(Received for publication, April 3, 1997)
From the Department of Biochemistry and Molecular
Biology and Complex Carbohydrate Research Center, University of
Georgia, Athens, Georgia, 30602 and the ¶ Department of Cell
Biology and Anatomy, University of Miami Medical School,
Miami, Florida 33101
Transformation of baby hamster kidney fibroblasts
by the Rous sarcoma virus causes a significant increase in the
GlcNAc(1,6)Man-branched oligosaccharides by elevating the activity
and mRNA transcript levels encoding
N-acetylglucosaminyltransferase V (GlcNAc-T V). Elevated
activity and mRNA levels could be inhibited by blocking cell
proliferation with herbimycin A, demonstrating that Src kinase activity
can regulate GlcNAc-T V expression. 5
RACE analysis was used to
identify a 3-kilobase 5
-untranslated region from GlcNAc-T V mRNA
and locate a transcriptional start site in a 25-kilobase pair GlcNAc-T
V human genomic clone. A 6-kilobase pair fragment of the 5
region
of the gene contained AP-1 and PEA3/Ets binding elements and, when
co-transfected with a src expression plasmid into HepG2
cells, conferred src-stimulated transcriptional enhancement upon a luciferase reporter gene. This stimulation by src
could be antagonized by co-transfection with a dominant-negative mutant of the Raf kinase, suggesting the involvement of Ets transcription factors in the regulation of GlcNAc-T V gene expression. The
src-responsive element was localized by 5
deletion
analysis to a 250-base pair region containing two overlapping Ets
sites. src stimulation of transcription from this region
was inhibited by co-transfection with a dominant-negative mutant of
Ets-2, demonstrating that the effects of the src kinase on
GlcNAc-T V expression are dependent on Ets.
The glycosylation of cell surface glycoproteins is a dynamic
process that can be regulated by agents that cause differentiation, such as retinoic acid (1) or transforming growth factor- (2), or by
those that induce cellular proliferation, for example, interleukin-1 or
tumor necrosis factor-
(3). In many instances, alterations of the
oligosaccharides on cell surface glycoproteins cause significant changes in the adhesive or migratory behavior of a cell (4, 5). An
induced alteration in the glycosylation of cell surface glycoproteins
that has been documented for many years concerns the significant
increase in oligosaccharide size caused by oncogenic transformation
using a variety of agents (6-14). This increase in size was found to
result mainly from an increase in the levels of asparagine-linked
oligosaccharides containing N-acetylglucosamine linked
1,6 to the
(1,6)-linked mannose in the trimannosyl core, (GlcNAc
(1,6)Man), and in many cases these oligosaccharides
express polylactosamine that can be sialylated (15-18). The
(GlcNAc
(1,6)Man) branch is synthesized by
N-acetylglucosaminyltransferase V (GlcNAc-T V),1 the enzyme whose activity is
significantly and selectively increased after transformation by tumor
viruses or isolated oncogenes (16, 19-22). Moreover, decreased
expression of the GlcNAc
(1,6)Man branch has been correlated with
decreased metastatic potential (23, 24), whereas the increased
expression of this branch appears in some instances to correlate with
the progression of invasive malignancies (25).
The transformation of baby hamster kidney (BHK) fibroblasts by the
src oncogene causes an increase in N-linked
oligosaccharide (GlcNAc(1,6)Man) branching, and the mechanism by
which this increase occurs has been under investigation in our
laboratories. To elucidate this mechanism, we examined GlcNAc-T V
enzyme activity and mRNA levels in BHK cells and their Rous sarcoma
virus-transformed counterparts (RSV-BHK) in the presence of the Src
kinase inhibitor, herbimycin A. The results from these experiments led
us to examine the 5
region of the human gene encoding GlcNAc-T V and
its increased expression caused by Src activity. Our results indicate
that the N-acetylglucosaminyltransferase V gene can be
transcriptionally activated by Src tyrosine kinase activity, and this
control is dependent on both the Raf-1 kinase and an Ets family
transcriptional activator.
