Human netrin-G1 isoforms show evidence of differential expression
Joanne M.A. Meerabuxa, Hisako Ohbaa, Masayuki Fukasawaa,., Yumiko Sutob,
Mika Aoki-Suzukia, Toshiaki Nakashibac, Sachiko Nishimurac,
Shigeyoshi Itoharac, Takeo Yoshikawaa,*
aLaboratory for Molecular Psychiatry, RIKEN Brain Science Institute, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan
bDepartment of Pediatric Cardiology, Tokyo Women’s Medical University, Tokyo 162-8666, Japan
cLaboratory for Behavioral Genetics, RIKEN Brain Science Institute, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan
Received 10 November 2004; accepted 5 April 2005
Available online 17 May 2005
The recently identified netrins-G1 and -G2 form a distinct subgroup within the UNC-6/netrin gene family of axon guidance molecules. In
this study, we determined the size and structure of the exon/intron layout of the human netrin-G1 (NTNG1) and -G2 (NTNG2) genes.
Northern analysis of both genes showed limited nonneuronal but wide brain expression, particularly for NTNG2. Reverse transcriptase PCR
detected nine alternatively spliced isoforms including four novel variants of NTNG1 from adult brain. A semiquantitative assay established
that major expression was restricted to isoforms G1c, G1d, G1a, and G1e in the brain and to G1c in the kidney. There is also evidence of
developmental regulation of these isoforms between fetal and adult brain. In conclusion, NTNG1 may use alternative splicing to diversify its
function in a developmentally and tissue-specific manner.
D 2005 Elsevier Inc. All rights reserved.
Keywords: Laminet-1; Laminet-2; Alternative splicing; Differential expression; Semiquantitative RT-PCR
The developing nervous system is dependent on the
actions of various secreted factors and membrane proteins
that allow axons to find their correct targets. The netrin
family was originally defined by netrin 1 and netrin 2, which
were isolated from vertebrates [1–3]. They are secreted
proteins, structurally related to the short arms of laminin g
. Netrin 4 is also a secreted protein, but is more similar to
the laminin h chains [5,6]. Recently, two diverged mole-
cules, netrin-G1 (Ntng1) and netrin-G2 (Ntng2), also called
laminet-1 and laminet-2, respectively, have been identified
from the mouse and included as family members [7–9].
These molecules differ from classical netrins by three main
features: (1) the presence of a glycosyl phosphatidylinositol
lipid (GPI) site for membrane anchorage, (2) the generation
of multiple isoforms, and (3) the failure to bind classical
netrin receptors [7–9]. No orthologues for these genes have
been found in Caenorhabditis elegans or Drosophila
melanogaster, prompting the suggestion that netrins-G1
and -G2 may provide a function in cell architecture that is
unique to vertebrates. Supporting this theory is the finding
that netrin-G1 shows genetic association with schizophrenia
. We set out to complete the genomic mapping and
cDNA structure determination for human netrin-G1 and all
its isoforms. We also examined the tissue distribution of
mRNA for NTNG1 and NTNG2, as well as the expression of
NTNG1 alternatively spliced transcripts.
First, mouse netrin-G1a cDNA sequence NM_030699
and G1d sequence AB038664, as well as human netrin-G1
expressed sequence tag (EST) clones BC030220 and
AB023193, were aligned with human genomic BAC clones
RP11-270C12, RP11-396N10, and RP11-436H6 (GenBank
Accession Nos. AC114491, AL590427, and AL513187,
respectively), using the NCBI BLAST 2 sequences algo-
0888-7543/$ - see front matter D 2005 Elsevier Inc. All rights reserved.
* Corresponding author. Fax: +81 48 467 7462.
E-mail address: firstname.lastname@example.org (T. Yoshikawa).
Genomics 86 (2005) 112 – 116
These comparisons identified a total of 10 exons for NTNG1
(Fig. 1), with the intron/exon boundaries shown in Supple-
mentary Fig. 1. All of the splice junctions conform to the
basic GT/AG rule although there are deviations from the
extended consensus sequence starting at exon 4. Translation
starts in exon 2 and encodes a maximally sized predicted
protein of 581 amino acids. The context of the initiator
methionine (GAAUUUAGAAUGU, in which A of the AUG
represents position 1) maintains the ?3 and ?9 positions
relative to the Kozak consensus (GCCGCGA/GCCAUGG)
. Six potential start sites, all in a poor context, lie
upstream. This suggests that NTNG1 may be regulated at the
translational stage. Translation of each exon sequence
revealed that exons 2 and 3 code for the VI domain, exons
4 and 5 for the V-1 domain, exon 6 and 7 for the unknown
domain , exons 8 and 9 for domains V-2 and V-3,
respectively, and exon 10 for the CVdomain, which contains
the GPI anchor (Fig. 1).
