraw from the cell cycle at about 14 d.p.c. (Sims
and Vaughn, 1979), these cells both at 16.5 d.p.c.
(Figure 6g,h), and in newborn dorsal root ganglia
neurons (not shown), expressed very high levels of
nek1. Within the newborn ganglion, due to the
di€erence in size and staining properties between
neurons and glial cells, the neuronal identity of the
expressing cells was unambiguously determined (not
shown).
Cranial ganglia can be classi®ed according to their
embryonic origin ± neural crest or placode. Interestingly, both neural crest-derived (superior ninth and
tenth), and placode-derived (inferior ninth and tenth,
and the acoustic) ganglia expressed nek1 (Figure 6e,f).
High levels were also present in the mixed trigeminal
ganglion (Figure 6a,b,e,f). Figure 6g,h also shows
strong signals over sympathetic trunk ganglia and
bronchial parasympathetic ganglia. In both sympathetic and parasympathetic ganglia, the high levels of
nek1 expression were maintained in newborn postmitotic neurons (not shown and Figure 7a,b).
Figure 6 In situ analysis of embryonic nek1 expression. (a) Bright- and (b) dark-®eld view of a sagittal section through the anterior
region of 11.5 d.p.c. mouse embryo. Expression is restricted to the trigeminal (Vg), acoustic (VIIIg) and inferior 9'th-10'th (IX ± Xgi)
cranial ganglia, and the otic vesicle (ov). (c) Bright- and (d) dark-®eld view of a transverse section through the trunk of 14.5 d.p.c.
embryo hybridized to nek1 antisense probe. (mc) motor column; (rg) dorsal root ganglia; (st) sympathetic trunk. (e) bright- and (f)
dark-®eld views of a sagittal section through the neck-head region of 17 d.p.c. embryo hybridized to nek1 probe. (VIIm) facial
motor ganglion; (IXg i,s) inferior (petrosal) and superior 9'th ganglion. (Xg i,s) inferior (nodose) and superior (jugular) 10'th
ganglion; (AH) anterior hypothalamus; (cd) cochlear duct; (md) medulla; (po) pond. (g) Bright- and (h) dark-®eld view of a sagittal
section through 17 d.p.c. embryo hybridized to nek1 probe. (he) heart; (ki) kidney; (li) liver; (lu) lung; (ne) nasal epithelium; (rg)
dorsal root ganglia; (sc) spinal cord; (sp) sphenopalatine ganglion; (st) sympathetic trunk. The arrow heads in h points to
parasympathetic bronchi ganglia. The calibration bar in a indicates 250 mm, and a ± f are of the same magni®cation. In g the bar
indicates 1 mm
1819
Expression of murine NIMA-related kinases
E Arama et al
1820
Consistent with the close developmental relationship
between adrenal chroman cells and sympathetic
neurons (Anderson and Axel, 1986), the embryonic
and adult adrenal also expressed nek1 (Figure 6g,h,
and not shown). The neurons and glial cells of the
autonomic enteric ganglia are derived from neural
crest cells which migrate to the gut from the vagal
and sacral regions of the neuraxis (Le Douarin,
1986). The ganglia comprise the myenteric plexus
between the longitudinal and circular muscles, and
closer to the lumen the submucous plexus. Moderate
levels of nek1 RNA were detected in the myenteric
plexus and lower levels in the submucous plexus (not
shown).
Discussion
The mammalian Nek kinases constitute a distinct
subfamily within the kinase superfamily and share
stretches of characteristic amino acids, as well as
general structure, with each other, and with the fungal
NIMA. To gain insight into the roles these genes may
serve in mammals, we isolated the murine nek2 gene
and compared its pattern of expression to that of nek1
during both gametogenesis and embryonic development.
