Integrin-dependent neuroblastoma cell adhesion and migration on laminin is regulated by expression levels of two enzymes in the O-mannosyl-linked glycosylation pathway, PomGnT1 and GnT-Vb.

Abstract

O-mannosyl-linked glycans constitute a third of all brain O-linked glycoproteins, and yet very little is understood about their functions. Several congenital muscular dystrophies with central nervous system defects are caused by genetic disruptions in glycosyltransferases responsible for the synthesis of O-mannosyl glycans. The glycosyltransferase GnT-Vb, also known as GnT-IX, is expressed abundantly in the brain and testis and is proposed to be the enzyme that branches O-mannosyl-linked glycans. In this study, we show in a human neuronal model that GnT-Vb expression enhances neurite outgrowth on laminin. GnT-Vb has been shown to perform both N-linked and O-mannosyl-linked glycosylation. To determine if the effect on neurite outgrowth was due to N-linked or O-mannosyl-linked glycosylation by GnT-Vb we suppressed the expression of glycosyltransferases important for the elongation of both N-linked and O-mannosyl-linked glycans using RNA interference. Our results suggest that GnT-Vb and PomGnT1, enzymes involved in the O-mannosyl glycosylation pathway, play an active role in modulating integrin and laminin-dependent adhesion and migration of human neuronal cells.

Full text

Page 1

Research Article
Integrin-dependent neuroblastoma cell adhesion and
migration on laminin is regulated by expression levels of two
enzymes in the O-mannosyl-linked glycosylation pathway,
PomGnT1 and GnT-Vb
Karen L. Abbotta,⁎, Karolyn Troupea, Intaek Leeb, Michael Piercea
aComplex Carbohydrate Research Center and Department of Biochemistry and Molecular Biology, University of Georgia,
Athens, GA 30605, USA
bDepartment of Internal Medicine, Division of Hematology, Washington University Medical School, St. Louis, MO 63110, USA
A R T I C L E I N F O R M A T I O NA B S T R A C T
Article Chronology:
Received 27 March 2006
Revised version received
11 May 2006
Accepted 16 May 2006
Available online 21 June 2006
O-mannosyl-linked glycans constitute a third of all brain O-linked glycoproteins, and yet
very little is understood about their functions. Several congenital muscular dystrophies with
central nervous system defects are caused by genetic disruptions in glycosyltransferases
responsible for the synthesis of O-mannosyl glycans. The glycosyltransferase GnT-Vb, also
known as GnT-IX, is expressed abundantly in the brain and testis and is proposed to be the
enzyme that branches O-mannosyl-linked glycans. In this study, we show in a human
neuronal model that GnT-Vb expression enhances neurite outgrowth on laminin. GnT-Vb
has been shown to perform both N-linked and O-mannosyl-linked glycosylation. To
determine if the effect on neurite outgrowth was due to N-linked or O-mannosyl-linked
glycosylation by GnT-Vb we suppressed the expression of glycosyltransferases important
for the elongation of both N-linked and O-mannosyl-linked glycans using RNA interference.
Our results suggest that GnT-Vb and PomGnT1, enzymes involved in the O-mannosyl
glycosylation pathway, play an active role in modulating integrin and laminin-dependent
adhesion and migration of human neuronal cells.
© 2006 Elsevier Inc. All rights reserved.
Keywords:
Human neuroblastoma cells
Integrin
Glycosyltransferase
Laminin
Cell adhesion
Cell migration
PomGnT1
GnT-Vb
GnT-IX
SH-SY5Y
Introduction
The genomic analysis of various neuronal disorders over the
past few years has shed light on the critical functions of
protein glycosylation during the development of the nervous
system. Human congenital disorders such as Walker Warburg
syndrome (WWS), muscle–eye–brain disease (MEB), and
Fukuyama congenital muscular dystrophy (FCMD) are caused
by inactivation of glycosyltransferases that function in the O-
mannosyl glycan processing pathway [1–4]. These disorders
are characterized by muscular dystrophy with neuronal
abnormalities. Particularly interesting are the defects in
neuronal migration and nerve–muscle adhesion noted in
muscle–eye–brain disease that are caused by inactivating
E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 2 8 3 7 – 2 8 5 0
⁎ Corresponding author.
E-mail address: kabbott@uga.edu (K.L. Abbott).
0014-4827/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.yexcr.2006.05.022
available at www.sciencedirect.com
www.elsevier.com/locate/yexcr

