 |
INTRODUCTION |
Polysialic acid is a unique carbohydrate of a linear homopolymer
of 
2,8-linked sialic acid, which may contain as many as 55 sialic
acid residues per polymer (1, 2). This unique glycan is mainly attached
to the fifth immunoglobulin-like domain of the neural cell adhesion
molecule (NCAM)1 via a
typical N-linked glycosylation (3, 4). NCAM is highly polysialylated in embryonic tissues. Although the majority of NCAM in
adult tissues lacks this unique glycan, polysialylated NCAM is present
in the olfactory bulb and hippocampus of adult brain where neuronal
regeneration persists (5, 6). Polysialic acid is thought to attenuate
the adhesive property of NCAM for homophilic interaction (7, 8),
thereby playing a role in neural cell migration (9), axonal growth,
path finding (10), and synaptogenesis (11). NCAM knockout mice exhibit
an abnormal formation of olfactory bulb and hippocampus and a defect in
spatial learning and memory, which may be caused by dysfunction of the hippocampus (12, 13). Such a phenotype appears to be due to the loss of
polysialic acid in these regions of the brain because treatment with
endoneuraminidase-N (endo-N), a specific glycosidase for polysialic
acid, resulted in the impairment of cell migration in the
subventricular zone of the olfactory bulb (14) and prevention of the
induction of long term potentiation, presumably by impairing the
induction of hippocampal synaptic plasticity (11). Taken together,
these results strongly suggest that polysialylated NCAM plays a
critical role in neural development and plasticity.
The cDNAs encoding polysialyltransferases have been cloned, and
these enzymes are called ST8Sia IV (PST) and ST8Sia II (STX) (15-19).
They belong to the 2,8-sialyltransferase gene family and are highly
homologous to each other, having 59% identity at the amino acid levels
(16, 17). By using an in vitro assay system, both ST8Sia IV
and ST8Sia II were shown to directly add polysialic acid to fetuin and
NCAM (20-23). This demonstrates that either ST8Sia IV or ST8Sia II
alone can form polysialic acid by adding the first 2,8-linked sialic
acid to 2,3- or 2,6-linked sialic acid in a glycoprotein
acceptor, followed by the multiple addition of 2,8-linked sialic
acid residues. However, it has not been determined if ST8Sia IV or
ST8Sia II can add polysialic acid more efficiently on a preformed
disialyl structure,
NeuNAc2
8NeuNAc23Gal
14GlcNAcR. In addition to
ST8Sia IV and ST8Sia II, three more 2,8-sialyltransferases have been
molecularly cloned. Among them, ST8Sia III (24) is more closely related
to ST8Sia IV or ST8Sia II than to ST8Sia I (GD3/GT3 synthase) (25-28)
or ST8Sia V (GT3 synthase) (29). These studies also demonstrated that
ST8Sia I and ST8Sia V utilize glycolipids as acceptors, whereas ST8Sia
III forms disialic and oligosialic acid in both glycoproteins and
glycolipids. On the other hand, it was shown previously that colominic
acid, a polymer of 2,8-linked sialic acid, did not serve as an
acceptor for polysialyltransferases (30). Moreover, it has been
reported that ST8Sia IV and ST8Sia II can add polysialic acid only on
sialylated N-glycans attached to NCAM because
N-glycans released from NCAM did not serve as acceptors
(22). These results suggest that ST8Sia IV and ST8Sia II cannot utilize
free N-glycans as acceptors.
In the present study, we first describe the unexpected discovery that
ST8Sia III can form polysialic acid. This discovery prompted us to
determine if polysialyltransferases and ST8Sia III differ in the mode
of actions and products formed. By using a newly developed assay, we
found that both polysialyltransferases and ST8Sia III form polysialic
acid on oligosaccharide acceptors and can form oligosialic and
polysialic acid also on the disialyl acceptor
(NeuNAc28NeuNAc23Gal14GlcNAcR). In contrast,
we demonstrated that ST8Sia IV and ST8Sia II are much more efficient in
polysialylation on NCAM than ST8Sia III. These results strongly suggest
that ST8Sia IV and ST8Sia II apparently evolved to participate in NCAM
polysialylation from a primordial enzyme that adds 2,8-sialic acid
to low molecular weight acceptors.
| |
EXPERIMENTAL PROCEDURES |
Isolation of a Full-length cDNA Encoding Human ST8Sia
III--
A cDNA harboring a catalytic domain of human ST8Sia III
was obtained as follows. First, cDNA was synthesized by reverse
transcriptase, SuperscriptTM II (Life Technologies, Inc.) using the
human fetal brain mRNA (CLONTECH) and a
3'-primer, 5'-TTGGAGTTCTTAGGCACAGTG-3', which is complementary to
nucleotides 1131-1151 (numbered 1-3 for the codon encoding the
initiation methionine) of the mouse ST8Sia III sequence (24). PCR was
carried out to obtain a catalytic domain using the formed cDNA as a
template and 5'-primer and 3'-nested primer. Both primers were adapted
from the mouse ST8Sia III sequence (nucleotides 113-130 and
1125-1147, respectively) (24), and BglII and
XhoI sites are included in these primers.