The processes of mammalian gametogenesis in males
and females share a general reductive division strategy,
while di€ering largely in their time schedule. In murine
Figure 7 nek1 expression in newborn and adult nervous system: (a) Bright- and (b) dark-®eld view of a cross section through the
accessory genital glands of one week old male mouse. The expression is restricted to parasympathetic ganglia (psg). (vd) vas
deference. (c) Bright- and (d) dark-®eld view of the posterior side of a transverse section through adult brain. Note the high signal
over the hypocampal pyramidal neurons (hp), subiculum (su), enthorhinal cortex (en) and Purkinje (Pu) cell layer, and the absence
of signal over cerebellar granular cells (gr) and the nonneuronal, layer 1 of the enthorihnal cortex (marked by *). (e) Bright- and (f)
dark-®eld view of a transverse section through the olfactory bulb. The expression is restricted to a band just super®cial to the mitral
cell layer (mi) (probably internal tufted neurons), and a band over the internal side of the periglomerular region (probably external
tufted neurons). Note the lack of signal over the olfactory granular region. The calibration bar in a indicates 250 mm, and a ± d are
of the same magni®cation. In e the bar indicates 100 mm
Expression of murine NIMA-related kinases
E Arama et al
males, the mitotic phase commences at around day 6
after birth and is continuously followed by the meiotic
phase. In contrast, in females, oogonia proliferate only
in the fetus, and begin meiosis I during late
embryogenesis. At birth, the oocytes are arrested at
late prophase of meiosis I (at diplotene), a stage
resembling the G2 phase of the mitotic division cycle.
Beginning at puberty, some of the arrested oocytes are
recruited into a growth phase and resume progress
through division I of meiosis. In both systems, the
onset of nek1 expression occurs earlier than that of
nek2. In the testis, nek1 levels accumulate at around
postnatal day 12 at the beginning of the leptotene/
zygotene stage, while a sharp rise in nek2 levels is
delayed until just prior to the commencement of the
meiotic divisions, at postnatal day 17 (late pachytene
stage). Similarly, nek1 is expressed in 15.5 d.p.c.
embryonic ovary (zygotene-early pachytene stages),
while no signi®cant nek2 expression is found at this
stage. In later stages of gametogenesis, the patterns
overlap and both genes are highly expressed in late
pachytene and diplotene spermatocytes, and in the
ovary from medium size oocyte to (at least) ovulation.
The early onset of nek1 expression is concurrent with
the condensation of the chromosomes at leptotene.
Intriguingly, NIMA overexpression leads to premature
chromatin condensation, even in the absence of cdc2
activity, and it has therefore been proposed that a
major role of NIMA in controlling entrance into
mitosis is through the regulation of chromatin
condensation (O'connell et al., 1994). The temporal
expression of nek2 in the gonads is reminiscent of the
expression patterns observed for several other genes
implicated in the control of entry into mitosis. For
example, cdc2 RNA and MPF kinase activity, are
highest in pachytene spermatocytes (Chapman and
Wolgemuth, 1994). The levels of Cdc25C RNA,
encoding a member of the Cdc25 phospatases which
dephosphorylate and activate p34cdc2 kinase, peak in
germ cells between the late pachytene spermatocyte
and round spermatid stages (Wu and Wolgemuth,
1995). The genes for the regulatory subunits of p34cdc2,
cyclinB2 and cyclinB1, are expressed in late pachytene
spermatocytes and in round spermatids, respectively
(Chapman and Wolgemuth, 1993).
Interestingly, nek1 and nek2 are expressed to only
low levels in mitotically dividing cells. For example,
both kinases are expressed to low levels in proliferating
spermatogonia compared to meiotic spermatocytes.
These results are somewhat surprising, given the high
similarity of these kinases to the cell cycle regulator
NIMA. Other genes involved in the regulation of the
cell cycle, including cdc2, cycB1 and cycB2, are also
expressed at higher levels in meiotic cells than in
mitotic cells (Chapman and Wolgemuth, 1993, 1994).
Several explanations for this ®nding can be suggested:
First, di€erent posttranscriptional control mechanisms
may operate in mitotic and meiotic cells, resulting in
comparable levels of kinase activity in the two systems.
Second, the kinases may have a role in both mitosis
and meiosis and lower levels of kinase activity may be
sucient to execute the mitotic role. It is also possible
that a rise in kinase activity is involved in signaling the
transition from the mitotic to the meiotic phase. A
similar quantitative model has recently been proposed,
correlating the levels of cdc2 ± cdc13 activity to the
progression through the cell cycle in ®ssion yeast (Stern
and Nurse, 1996). And third, Nek1 and/or Nek2 may
not have a role in the mitotic cell cycle. For Nek2, this
option is less likely as it has been reported that human
nek2 RNA, and kinase activity, ¯uctuate in HeLa cells
in a cell cycle-dependent way (Fry et al., 1995).