Page 2

mutations in the peptide-O-mannose β-1,2-acetylglucosmi-
nyltransferase known as PomGnT1 [3], the glycosyltransferase
that transfers N-acetylglucosamine (GlcNAc) in beta(1,2)
linkage onto Ser/Thr-O-mannosyl residues, allowing exten-
sion of O-mannosyl glycans. O-linked mannose structures
comprise approximately 25% of the total O-linked glycans in
brain[5].Thesestructureshavebeencharacterizedfromrabbit
and sheep brain and partial structural information has been
obtained for the J1 glycoprotein from the mouse nervous
system [6–8], as well as alpha-dystroglycan in smooth muscle
[9].
The glycosyltransferase termed GnT-Vb, also known as
GnT-IX, is a recently discovered enzyme that is selectively
expressed in neuronal cells and participates in the synthesis
of O-mannosyl glycans [10–13]. This enzyme catalyzes the
transfer of GlcNAc in β(1,6) linkage to the peptide-O-linked-
Man, but only if the β1,2-linked GlcNAc formed by PomGnT1 is
present (Fig. 2A). Studies in vitro reveal that GnT-Vb prefers O-
mannosyl acceptors, but can also transfer to the Man (β1,2)
GlcNAc found in non-galactosylated N-linked glycans during
biosynthesis [10]. However, its homolog, GnT-Va, is only
capable of transferring GlcNAc in β(1,6) linkage to N-linked
glycans. GnT-Vb activity is expected to be required for the
expression of the majority of both O-linked polylactosamine
and HNK-glycans in brain [10], both of which function in
neuronal cell adhesion. Since GnT-Vb shows selective expres-
sion during neuromorphogenesis in the mouse (Matthews,
unpublished data), it may function in neuronal cell adhesion
and migration.
Neuroblastoma cells of the N-type (neuronal), such as SH-
SY5Y, are typically loosely adherent to substrates, express
neuronal cell surface markers, can be induced to differentiate,
and are inherently tumorigenic in nude mice [14–17]. In the
present study, SH-SY5Y cells were used to study the potential
role of GnT-Vb glycosylation in the regulation of cell adhesion,
differentiation, and migration. We find that increased GnT-Vb
expression leads to enhanced neurite outgrowth in SH-SY5Y
cells plated on laminin. RNA interference studies show that
reducing the expression levels of two glycosyltransferases
acting in the O-mannosyl glycan pathway, GnT-Vb and
POMGnT-I, using small interfering RNAs [18], leads to
increased adhesion to laminin, as well as reduced migration,
and impaired neurite outgrowth on this extracellular matrix
glycoprotein. The phenotype observed in GnT-Vb suppressed
cells could be rescued by expression of a form of GnT-Vb not
targeted for RNA interference. We also show that suppression
of GnT-Vb expression leads to significant changes in β1
integrin transcript and protein levels. Overall, these data
indicate that GnT-Vb and PomGnT1 expression levels play an
active role in modulating integrin and laminin-dependent
adhesion and migration of human neuronal cells.
Materials and methods
Antibodies and chemicals