Cells were grown to
confluency and harvested in 50 mM MES 6.5, 150 mM NaCl, and lysed by addition of Trition X-100 to 1%. Lysates were assayed according to the method of Palcic et
al. (22). Briefly, 106 cpm of
UDP-[3H]GlcNAc (25 cpm/pmol) and 10 nmol of synthetic
trisaccharide acceptor for GlcNAc-T V (octyl
6-O-[2-O-(2-acetamido-2-deoxy--D-glucosyl-pyranosyl)-
-D-mannopyranosyl]-
-D-glucopyraoside) were dried under vacuum in a 1.5-ml microcentrifuge tube. Extracts of
various protein concentrations were prepared, and 10 µl were added to
the assay tube. Assays were incubated at 37 °C for 4 h and
quenched by the addition of 500 µl of water. Radiolabeled product was
isolated on a C18 Sep-Pak (Waters) column, eluted in 2 ml
of methanol, and counted in a scintillation counter. Assays were
performed in duplicate or triplicate, at two or three protein concentrations, and specific activity was calculated by linear least
squares regression analysis of the data.
20 µg of total RNA was electrophoresed on a 1% formaldehyde-agarose gel and blotted to nylon. A 1-kb fragment of a partial rat GlcNAc-T V cDNA clone was random prime-labeled (26), and the blot was probed according to the method of Church and Gilbert (27). Data were collected and quantitated with a PhosphorImager.
Phosphotyrosine QuantitationCells were harvested by the addition of 1 ml of preheated SDS-polyacrylamide gel electrophoresis sample buffer to a 10-cm plate and shearing through a 20 gauge needle. Protein concentrations were determined on trichloroacetic acid precipitates using the BCA reagents (Pierce). 20 µg of protein were electrophoresed on a 4-20% gradient acrylamide gel and transferred to nitrocellulose (Bio-Rad). Blots were probed with an anti-phosphotyrosine antibody (a kind gift from Dr. Bart Sefton) followed by a goat anti-mouse horseradish peroxidase conjugate. Bands were detected using the ECL reagents (Amersham Corp.) and quantitated by scanning densitometry.
5Marathon Race Ready cDNA from human
whole brain (CLONTECH) was used as a template in a
5 RACE PCR according to the manufacturer's instructions. A 3
PCR
primer (303, CCTGGACCTCAGCAAAAGGTACATCAAGGC) designed near the 5
end
of the published human cDNA sequence was used along with the 5
anchor primer in a primary round of RACE PCR. Products were separated
on a 1% Tris-acetic acid-EDTA-agarose gel and blotted to
nitrocellulose. A nested PCR product was generated using the primers
501 (GAATGGAAGTGAGGGAAGGC) and 305 (GGAAGTTGTCCTCTCAGAAGCTGGGCTTT) and
a genomic clone template. This product was random prime-labeled (26)
and used to probe the membrane to which the RACE PCR product was
transferred (see Fig. 6). To improve yields of authentic products, a
secondary round of RACE PCR was then performed. A nested GlcNAc-T V 3
primer (305) and nested 5
anchor primer were used in the secondary
round of RACE PCR using as templates the products from the primary RACE
PCR. The products from this round of RACE PCR were directly subcloned
into the TA cloning vector (Invitrogen), and clones were sequenced.
Genomic Clone Isolation
The human GlcNAc-T V promoter
sequences were isolated from a human genomic library cloned in the
-FIX II vector (Stratagene, Inc.). A 687-bp EcoRI
fragment containing the 5
-untranslated region of the rat cDNA (28)
was used as a probe to screen this library using standard plaque
hybridization procedures. Screening 5 × 106 phage
plaques yielded two overlapping genomic clones that span a region of
approximately 30 kb.
Sequencing of RACE PCR clones and human genomic clones was performed using the Applied Biosystems, Inc. reagents by the UGA Molecular Genetics Instrumentation Facility.