Interestingly, high levels of Nek2 activity were found
in HeLa cells in S, and G2 and lower levels in M and
G1, suggesting either a role in DNA synthesis or a
preparatory role for mitosis. The testicular system in
which DNA synthesis in spermatogonia is separated
temporally and spatially from the meiotic division of
the spermatocytes serves as an excellent model system
to distinguish between the two processes. The low
levels of nek2 RNA in spermatogonia and the sharp
rise of nek2 levels in late spermatocytes favor a role for
Nek2 in the meiotic divisions rather than in the S
phase. However, to gain better insight into the
possible roles of these genes in meiosis, the protein
pro®les of expression should also be determined.
Surprisingly, nek1 expression is not con®ned to
mitotic and meiotic cells. Very high expression was
observed in peripheral and motor neurons. Much lower
levels were also observed in other parts of the central
nervous system. This expression is not dependent on
the embryonic origin of the ganglia as both neural crest
and placode-derived peripheral ganglia exhibited
similar high levels. Quite unexpectedly, high levels of
nek1 RNA were detected in postmitotic neurons.
Moreover, we were unable to detect signi®cant
expression in actively dividing, immature neurons as
migrating neural crest cells or cells located in the
periventricular zone of the developing spinal cord,
suggesting that the onset of nek1 expression is
concurrent with the time that speci®c neurons withdraw from the cell cycle and di€erentiate. This
expression pattern suggests that Nek1 kinase is also
involved in postmitotic processes in speci®c neurons.
Although these data do not provide a direct indication
of the precise function(s) of Nek1 in the nervous
system, it is noteworthy that the two major sites of
nek1 expression in the nervous system, peripheral and
motor neurons, have in common very long axons
suggesting that Nek1 kinase might be involved in some
aspects of axon growth and maintenance.
In conclusion, the present study demonstrates that
the nek1 and nek2 genes are di€erentially expressed
during meiosis, and that nek1 is also expressed in
speci®c neurons. This situation is reminiscent of the
CDKs cell-cycle regulators. Only one CDK appears to
be involved in cell cycle regulation in budding and
®ssion yeast (Elledge, 1996) while diverged, multiple
members exist in higher eukaryotes. In higher
eukaryotes, the di€erent members play distinct
functions in certain points of the cell cycle (reviewed
by Sherr, 1993; Pines, 1995), and additional role in
postmitotic neurons has been demonstrated for CDK5
(Tsai et al., 1994; Lew et al., 1994), and have also been
suggested for other CDKs (Freeman et al., 1994).
Similarly, only one member of the NIMA family
probably exists in Aspergillus (Osmani and Ye, 1996),
while mammals have at least six NIMA-related genes
(Letwin et al., 1992; Schultz et al., 1994; Lu and
Hunter, 1995). The expression patterns reported here
extend this similarity and suggest distinct functions for
the mammalian NIMA-related kinases in di€erent
1821
Expression of murine NIMA-related kinases
E Arama et al
1822
stages of meiosis and mitosis, as well as an additional
role in di€erentiated neurons.
Materials and methods
RT ± PCR and nek2 cDNA cloning
Reverse transcription of poly (A)+ mRNA from adult testis
was done using random hexamers (Boehringer Mannheim)
as primers. For the PCR reaction, degenerate oligonucleotide primers, designed according to conserved amino acids
in NIMA-related kinases, were synthesized. The sense
primer represented the motif (L/M)KHPNIV [TGAA(A/
G) CA(T/C) CCN AA(A/G) AT(T/C) GT], and the antisense primer the motif GTPYY(L/V/M) [TAN (A/
G)TA (A/G)TANGG NGT NCC]. The PCR was carried
out using Taq DNA polymerase (Boehringer Mannheim)
and an automated thermal cycler (MJ Research). The PCR
cycle was 1 min at 948C, 1 min at 408C and 1 min at 638C
for 36 cycles. The products were separated by electrophoresis on 1.5% agarose gel, extracted using QIAquick Gel
Extraction kit (QIAGEN), and ligated into the T-cloning
Kit based on the pUC57 vector (MBI Fermentase). The
clones were sequenced using the Dye Terminator Cycle
Sequencing Kit (Perkin Elmer) and an Applied Biosystems
Model 373 automated DNA sequencer at the DNA
Sequencing Unit, Biological Services, Weizmann Institute.