ERK1/2 monoclonal antibody and all HRP-labeled anti-mouse
and anti-rabbit IgG were obtained from Santa Cruz Biotech-
nology. Polyclonal antibody against β1 integrin was obtained
from Chemicon. The HA epitope antibody was obtained from
Roche. Laminin from human placenta was obtained from
Sigma. All trans retinoic acid was obtained from Sigma. NHS-
LC–biotin was a product of Pierce, and streptavidin–HRP was
obtained from Vector Laboratory.
Plasmids
Targets for RNA interference were selected using the predic-
tion software available from Oligoengine. The following
oligonucleotide sequences were annealed and cloned into
the pSUPER.retro.neo.gfp vector: VB-coding +: 5′-GATCCCCTC-
CAC GCCTACATCCAGCATTCAAGAGATGCTGGATG-
TAGGCGTGGATTTTTA-3′,
AAATCCACGCCTACATCCAG
GGATGTAG GCGTGGAGGG-3′, VB-3′UTR +: 5′-GATCCC-
CGGTGTCGGACTGC TCAGAGTTCAAGAGACTCTGAGCAGT
CCGACACCTTTTTA-3′, VB-3′UTR-: 5′-AGCTTAAAAA-
GGTGTCGGACTGCTCAGAGTCTCTTGAACTCTGAGCAGT
CCGACACCGGG-3′, Control +: 5′-GATCCCCTTCTCCGA-
ACGTGTCACGTTT CAAGAGAACGTGA CACGT TCGGA-
GAATTTTTA-3′, Control-: 5′-AGCTTA AAAATTCTCCGAACGTG
TCACGTTCTCTTGAAACGTGACACGTTCGGAG AAGGG-3′,
PomGnT1 +: 5′-GATCCCCGCAAAGTTGAGGGCTATGGTTCAAG
AGACCATAGCCCTCAACTTTGCTTTTTA-3′, PomGnT1-: 5′-
AGCTTAAAAAG CAAAGTTGAGGGCTATGGTCTCT TGAACCA-
TAGCCCTCAACTTTGCGGG-3′, GnT-Va +: 5′-GATCCCCGA-
G A A A G C G G A A G A A A G T C T T C A A G A G A G A C T T T
CTTCCGCTTTCTCTTTTTA-3′,
AAAAGAGAAAGCGGAA GAAAGTCTCTCTTGAAGACTTT-
CTTCCGCTTTCTCGGG-3′. The creation of a hemagglutinin
(HA)-epitope-GnT-Vb fusion expression plasmid was a 2-step
process: first, oligonucleotides encoding the following HA
epitope (MGSYPYDVPDYASLEF) were designed to have BamH1
and EcoRI overhangs. The annealed oligos were cloned into
BamHI and EcoRI cut pCDNA3.1+neo (Invitrogen Corporation,
Carlsbad, CA) creating the plasmid 3.1-HA. The cDNA encod-
ing GnT-VB was cut from pCD-VB [13] using EcoRI and NotI
restriction enzymes and cloned into 3.1-HA creating the
plasmid 3.1-HA-Vb. All plasmids created for this study were
sequenced to verify accuracy.
VB-coding-: 5′-AGCTTAA-
CATCTCTTGAATGCT
GnT-Va-: 5′-AGCTTA-
Cell culture and transfections
SH-SY5Y cells were grown in Dulbecco's modified Eagle
medium with 4.5 g/l Glucose, 2 mM L-glutamine, 10% fetal
calf serum (Hyclone, Logan, UT), penicillin and streptomycin
(100 μg/ml each) at 37°C, 5% CO2. For transfection, SH-SY5Y
cells were grown to 70% confluence and transfected using
Lipofectamine 2000 (Invitrogen Corporation, Carlsbad, CA)
according to the manufacturer's instructions. To create stable
lines, the cells were maintained with 800 μg/ml G418 for 2
weeks and used for 6 weeks following selection.
Real-time PCR
Total RNA was isolated using TRIZOL (Invitrogen) reagent.
Two micrograms of total RNA was reverse transcribed using
the iScript cDNA synthesis kit (Bio-Rad). The primers used for
the amplification of GnT-Va and GnT-VB have been described
2838
E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 2 8 3 7 – 2 8 5 0