Luciferase Reporter ConstructionFor pGL2-TV1, a 6-kb XhoI-SacI fragment of the genomic clone was band-purified (Sephaglass, Pharmacia Biotech Inc.) and subcloned into XhoI-SacI sites of the pGL2-Basic vector (Promega). For pGL2-TV2 through pGL2-TV4, PCR products were generated using the genomic clone as a template and subcloned into the TA vector. TA clones of the correct orientation were cut with XhoI-SacI, and inserts were cloned into XhoI-SacI sites of pGL2-Basic.
Promoter Activity DeterminationsSV40--galactosidase (2 µg) and reporter constructs (2 µg) ± effector plasmids (2 µg)
were transfected by the calcium-phosphate precipitation method (29)
into 50% confluent cultures of HepG2 cells grown in 6-well culture
plates. 40 h post-transfection, cell lysates were prepared and
assayed for
-galactosidase and luciferase (Promega). Luciferase
activity was normalized to vector-dependent
-galactosidase activity.
The plasmids encoding the Raf-1 kinase and its dominant negative form (30) were kind gifts from Dr. Ulf Rapp. Plasmids encoding Ets-2 and its dominant negative form (31) were kinds gifts from Dr. K. E. Boulukos. The v-src expression plasmid was a kind gift from Dr. Tony Hunter.
Earlier experiments utilizing BHK and RSV-BHK cells metabolically
radiolabeled with [2-3H]mannose indicated at least a
2-fold increase in the total amount of (2,6)-substituted mannose in the
RSV-BHK cells, normalized to total mannose-labeled glycopeptides.
Although the specific activity of GlcNAc-T V was increased over 6-fold
in the RSV-transformed cells, no significant differences in the kinetic
properties of GlcNAc-T V in the transformed cells could be detected.
These results suggested that the increases in GlcNAc(1,6)Man levels
after transformation were most likely not due to post-translational
effects on the enzyme (22). The specific activity of GlcNAc-T V and its
mRNA levels were measured, therefore, to determine if the increase in GlcNAc-T V activity in the transformed cells could result from differences in mRNA levels. GlcNAc-T V activity was assayed under optimal conditions in BHK and RSV-BHK cells using a synthetic trisaccharide acceptor. The transformed BHK cells demonstrated a
GlcNAc-T V enzyme specific activity 6-fold higher than the
untransformed BHK cells. By contrast, no difference was seen in the
specific activity of another N-acetylglucosaminyltransferase
that functions in the synthesis of N-linked
oligosaccharides, GlcNAc-T I (Fig. 1), indicating the
specificity of Rous sarcoma virus transformation on GlcNAc-T V
activity. To investigate the possibility that the difference in
GlcNAc-T V specific activity is associated with a difference in
steady-state mRNA levels, Northern blots were performed using a
fragment of a cDNA encoding GlcNAc-T V. Compared with BHK cells,
RSV-transformed cells were found to have a 6-fold increase in the
expression levels of both 8.7 and 9.3 kb GlcNAc-T V transcripts, but no
change in either GAPDH or GlcNAc-T I transcripts (Fig.
2). Although not apparent in Fig. 2, PhosphorImager
quantitation demonstrated an equivalent increase in both GlcNAc-T V
mRNA transcripts. These results demonstrated that the elevation of
enzyme activity in the RSV-transformed cells was a result of either
transcriptional activation or increased mRNA stability and argue
against postranslational modifications of the enzyme causing a
significant increase in its catalytic activity.
To obtain convincing evidence that the differences in GlcNAc-T V expression result from src tyrosine kinase activity, we made use of a src-selective tyrosine kinase inhibitor, herbimycin A, a metabolite produced by Streptomyces sp. MH237-CF8. This inhibitor was first identified for its ability to reverse the transformed morphology of Rous sarcoma virus-infected rat kidney cells (32), and this reversion of morphology was associated with a reduction in total cellular phosphotyrosine levels (33). Herbimycin A was unable, however, to reverse the transformed morphologies induced by the ras, raf, or myc oncogenes, demonstrating its specificity for the src family of tyrosine kinase oncogenes (34). Herbimycin A is also able to reverse src-stimulated expression of the glucose transporter gene (35) and to cause a reversible Go contact arrest in src-transformed normal rat kidney cells (36). We utilized herbimycin A, therefore, to test the hypothesis that the expression of GlcNAc-T V is positively regulated by the src tyrosine kinase.