One of the clones (PCRp in Figure 1a) encoded a putative
kinase related to the NIMA family and was used as a
probe to isolate cDNA clones from a murine adult testis
cDNA library. Several clones were isolated, and four of
them (Figure 1a), were partially or entirely sequenced. The
reported sequence was con®rmed on both strands.
RNA extraction and Northern analysis
Total RNA was obtained from the various tissues using the
LiCl/urea precipitation method, as described in Au€ray
and Rougeon (1980, Procedure C). The RNA was
separated on 1.2% agarose, and transferred to a Magna
nylon membrane (MSI). The ®lters were hybridized with a
randomly primed nek1, nek2 or GAPDH cDNA probes.
Clone 23 (Figure 1a in Letwin et al, 1992) including the tail
region and 3' untranslated sequences, was used as nek1
probe. The nek2 probe was the Sacl-clone fragment from
clone mt210 (Figure 1a), including 3' untranslated
sequences. Hybridization was done according to standard
procedures (Sambrook et al., 1988), and the ®lters were
washed in 0.16SSC and 0.1% SDS at 608C. The ®lters
were exposed to Kodak X-OMAT ®lm with intensifying
screens.
In situ RNA analysis
In situ hybridization was carried out on embryonic and
adult 10 mm cryosections as previously described (Motro et
al, 1991). To allow comparison of the levels of RNA
expression in the di€erent stages of the prepuberal and
adult testis, all of the testes from the di€erent time points
were embedded in a single cryomold, cosectioned, and
processed on a single slide. Post-hybridization washings
included treatment with 50 mg/ml RNase A at 378C for
30 min, and two stringent washes of 20 min each at 608C in
0.16SSC. The slides were dipped in Kodak NTB-2
emultion, exposed for 5 ± 6 days, developed, and stained
with toluidine blue.
Sequence comparisons
A search for sequences similar to nek2 was conducted
against the GenBanklEMBL library database using the
FASTA program of the University of Wisconsin Genetic
Computer Group (GCG) Sequence Analysis Software
Package. The percentage of similarity between sequences
was determined using the Gap program. Comparison of the
sequence of the murine Nek2 kinase to those of human
Nek2 and NIMA was done using the Pileup program of the
GCG package.
Acknowledgements
We gratefully acknowledge the expert advice from Jeremy
Don on staging the testicullar tubules. We also thank Ron
Wides and Don Katco€ for critical reading of the
manuscript, and Uri Nir for providing testis cDNA
libraries. This research was supported by the Israel Science
Foundation founded by the Israel Academy of Sciences
and Humanities- Center of Excellence Program. BM is an
Israel Cancer Research Fund fellow.
References
Amon A, Surana U, Muro€ I and Nasmyth K. (1992).
Nature, 355, 368 ± 371.
Anderson DJ and Axel R. (1986). Cell, 47, 1079 ± 1090.
Arion D, Meijer L, Brizuela L and Beach D. (1988). Cell, 55,
371 ± 378.
Au€ray C and Rougeon F. (1980). Eur. J. Biochem., 107,
303 ± 314.
Bellve AR, Cavicchia JC, Millette CF, O'brien DA,
Bhatnagar YM and Dym M. (1977). J. Cell Biol., 74,
68 ± 85.
Borum K. (1961). Exp. Cell. Res., 24, 495 ± 507.
Cance WG, Craven RJ, Weiner TM and Liu ET. (1993). Int.
J. Cancer, 54, 571 ± 577.
Chapman DL and Wolgemuth DJ. (1993). Development, 118,
229 ± 240.
Chapman DL and Wolgemuth DJ. (1994). Dev. Biol., 165,
500 ± 506.
Elledge SJ. (1996). Science, 274, 1664 ± 1672.
Freeman RS, Estus S and Johnson EM Jr. (1994). Neuron,
12, 343 ± 355.
Fry AM and Nigg EA. (1995). Curr. Biol., 5, 1122 ± 1125.
Fry AM, Schultz SJ, Bartek J and Nigg EA. (1995). J. Biol.