Page 3

previously [13]. The primer set for G3PDH was 5′-CAGCCT-
CAAGATCAT CAGCA-3′ and 5′-GTCTTCTGGGTGGCAGTGAT-
3′. The primer set for PomGnT1 was 5′-CTGCGGAGATTCTGT-
CAGACAGGG-3′and5′-CATTGGCTTCACTGATGGCTCGCC.The
primer set for human β1 integrin was 5′-GGATATTACTCAGAT
CCAACCACAGC-3′ and 5′-GGTCAATGGGATAGTCTTCAGC-3′.
The primer set for human α1 integrin was 5′-GGAGGATTTC-
TGGCTTGTGGGC-3′ and 5′-GGCAATGGAATTCACGACTTG-3′.
Real-timePCRwasperformedasdescribedpreviously[13]using
theiQTMSYBRGreenSupermix(Bio-Rad).
Cell adhesion assays
Cell adhesion was determined according to the method
previously described [19]. Ninety-six-well plates were coated
with 10 μg/ml human placental laminin (Sigma). Additional
wells were coated with 1% bovine serum albumin in PBS
(minimal adhesion-negative control) and 1 mg/ml poly-d-
lysine (maximal adhesion-positive control). The plates were
left to dry at 37°C for 1 h. Each well was washed with 1× PBS
with calcium and magnesium and blocked with 1% BSA and
incubated at 37°C for 30 min. In each well, 4×104cells were
seeded in serum-free DMEM. Fluorescent images of adherent
cells were captured 45 min after plating. For quantification of
cell adhesion, the plates were incubated at 37°C for a total of
30 min then washed gently three times with 1× PBS to remove
any unattached cells. Each well was fixed with 3.5%
formaldehyde for 15 min and then stained with 0.5% crystal
violet solution overnight. The stained wells were washed
twice with 1× PBS and the absorbance was measured at
595 nm using an automated microtiter plate spectrophot-
ometer. The data are expressed as the mean of triplicate
wells performed in three separate experiments. For β1
integrin blocking experiments, the monoclonal antibody
mAb13 or normal mouse IgG was added at a concentration
of 30 μg/ml to the cells for 30 min at 37°C prior to plating on
laminin.
Scratch wound assay
SH-SY5Y cells expressing control or gene specific RNAi stable
cell lines were seeded into four well chamber slides coated
with laminin in OPTI-MEM. After the 24 h, a clear area was
scraped in the monolayer with a 200 μl plastic tip. After
being washed with OPTI-MEM, the chamber slides were
incubated for 4 or 6 h at 37°C in regular growth media.
Migration of the cells into the wounded area was monitored
by microscopy.
Transwell migration assay
Migration assays were performed using 24-well transwell
units with 8-μm pore size polycarbonate inserts. The under-
sides of the inserts were coated with laminin (10 μg/ml) at 37°C
overnight then blocked with 1% BSA. SH-SY5Y cells (7×104)
were seeded in 400 μl OPTI-MEM media in the upper
compartment while 500 μl OPTI-MEM supplemented with 2%
fetal bovine serum was placed in the lower chamber. Cells
were allowed to migrate for 6 h in a humidified incubator at
37°C with 5% CO2. Cells on the inside of the membrane were
removed and the cells that had migrated to the underside of
the membrane were fixed and stained with crystal violet
overnight. Five random fields were counted from two inde-
pendent experiments.
Neurite differentiation
All trans retinoic acid (ATRA) was used to promote axon
growth in the SH-SY5Y cell lines. Six-well plates were coated
with laminin as described for the 96-well plates. Stable SH-
SY5Y RNAi expressing cells plated at 50% confluence were
incubated in the absence or presence of 10 μM of retinoic
acid (RA) in OPTI-MEM media containing an additional 2%
FBS. The cells in the absence of ATRA were incubated with
DMSO as a negative control. Twenty-four hours after the
addition of ATRA cells with neurite length greater than 2 cell
bodies were counted. Results shown are from the analysis
from 3 separate experiments totally >1000 cells for each
stable cell line.
Immunoblotting
Cells were harvested in RIPA buffer (1× PBS, 1% Nonidet P-40,
0.5% sodium deoxycholate, 0.1% SDS). Twenty micrograms
total lysate protein was boiled in SDS sample buffer and
separated on 4–12% Bis Tris polyacrylamide gels (Invitrogen)
before transfer to PVDF membrane. Membranes were blocked
overnight at 4°C in 5% nonfat milk/1×TBS (TBS Blotto). After
blocking the membranes were incubated in primary anti-
bodies (1:1000) diluted in 1× TBS/.05% Tween 20/5% nonfat
milk (TBST Blotto) for 1 h at room temperature. After washing,
the membranes were incubated with HRP-conjugated second-
ary antibody diluted (1: 10,000) in TBST at room temperature
for 30 min. Washed blots were developed in Western Light-
ening Plus substrate (Perkin Elmer) and exposed to Kodak
Biomax chemiluminescent film.
Cell surface labeling and immunoprecipitation
Subconfluent cells were washed twice in ice-cold 1× PBS and
incubated with NHS-LC-biotin (1 mg/ml) in PBS at 4°C for
30 min on a rocking platform. Cells were washed twice with
cold PBS and lysed in RIPA buffer supplemented with
protease inhibitors for 30 min on ice. For immunoprecipita-
tion lysates (300 μg) were precleared by incubating in
protein A/G agarose for 1 h to remove non-specific binding.
After protein concentration quantification, 150 μg of lysates
were incubated in the presence of either 2 μg β1 integrin
antibody (Chemicon) or 50 μl streptavidin agarose overnight
at 4°C.
Results
GnT-Vb expression levels regulate neurite outgrowth of
SH-SY5Y neuroblastoma cells
The SH-SY5Y cell line has been extensively used to study
neuronal differentiation [20–22]. Notably, SH-SY5Y cells plated
onlaminin can be inducedto differentiateandextend neurites
2839
E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 2 8 3 7 – 2 8 5 0