Northern analysis using NTNG1 exon 3 as a probe
detected two groups of transcripts. The upper band at 4.4 kb
may represent a partially spliced transcript, AB023193,
which comprises spliced exons 1 to 5 with a retained intron
5. The lower diffuse band at 3.2 kb should represent the
alternatively spliced transcripts. Both sized transcripts are
present in the brain and kidney with possible weak
expression in the spleen, liver, small intestine, placenta,
and lung (Fig. 2). In the brain, expression was strongest in
the cerebral cortex followed by the occipital pole, frontal
lobe, temporal lobe, and putamen. Strong expression has also
been detected in the thalamus and inferior colliculus .
For the determination of NTNG2 structure, the mouse
netrin-G2 cDNA sequence NM_133500.1 and a human
NTNG2 EST sequence NM_032536.1 were aligned with
human BAC clones RP11-479K20, RP11-203M2, RP11-
5N16, and RP11-738I14 (AL159997.14, AL353701.15,
AL353631.17, and AL354735.14, respectively) in the
manner previously stated. This analysis detected eight exons
Fig. 1. Genomic organization and isoforms of NTNG1. (A) Upper diagram shows the genomic layout of NTNG1 exons and their encoded domains. The lower
diagram shows the exons that encode the truncated transcript AB023193. The span of each domain is indicated by an inverted ‘‘T’’. Brackets denote intron sizes.
Colored boxes denote coding exons, and white boxes, untranslated regions. (B) Exonic composition of the nine alternatively spliced isoforms and the database-
derived human transcript AB023193. Asterisks denote newly detected isoforms. Diagram is not to scale.
J.M.A. Meerabux et al. / Genomics 86 (2005) 112–116
for NTNG2 (data shown in Supplementary Fig. 2, since it
was superseded by the sequence OTTHUMG00000020835
located at http://www.ensembl.org/Homo_sapiens/). All
splice junctions conform to the GT/AG rule but few conserve
the longer consensus sequence (Supplementary Fig. 1). The
initiator methionine context maintains the ?6 and essential
?3 position (UGCGCAGCCAUGC, in which A of the AUG
represents position 1) and three upstream start sites also
maintain the ?3 position, suggestive of translational control
as for NTNG1 . Exons 2 and 3 code for domain VI, and
exons 4, 5, 6, and 7 code for domains V-1, V-2, V-3, and CV ,
respectively, producing a predicted protein of 530 amino
acids (Supplementary Fig. 2). An exon 3-derived probe
identified two diffuse bands, spanning 4.4 to 3.6 kb, in the
brain and peripheral blood leukocytes (Fig. 2). A diffuse
band of approximately 3.6 kb was observed in heart and
skeletal muscle with very low signal in placenta and lung
tissues. In the brain, two transcripts were observed in all
regions tested, with the possible exception of the cerebellum
and medulla, where only the smaller sized transcript was
present. This differential detection of NTNG2 transcripts
suggests alternative splicing of the gene although this feature
was not examined in this study.
A fluorescently labeled probe derived from exons 1 to 5
of netrin-G1 showed a single hybridization to chromosome 1
at p13.3 (Supplementary Fig. 3), in keeping with its database
assignment. A netrin-G2 probe (exons 2–6) showed
fluorescent signal at 9q34 in all cells, with weak signal at
12q24.3 in 20% of the cells (Supplementary Fig. 3). The
probe sequence showed no significant similarity to NTN4,
which maps to 12q22–q23, nor to related family members
NTN1 and NTN2L (17p13.1 and 16p13.3, respectively)
[5,12,13], indicating a novel but related sequence of netrin-
G2 at this locus.