First, to monitor the effects of src kinase and demonstrate
its inhibition by herbimycin A, we measured total cellular
phosphotyrosine levels by performing Western blots with an
-phosphotyrosine antibody on extracts made from cells treated for
24 h with various concentrations of the drug. These results
demonstrate that, as expected, herbimycin A caused a
dose-dependent decrease in cellular phosphotyrosine levels
(Fig. 3). At a concentration of 1 µg/ml, herbimycin A
caused a reversal of the RSV-transformed morphology and a complete
inhibition of cell division (data not shown). Consistent with the
drug's effect of inhibiting the src kinase (33) and
blocking cell division in Go (36), herbimycin A caused a
dose-dependent decrease in GlcNAc-T V enzyme specific
activity in RSV-transformed BHK cells (Fig. 4).
Interestingly, although the drug blocked cell division and caused a
modest decrease in phosphotyrosine levels in the untransformed BHK
cells (data not shown), it had little effect on the expression level of
GlcNAc-T V enzyme activity in confluent cultures of untransformed cells
(Fig. 4). This result suggests that regulation of GlcNAc-T V expression
is complex, with both src-dependent and
src-independent factors. Herbimycin A had no effect on the
specific activity of GlcNAc-T I (data not shown), arguing against
nonspecific toxic effects on the transformed cells and confirming that
GlcNAc-T I is not regulated by src. To determine if the
inhibition of expression of GlcNAc-T V enzyme activity by herbimycin A
was a result of inhibiting the expression of the mRNA encoding the
enzyme, Northern blots were performed on RNA samples prepared from
RSV-transformed cells treated with various concentrations of the drug.
Similar to its effects on GlcNAc-T V enzyme specific activity,
herbimycin A caused a decrease in GlcNAc-T V message levels in the
RSV-BHK cells in a dose-dependent manner (Fig.
5). Taken together, these results indicate that
expression of the GlcNAc-T V mRNA in the src-transformed
cells is under the control of the src tyrosine kinase.
To elucidate the mechanism by which src induces the
expression of GlcNAc-T V, we isolated the 5-flanking region of the
gene and analyzed this region for promoter activity. The GlcNAc-T V message is approximately 9 kb in most rodent and human tissues, with
brain having high expression levels. To locate a promoter for GlcNAc-T
V, 5
RACE PCR techniques were used to isolate and sequence the 5
end
of the message from human brain. RACE PCR products were first generated
from the genomic clone using the 303 primer (designed against the 5
end of the human GlcNAc-T V cDNA sequence) and then analyzed by
Southern blotting. GlcNAc-T V-specific sequences were detected using a
nested 501-305 PCR product as the hybridization probe. Multiple bands
were detected, the longest of which was 2.9 kb (Fig. 6).
PCR products were ligated into the TA cloning vector, and clones
corresponding to the 600- and 1200-bp products were obtained and found
to overlap and differ only in the length of their 5
ends. To isolate a
product that encompassed all of the 5
-untranslated region of the
GlcNAc-T V message, a second round of RACE PCR was then performed using two nested primers designed near the 5
end of the 0.6 kb clone, and
the resulting products were subcloned and sequenced. Clones corresponding to the 1.8- and 2.9-kb bands were isolated from this
second round of RACE PCR and sequenced. As before, these clones were
found to differ only in the length of their 5
ends. A third round of
RACE PCR using a 3
primer designed to the sequence near the 5
end of
the 2.9-kb clone produced a single product of the expected size (70 bp). The assembled sequence of the RACE clones was co-linear with the
sequence of the genomic clone in this region, indicating that no
splicing events occurred in the 5
-UTR of the message from human brain.