Chem., 270, 12899 ± 12905.
Gautier J, Norbury C, Lohka M, Nurse P and Maller J.
(1988). Cell, 54, 443 ± 449.
Hanks SK and Hunter T. (1995). FASEB J., 9, 576 ± 596.
Kozak M. (1984). Nucl. Acid Res., 12, 857 ± 870.
Le Douarin NM. (1986). Science, 231, 1515 ± 1522.
Letwin K, Mizzen L, Motro B, Ben-David Y, Bernstein A
and Pawson T. (1992). EMBO J., 11, 3521 ± 3531.
Levedakou EN, He M, Baptist EW, Craven RJ, Cance WG,
Welcsh PL, Simmons A, Naylor SL, Leach RJ, Lewis TB,
Bowcock A and Liu ET. (1994). Oncogene, 9, 1977 ± 1988.
Lew J, Huang Q-Q, Qi Z, Winkfein RJ, Aebersold R, Hunt T
and Wang JH. (1994). Nature, 371, 423 ± 426.
Lohka MJ, Hayes MK and Maller JL. (1988). Proc. Natl.
Acad. Sci. USA, 85, 3009 ± 3013.
Lu KP and Hunter T. (1995). Cell, 81, 413 ± 424.
Lu KP and Means AR. (1994). EMBO J., 13, 2103 ± 2113.
Minshull J, Pines J, Golsteyn R, Standart N, Mackie S,
Colman A, Blow J, Ruderman JV, Wu M and Hunt T.
(1989). J. Cell Sci. Suppl., 12, 77 ± 97.
Expression of murine NIMA-related kinases
E Arama et al
Morgan DO. (1995). Nature, 374, 131 ± 134.
Motro B, van der Kooy D, Rossant J, Reith A and Bernstein
A. (1991). Development, 113, 1207 ± 1221.
Nurse P and Bissett Y. (1981). Nature, 292, 558 ± 560.
O'Connell MJ, Norbury C and Nurse P. (1994). EMBO J.,
13, 4926 ± 4937.
Osmani AH, McGuire SL and Osmani SA. (1991). Cell, 67,
283 ± 291.
Osmani SA, Pu RT and Morris NR. (1988). Cell, 53, 237 ±
244.
Osmani SA and Ye XS. (1996). Biochem. J., 317, 633 ± 641.
Peters H. (1969). Acta Endocrin., 62, 98 ± 116.
Pines J. (1995). Adv. Cancer Res., 66, 181 ± 212.
Russell LD, Ettlin RA, Sinha Hikim AP and Clegg ED.
(1990). Histological and Histopathological Evaluation of
the Testis. Cache River Press.
Sambrook J, Fritsch EF and Maniatis T. (1989). Molecular
Cloning ± A Laboratory Manual. 2nd ed., Cold Spring
Harbor Laboratory Press.
Sims TJ and Vaughn JE. (1979). J. Comp. Neurol., 183, 707 ±
720.
Schultz SJ, Fry AM, Sutterlin C, Ried T and Nigg EA.
(1994). Cell Growth Di€er., 5, 625 ± 635.
Schultz SJ and Nigg EA. (1993). Cell Growth Di€er., 4, 821 ±
830.
Sherr CJ. (1993). Cell, 73, 1059 ± 1065.
Sherr CJ and Roberts JM. (1995). Genes Dev., 9, 1149 ± 1163.
Simanis V and Nurse P. (1986). Cell, 45, 261 ± 268.
Sorger PK and Murray AW. (1992). Nature, 355, 365 ± 368.
Stueland CS, Lew DJ, Cismowski MJ and Reed SI. (1993).
Mol. Cell. Biol., 13, 3744 ± 3755.
Stern B and Nurse P. (1996). Trends Genet., 12, 345 ± 350.
Tsai L-H, Dellale I, Caviness VS Jr, Chae T and Harlow E.
(1994). Nature, 371, 419 ± 423.
Wu S and Wolgemuth DJ. (1995). Dev. Biol., 170, 195 ± 206.
Ye XS, Xu G, Pu RT, Fincher RR, McGuire SL, Osmani AH
and Osmani SA. (1995). EMBOJ., 14, 986 ± 994.
1823