Previous studies reported six mouse Ntng1 isoforms that
retain the GPI lipid anchor [7,9]. To detect human alternative
splice forms, we performed RT-PCR analysis using adult
whole brain-derived cDNA (Clontech, Palo Alto, CA, USA)
with primers designed to exons 4 and 10. These exons were
chosen because they flank the variable region observed in the
mouse and select for transcripts containing the GPI lipid
anchor . Amplicons were cloned and sequenced to
determine their precise size and exonic composition. A total
of nine human isoforms denoted as G1a–G1o, in keeping
with the original mouse nomenclature [7,9], were identified
(Fig. 1). Mouse isoforms G1f, G1h, G1i, and G1j were not
detected in this study since they lack the GPI domain-
encoding exons. Isoforms G1a, G1b, G1c, G1e, and G1f
have already been detected in the mouse [8,9]. The large
number of NTNG1 isoforms detected in human brain reflects
findings that between 40 and 60% of genes undergo
alternative splicing, with the brain showing the highest
levels of exon skipping [14,15]. The exon skipping seen in
NTNG1 results in essentially the same amino acid sequence
for each included domain but with single amino acid changes
in some isoforms at points where exon fusions form a new
triplet (Supplementary Fig. 4). These spliced variants are
supported by the semiquantitative detection of differential
Fig. 2. Northern analyses of NTNG1 and NTNG2. The blots were purchased from Clontech. Probes for both genes were derived from exon 3 (NTNG1-F,
GTTTGATTTTGAAGGAAGACA, 3Vend at nt 338, -R, AAAAACGCGAACCTGTC, 3Vend at nt 680, BC030220; and NTNG2-F, CCACCTACTGGCA-
GAGCATCA, 3Vend at nt 331, -R, CGATGTTGGAGATGGCGTAGAAGT, 3Vend at nt 821, AB058760). h-Actin was used as a control probe.
J.M.A. Meerabux et al. / Genomics 86 (2005) 112–116
expression between the brain and the kidney (Fig. 3). The
isoform amplicons were identified by direct comparison of
the computer-generated sizes with previously sequenced
clones. The relative expression of each isoform was
calculated by first normalizing total expression of netrin-
G1 in each cDNA pool (using primers to exon 3) and then
determining the ratio of each isoform relative to the total
isoform pool. In the kidney, isoform G1c is the major
transcript, with low levels of G1l. Isoform G1c was shown to
bind to the netrin-G1 ligand (NGL-1), promoting the
outgrowth of thalamic neurons in culture . Interestingly,
Northern analysis shows no expression of NGL1 in the
kidney , so in this tissue, G1c appears to require a
differentligand. Althoughno human locus forkidney disease
with mouse chromosome 3, where a modifier locus for renal
vascular disease lesions has been identified .
Additionally, there is evidence of developmental regu-
lation of NTNG1 in the brain (Fig. 3). Of the nine isoforms
detected in adult brain, only five isoforms were present in
fetal brain, with G1c and G1d being the major species. Fetal
brain shows higher levels of G1d and the minor transcript
G1e relative to adult brain, as well as lower levels of G1c and
the minor transcript G1a. While G1a most closely resembles
the classical netrins in structure, G1c is the most abundantly
expressed isoform and it differs from G1d by the inclusion of
an unknown domain coded for by exons 6 and 7. These
exons have not been previously identified in human NTNG1
ESTs, but are conserved among mouse, rat, and chicken. No
specific sequences or regulatory elements that drive alter-
native splicing in NTNG1 have been identified to date, nor
have any mutations been associated with specific isoforms.
The truncated AB023193 encodes a protein that would
terminate after domain V-1. Its easy detection by Northern
analysis suggests that this transcript may not undergo
nonsense-mediated decay . It remains to be seen how
or whether this isoform, which is similar in sequence to G1c,
plays a role in cellular migration.
Netrin-G1 null animals show reduced prepulse inhibition,
strongly suggesting that the gene may be important in the
maintenance of neuronal plasticity associated with sensory
motor gating and or cognitive functioning . Reduced
prepulse inhibition is a behavioral paradigm associated with
schizophrenia , raising the possibility that NTNG1 plays
a role in the pathophysiology of this disease. Indeed, an
association study from our laboratory suggests that NTNG1
does show association with schizophrenia, possibly via a
mechanism that alters the ratio of isoform expression .
This could be analogous to pathologies associated with the
tau protein. Mutations that affect splicing, particularly the
Fig. 3. Tissue-dependent expression patterns of NTNG1 isoforms. Left shows an Agilent gel (Agilent Technologies, Palo Alto, CA, USA) containing PCR
products representing alternatively spliced NTNG1 isoforms derived from human fetal brain (22 weeks), adult brain, and adult kidney (Clontech). The exon 4
primer is F, CCCCATCCCCAAAGGCACTGC, and exon 10 primer, R, GTCGGAGCCGCAGCTGCCAGC. The isoform bands were distinguished by size
comparison with previously sequenced clones. The bar chart shows the relative expression of the major NTNG1 isoforms detected in human fetal brain, adult
brain, and adult kidney. For the semiquantitative assay, PCR amplicons were loaded onto an Agilent 2100 Bioanalyzer utilizing a DNA 1000 LabChip kit
(Agilent Technologies). The relative expression of each isoform was calculated by determining the fraction of each isoform from the total quantity of detected
isoforms. Values are an average of two independent experiments.