These results demonstrate, therefore, that the location of the 5
-most
transcriptional start site utilized in brain is located approximately
2.9 kb upstream of the ATG, corresponding to the band of this size
shown in Fig. 6.
To examine the 5 region flanking the 5
-most transcriptional start
site, a 6-kb SacI-XhoI genomic fragment
containing 848 bp of the brain 5
-UTR and 5.5 kb of 5
-flanking
sequence (depicted in Fig. 7), designated pGL2-TV1, was
cloned into the luciferase expression vector pGL2-basic. The activity
of this region as a promoter and its responsiveness to src
were then examined in transiently transfected HepG2 cells. pGL2-TV1 was
found to act as a weak promoter, shown in Fig. 8,
consistent with the low levels of GlcNAc-T V transcript observed in
HepG2 cells and most tissues. Moreover, this DNA fragment conferred
transcriptional responsiveness to src when co-transfected
with a src-containing expression plasmid (Fig. 8). The
similarity between the increases in GlcNAc-T V expression in the
RSV-transformed cells and src stimulation of transcription from the GlcNAc-T V promoter in HepG2 cells suggests that
transcriptional control is likely the most important regulatory
influence of src on GlcNAc-T V activity.
Transformation and some transcriptional activation by src occurs via the MAPK pathway in a Raf-1-dependent manner. For example, a dominant-negative Raf-1 mutant suppresses src transformation of BALB/c mouse fibroblasts (37) and is able to block src-stimulated transcriptional activation of the EGR gene (38). Furthermore, this mutant is able to block serum or Ras stimulation of the transcription of an AP-1/Ets-driven gene (30). If transcriptional activation of GlcNAc-T V by src occurs at least in part via the MAPK pathway, we reasoned that the activation should be inhibited by the dominant-negative Raf-1 mutant. Consistent with this hypothesis, the transcriptional stimulation of the GlcNAc-T V promoter by src was significantly inhibited when co-transfected with a plasmid encoding a dominant-negative mutant Raf, RafC4B-DN (Fig. 8). Because proliferation often occurs in a Raf-dependent manner, this result is consistent with correlations noted between GlcNAc-T V expression and cellular proliferation in nontransformed cells (39, 40) and may predict in certain cell types a general association between GlcNAc-T V enzyme activity and cell proliferation.
To map more closely the region of the GlcNAc-T V promoter responsible
for transcriptional activation by src, a series of 5 deletions containing 70 bp of the brain 5
-UTR and different amounts of
the 5
-flanking region were constructed by PCR amplification from the
genomic clone, the boundaries of which are depicted in Fig. 7. Promoter
fragments were cloned into the pGL2-basic vector and tested for basal
promoter activity and src responsiveness. Based on the
results from several sets of experiments, both pGL2-TV2, which
contained about 1.2 kb, and pGL2-TV3, which contained 739 bp, were both
found to be weakly active as promoters and transcriptionally responsive
to src (Fig. 9). The pGL2-TV4 construct
containing 339 bp, however, was found to be inactive as a basic
promoter and completely unresponsive to src. These results
suggest a requirement for the two overlapping PEA-3 sites located near
the transcriptional start site, contained in pGL2-TV3, in the
src-mediated transcriptional activation of the GlcNAc-T V
gene (Figs. 7 and 10).
PEA-3 sites are bound by the Ets family of transcriptional activators
and mediate transcriptional activation in response to mitogenic signals
from plasma membrane-associated tyrosine kinase oncogenes (41). To
obtain more direct evidence for the involvement of an Ets family
transcription factor in src-stimulated transcription of the
GlcNAc-T V gene, we exploited the availability of an ets-2 expression plasmid along with an expression plasmid that encoded a
truncated ets-2 with a dominant-negative activity,
ets-21-328 (31). Co-transfection of pGL2-TV3 with
ets-2 resulted in stimulation of transcription, whereas
co-transfection with ets-2
1-328 was able to block
src-transcriptional stimulation (Fig. 11).