Page 4

by adding micromolar amounts of all trans retinoic acid
(ATRA). To test if GnT-Vb-mediated glycosylation functions in
human neuroblastoma cell differentiation in vitro, GnT-Vb
was cloned into the expression vector pCDNA 3.1-HA and
constitutively expressed in SH-SY5Y cells. There were no
noticeable morphology differences between vector-trans-
fected (3.1-HA) and GnT-Vb-transfected (3.1-HA-VB) cells
grown on plastic cell culture dishes (data not shown). Both
types of transfectants were induced to differentiate by plating
on laminin-coated dishes in the presence of 10 μM ATRA. GnT-
Vb-expressing cells displayed markedly enhanced neurite
outgrowth at 24 h compared to mock-transfected control
cells (Fig. 1A). Real-time PCR measurement of GnT-Vb
transcript levels revealed an 18-fold increase in the level of
GnT-Vb expressioncomparedwith mock transfectedcells (Fig.
1B). To determine if expression of GnT-Vb as a fusion protein
with the HA epitope translated into increased GnT-Vb enzyme
activity we measured the specific activity from cell lysates as
described previously [23]. GnT-Vb-expressing cells showed a
large increase in activity compared with mock transfected
cells (Fig. 1C). Several neurite outgrowth experiments were
performed, andcumulative resultscollectedat 24hafter ATRA
addition indicate that GnT-Vb expression leads to >50% of
cells showing neurite extensions greater than 2× the cell body
length while only 15% of cells transfected with the vector
alone show neurites greater than 2× the cell body (Fig. 1D).
This effect was qualitatively similar to NGF-induced enhanced
neurite outgrowth in the pheochromocytoma cell line PC12
after increased GnT-Vb expression [23].
Suppression of endogenous levels of glycosyltransferases
GnT-Va, GnT-Vb, and PomGnT by expressing short hairpin
RNAs, inhibits ATRA-induced neurite outgrowth
InvitrosubstratespecificityassaysdemonstratethatGnT-Vbcan
add β(1,6) GlcNAc to both N-linked and O-mannosyl linked
glycans, while GnT-Va can add β(1,6) GlcNAc only to N-linked
glycans (Fig. 2A) [10,11,13,24]. Moreover, PomGnT1 is specific for
theO-mannosylglycosylationpathwayandaddsGlcNAcinβ(1,2)
linkage to the O-mannosyl glycan prior to GnT-Vb addition of
GlcNAcinβ(1,6)linkage(Fig.2A)[3,10].Todetermineiftheeffects
ofGnT-Vbexpression on neurite outgrowth were likelydueto N-
linked versus O-linked glycosylation, we established stable
knockdown cell lines using RNA interference [18,25]. Prior to
suppressing gene expression, the endogenous levels of GnT-Va,
GnT-Vb, and POMGn-T1 were measured in SH-SY5Y cells using
real-time PCR. PomGnT1 is expressed most abundantly, with
GnT-Va and GnT-Vb being expressed at lesser levels (Fig. 2B).
SeveralhairpinRNAsequencescomplementarytoGnT-Vb,GnT-
Va, and PomGnT1 were selected using the Oligoengine search
engine and cloned into the pSUPER vector, which constitutively
expresses hairpin RNAs from the human H1 promoter and
containstheneomycinresistancegeneallowingfortheselection
of a stable population of cells [26]. The location of the targeting
Fig. 1 – SH-SY5Y cells expressing GnT-Vb show increased neurite outgrowth following retinoic acid induced differentiation.
(A) SH-SY5Y cells stably expressing either pcDNA 3.1-HA or pcDNA3.1-HA-VB were plated on laminin-coated 6-well culture
plates at 50% confluence. Cells were placed in OPTI-MEM supplemented with 2% FBS with 10 μM ATRA. Phase contrast images
were captured 24 h after drug addition, bar 50 μm. (B) Real-time PCR measurement of GnT-Vb transcript levels from vector (3.1-
HA) and GnT-Vb expressing (3.1-HA-VB) SH-SY5Y cells normalized to GAPDH. (C) Enzymatic assay for GnT-Vb activity using
radiolabeled UDP-[3H]GlcNAc and synthetic trisaccharide acceptor in the presence of Mn+2. (D) Cumulative data representing
the percentage of cells displaying neurites greaterthan 2× the cell body length collectedfrom 3 separate experimentsand >1000
cells analyzed. Each bar represents the (±SD) of triplicate determinations.
2840
E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 2 8 3 7 – 2 8 5 0