J.M.A. Meerabux et al. / Genomics 86 (2005) 112–116
inclusion of exon 10, lead to skewed ratios of tau isoforms Download full-text
and, by multistep mechanisms, to several neurodegenerative
disorders such as Alzheimer disease [22,23]. Further
analysis is necessary to determine the cellular specificity
of NTNG1 isoforms and their roles in the development of
sensory motor systems.
We thank Dr. Kazuo Yamada for advice with this project.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.ygeno.2005.
 T.E. Kennedy, T. Serafini, J.R. de la Torre, M. Tessier-Lavigne,
Netrins are diffusible chemotropic factors for commissural axons in
the embryonic spinal cord, Cell 78 (1994) 425–435.
 T. Serafini, et al., The netrins define a family of axon outgrowth-
promoting proteins homologous to C. elegans UNC-6, Cell 78 (1994)
 A.W. PuschelPuschel, Divergent properties of mouse netrins, Mech.
Dev. 83 (1999) 65–75.
 H. Hutter, et al., Conservation and novelty in the evolution of
cell adhesion and extracellular matrix genes, Science 287 (2000)
 M. Koch, et al., A novel member of the netrin family, beta-netrin,
shares homology with the beta chain of laminin: identification,
expression, and functional characterization, J. Cell Biol. 151 (2000)
 Y. Yin, J.R. Sanes, J.H. Miner, Identification and expression of mouse
netrin-4, Mech. Dev. 96 (2000) 115–119.
 T. Nakashiba, et al., Netrin-G1: a novel glycosyl phosphatidylinositol-
linked mammalian netrin that is functionally divergent from classical
netrins, J. Neurosci. 20 (2000) 6540–6550.
 T. Nakashiba, S. Nishimura, T. Ikeda, S. Itohara, Complementary
expression and neurite outgrowth activity of netrin-G subfamily
members, Mech. Dev. 111 (2002) 47–60.
 Y. Yin, J.H. Miner, J.R. Sanes, Laminets: laminin- and netrin-related
genes expressed in distinct neuronal subsets, Mol. Cell. Neurosci. 19
 M. Fukasawa, et al., Case–control association study of human netrin
G1 gene in Japanese schizophrenia, J. Med. Dent. Sci. 51 (2004)
 M. Kozak, Initiation of translation in prokaryotes and eukaryotes,
Gene 234 (1999) 187–208.
 J.A. Meyerhardt, et al., Netrin-1: interaction with deleted in colorectal
cancer (DCC) and alterations in brain tumors and neuroblastomas, Cell
Growth Differ. 10 (1999) 35–42.
 T.J. Van Raay, et al., The NTN2L gene encoding a novel human netrin
maps to the autosomal dominant polycystic kidney disease region on
chromosome 16p13.3, Genomics 41 (1997) 279–282.
 B. Modrek, C. Lee, A genomic view of alternative splicing, Nat.
Genet. 30 (2002) 13–19.
 G. Yeo, D. Holste, G. Kreiman, C.B. Burge, Variation in alternative
splicing across human tissues, Genome Biol. 5 (2004) R74.
 J.C. Lin, W.H. Ho, A. Gurney, A. Rosenthal, The netrin-G1 ligand
NGL-1 promotes the outgrowth of thalamocortical axons, Nat.
 T.H. Le, et al., Modifier locus on mouse chromosome 3 for renal
vascular pathology in AT1A receptor-deficiency, Hypertension 43
 J. Lykke-Andersen, mRNA quality control: marking the message for
life or death, Curr. Biol. 11 (2001) R88–R91.
 S. Nishimura, K. Seki, A. Watanabe, K. Niimi, T. Yoshikawa, T.
Nakashiba, et al., Axonal netrin-G1 regulates NMDA-receptor depend-
ent synaptic functions in selected neural circuits, Abstr. Soc. Neurosci.
 M.A. Geyer, K. Krebs-Thomson, D.L. Braff, N.R. Swerdlow,
Pharmacological studies of prepulse inhibition models of sensorimotor
gating deficits in schizophrenia: a decade in review, Psychopharma-
cology (Berlin) 156 (2001) 117–154.
 M. Aoki-Suzuki, et al., A family-based association study and gene
expression analyses of netrin-G1 and -G2 genes in schizophrenia,
Biol. Psychiatry 57 (2005) 382–393.
 L. Buee, T. Bussiere, V. Buee-Scherrer, A. Delacourte, P.R. Hof, Tau
protein isoforms, phosphorylation and role in neurodegenerative
disorders, Brain Res. Brain Res. Rev. 33 (2000) 95–130.
 C.J. Lee, K. Irizarry, Alternative splicing in the nervous system: an
emerging source of diversity and regulation, Biol. Psychiatry 54
J.M.A. Meerabux et al. / Genomics 86 (2005) 112–116