This result demonstrates that this region is capable of being
transcriptionally stimulated in response to ets-2 expression
and suggests that src transcriptionally activates this
region through a factor that binds to the PEA-3 site(s).
A major regulatory mechanism for controlling the expression of
glycan structures is alteration in the activities of the enzymes involved in their synthesis. Changes in glycosyltransferase activity may occur by several different mechanisms, including post-translational interactions with specifier proteins (42) and translational control of
glycosyltransferase transcripts. For example, the
(1,4)-galactosyltransferase is expressed at low to moderate levels
in most tissues as a 4.1-kb transcript. However, high level expression
in lactating mammary gland is accompanied by a switch from one
transcriptional start site to another, which is more proximal to the
translational start site, giving rise to the expression of a 3.9-kb
transcript (43). The longer transcript is predicted to form a stable
hairpin loop structure in the 5
-untranslated region, which may reduce
translation efficiency. The shorter transcript is predicted to be
unable to form any significant 5
structure and therefore is thought to be translated with greater efficiency. More direct evidence of this
kind of regulation can be found in a comparison between
(2,6)-sialyltransferase cDNAs from different tissues. The
cDNAs from transcripts encoding an
(2,6)-sialyltransferase
isolated from placenta and B cells encode identical proteins, but their
5
-untranslated regions diverge from each other approximately 180 bp
upstream from the ATG. The 5
-untranslated regions confer different
in vitro translational efficiencies on the respective
mRNAs. The placental transcript is translated with what appears to
be a 10-fold greater efficiency than is the B cell transcript.
Furthermore, an artificial cDNA construct lacking the 400-bp
5
-untranslated region is translated with an apparent 10-fold greater
efficiency than the B cell transcript, indicating that both
5
-untranslated regions function with different efficiencies as
translational suppressors (44).
Another mechanism for the regulation of glycosylation is the alteration
in the stability of the mRNAs that encode glycosyltransferases. F9
teratocarcinoma cells treated with retinoic acid plus dibutyryl cyclic
AMP differentiate into parietal endoderm-like cells and express 20-fold
greater specific activity of the (1,4)-galactosyltransferase. This
increased expression of activity was found to be accompanied by a
similar increase in the expression levels of the mRNA encoding the
enzyme. However, promoter activity determinations based on luciferase
reporter activity, as well as nuclear run-off assays both indicated
there to be no significant increase in the transcription rate of the
gene. The elevated expression of the mRNA was found to be
associated with a marked increase in the half-life of both the short
and long transcripts encoding the enzyme (45). In a similar study an
effect on the stability of the mRNA for
N-acetylglucosaminyltransferase V has been shown to account
for an increase in mRNA levels in B-16 mouse melanoma cells treated
with transforming growth factor-
(2).
Finally, control of glycan structure through control of the
transcriptional activity of glycosyltransferase genes has now been
described in several systems. Early indications of this regulatory mechanism were found in a functional analysis of a promoter of the
(2,6)-sialyltransferase gene. This transferase is highly expressed
in liver, and its promoter was found to contain binding sites for
liver-enriched transcription factors. These sites and their cognate
transcription factors were found to be functional in luciferase
reporter transactivation assays in Hep G2 cells but were not functional
in Chinese hamster ovary cells. These transcription factors and their
binding sites are thought to be responsible for the liver-restricted
expression of this message (46, 47). Also, retinoic acid-induced
differentiation of F9 teratocarcinoma cells has been documented to be
associated with an increase in the activity of the murine
(1,
3)-galactosyltransferase and a switch in the expression of terminal
capping structures from
(2, 3)-linked sialic acid to galactose. The
increase in the activity of the transferase is associated with an
elevation in the steady-state levels of the mRNA and a similar
increase in the transcription rate of the gene (1).