Page 5

sequences that were selected is shown in Fig. 3A, and stable
RNAi-expressing cell lines were produced that expressed
sequences based on these locations: VB-coding region, VB-3′
UTR,PomGnT1,GnT-Va,aswellascontrol(scrambledsequences
targetingnospecifictranscript).Real-timePCRanalysiswasused
to measure the levels of each glycosyltransferase transcript to
determine gene suppression levels. In the two GnT-Vb-sup-
pressed stable cell lines, the levels of GnT-Vb mRNA were
reducedbyatleast90%intheVB-codingRNAicellsand65%inthe
VB-3′UTRRNAicells(Fig.3B).A72%reductioninPomGnT1mRNA
levels and a 60% reduction in GnT-Va mRNA levels were also
measured(Fig.3B).DespitethefactthattheVB-codingRNAi cells
had a greater reduction in mRNA levels (90%) than the VB-3′ UTR
RNAicells(65%),weobservedsimilarphenotypicchangesinboth
cell lines. These data would suggest that a 60% reduction in
transcript levels is sufficient for effectively reducing target gene
expression in RNA interference-mediated stable gene suppres-
sionexperiments.Furthermore,wehaveevaluatedtheefficacyof
GnT-Va suppression in MDA-MB231 cells, a cell type in which all
β(1,6) branched glycans would be due to the activity of GnT-Va.
The presence of the N-linked β(1,6) branched glycan added by
GnT-Va is specifically recognized by a carbohydrate binding
protein (lectin) known as L-PHA (L-phytohemagglutinin) [27].
Comparing L-PHA reactivity between control and GnT-Va
suppressed cells, we were able to confirm that a 50% reduction
in GnT-Va mRNA was sufficient to reduce the presence of β(1,6)
branched glycan structures by >75% (data not shown). The
results of these experiments confirm both the specificity and
efficacy of GnT-Va suppression using the selected RNA inter-
ference target sequence.
Suppression of both GnT-Vb and PomGnT1 expression impairs
the ability of human neuroblastoma cells to differentiate,
increases adhesion, and reduces migration on laminin
Separate populations of SH-SY5Y cells stably expressing
Control RNAi, GnT-Vb-coding RNAi, GnT-Va RNAi, or
PomGnT1 RNAi were each induced to differentiate on
Fig. 2 – Glycosyltransferase synthetic pathway and relative abundance of GnT-Va, GnT-Vb, and PomGnT1 in SY-SY5Y cells
(A) PomGnT1 catalyzes the transfer of GlcNAc in β(1,2) linkage to the O-linked mannose structure [3]. GnT-Vb transfers to the
O-linked mannose in β(1,6) linkage after PomGnT1 [10]. GnT-Va catalyzes the transfer of GlcNAc in β(1,6) linkage to α(1,6)
mannose arm of the N-linked glycan leading to poly-N-acetyllactosamine extension [49]. GnT-Vb can transfer GlcNAc in β(1,6)
linkage to either the α(1,3) or α(1,6) mannose arm of the N-glycan [12,13]. (B) Relative expression levels of each
glycosyltransferase in SH-SY5Y cells. Endogenous levels of GnT-Vb, GnT-Va, and PomGnT1 transcripts were measured using
real-time PCR and data are expressed as relative levels normalized to GAPDH.
2841
E X P E R I M E N T A L C E L L R E S E A R C H 3 1 2 ( 2 0 0 6 ) 2 8 3 7 – 2 8 5 0