The association between oncogenic transformation and increases in
GlcNAc-T V enzyme activity and expression of GlcNAc(1,6)Man branches
on glycoproteins raises the question concerning the mechanism by which
oncogenes and proliferative signals in general up-regulate the
synthesis of GlcNAc
(1,6)Man branches. From our results it is
reasonable to conclude that the increases in GlcNAc-T V enzyme activity
seen in src-transformed cells are not due to an increase in
the catalytic efficiency of the glycosyltransferase; rather, they are
due to an increase in the expression of the GlcNAc-T-V mRNA and
subsequently an increase in the number of GlcNAc-T V enzyme
molecules/cell. The elevation in steady-state mRNA levels could be
a result of increased mRNA stability or an increase in the
transcription rate of the gene. Our results demonstrate that a 5
fragment of the GlcNAc-T V gene is able to confer
src-transcriptional responsiveness upon a heterologus
reporter gene, indicating that the mechanism by which src
increases steady-state mRNA levels is most likely transcriptional
activation. Also, the 5
-flanking region of the GlcNAc-T V gene was
found to contain AP-1 and PEA-3 binding sites, the presence of at least
two being required for src-stimulated transcriptional
enhancement. These data suggest that cooperation may be required
between AP-1 and Ets transcription factors to render the gene fully
responsive to oncogenic stimulation. Finally, transcriptional
stimulation by src was found to be significantly dependent
on the Ras-Raf pathway, again implicating transcription factors that
are activated by the MAPK pathway and suggesting a general association,
at least in some cell types, between GlcNAc
(1,6)Man branching and
cellular proliferation. Other studies have shown that particular
glycosyltransferase activities are increased after ras-transformation; notably,
(2,6)-sialyltransferase activity in NIH3T3 cells is
specifically increased by ras expression, and this increase
is due to increased levels of the mRNA encoding this enzyme (49,
50).
The regulation of the expression of the GlcNAc-TV gene is quite
complex, involving the use of multiple promoters and alternative splicing. The brain transcript described in this paper begins at a
promoter about 3 kb upstream of the translation initiation site,
generating a long 5-untranslated sequence that is unspliced and
colinear with the genomic sequence. This transcript and its promoter
are different from those observed in HuCC-T1 human bile duct carcinoma
cells (48). The GlcNAc-T V mRNAs in these cells appear to be
initiated from multiple promoters, one of which is about 1.4 kb
downstream of the brain promoter. These mRNAs are also
differentially spliced in the 5
-UTR to generate messages with much
shorter 5
-UTRs. There are also data to suggest that there are
additional promoters and alternative splices used by A431 cells when
expressing this gene.2 The precise
identification of the cis-acting sequences that mediate the
activation of GlcNAc-T V transcription from these alternate promoters
will require further investigation. Other regulatory mechanisms may
also be operative in some cell types, including translational control.
The GlcNAc-T V transcript expressed in human brain and in HepG2 cells
is approximately three times larger than the coding region, indicating
the presence of extensive 5
- and 3
-untranslated sequences. The
transcript from human brain was found to have a 5
-untranslated region
of approximately 3 kb, the longest 5
-UTR of any glycosyltransferase
observed to date. This 5
-UTR may function as a regulator of
translation efficiency, as has been suggested in the case of the
mammary-specific transcript for the
(1,4)-galactosyltransferase and
demonstrated for the placental and B cell forms of the
(2,6)-sialyltransferase.
Of what consequence is the increased expression of GlcNAc(1,6)Man
branches during proliferation? One possibility is that these branches
expressed on cell surface proteins that function in cell adhesion can
significantly alter the adhesiveness of cells to the extracellular
matrix or to each other, allowing them to become more migratory.
Evidence in support of this hypothesis has been observed in the case of
mink lung epithelial cells stabily transfected with and overexpressing
mouse GlcNAc-T V (51). The cells overexpressing GlcNAc-T V show an
altered morphology show altered rates of migration in an in
vitro assay. Moreover, these cells show less adhesion to
laminin-coated surfaces, compared with the nontransfected controls,
suggesting an effect of GlcNAc-T V expression on laminin adhesion.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) AF004882.
In memoriam, E.A.P.
We are grateful to Dr. Kelley Moremen for many insightful discussions and suggestions.