To obtain the human 5'-upstream cDNA sequence, 5'-RACE was
performed by the RACE System (Life Technologies, Inc.) using primers that are complementary to nucleotides 225-244 and 211-230,
respectively, in the human ST8Sia III sequence determined above. The
RACE product was digested at a SalI site in the anchor
sequence and at an internal XbaI site and cloned into the
same sites of pBluescript II (Stratagene).
To construct an intact form of human ST8Sia III, 5'-end sequence was
amplified by PCR using the vector harboring the 5'-RACE product
as a template. The upstream primer is
5'-CGTAAGCTTACACGCCAGCGAGCT-3' (HindIII is underlined), in which the last 16 nucleotides correspond to nucleotides 
72 to 57. The downstream
primer was adapted from the flanking vector sequence. The PCR product
thus obtained was digested by HindIII and XbaI,
and this fragment was appended to the 5'-end XbaI site
(nucleotides 201-206) of a large cDNA fragment (encompassing
nucleotides 113-1145), producing a full-length cDNA encoding human
ST8Sia III. The ligated cDNA was cloned into HindIII and
XhoI sites of pcDNAI (Invitrogen), resulting in
pcDNAI-ST8Sia III.
To obtain the 3'-end sequence, P1 plasmid human genomic DNA library was
screened using PCR as described (23, 28). Upstream and downstream
primers for this PCR correspond to nucleotides 890-911 and nucleotides
1110-1130 of the human ST8Sia III sequence, respectively. Purified DNA
from one of the P1 clones was cloned into pUC19 vector, and the plasmid
containing the ST8Sia III sequence was identified by PCR. By using the
above upstream primer, its nucleotide sequence was determined.
Construction of a Soluble Chimeric ST8Sia III--
pcDNAI-A
encoding a signal peptide and an IgG binding domain of protein A was
prepared as described before (4). A cDNA fragment encompassing
nucleotides 113 to 1145 of human ST8Sia III was obtained as described
above and digested by BglII and XhoI. It was then
ligated to BamHI site of the 3'-end of pcDNAI-A insert
and XhoI site of the vector, resulting in
pcDNAI-A·ST8Sia III.
Other Plasmids--
pcDNAI-ST8Sia IV harboring cDNA
encoding the full-length human ST8Sia IV and pcDNAI-A·ST8Sia IV
harboring cDNA encoding a soluble chimeric ST8Sia IV were
constructed as described previously (17, 20). Similarly,
pcDNAI-ST8Sia II and pcDNAI-A·ST8Sia II were constructed as
described (4, 23). pIG-NCAM·IgG encoding a soluble NCAM fused with
hinge and constant regions of human IgG (31) was kindly provided by Dr.
David Simmons, Oxford University.
Oligosaccharides--
Colominic acid (average degree of
polymerization = 30) was purchased from Sigma. N-Glycans
from bovine fetuin (A3) were purchased from Oxford Glycosystems. This
oligosaccharide has the structure of
NeuNAc23/6Gal14GlcNAc12Man16[NeuNAc23/6Gal14GlcNAc14(NeuNAc23/6Gal14GlcNAc12)Mana13]Man14GlcNAc14GlcNAc. NeuNAc23Gal 14GlcNAc1 octyl and
NeuNAc28NeuNAc23Gal14GlcNAc1octyl were
synthesized as described previously (32). The molecular mass of
these compounds was confirmed by matrix-assisted laser desorption
ionization-time of flight (MALDI-TOF) mass spectrometry. Gal14GlcNAc16Man16Man1O(CH2)7CH3
(octyl), compound 1, was synthesized as described (33, 34).
NeuNAc23Gal 1 4GlcNAc16Man16Man1octyl, compound 2, and
NeuNAc26Gal14GlcNAc16Man16Man1octyl, compound 3, were enzymatically synthesized from synthetic
compound 1 using rat 2,3-sialyltransferase (Calbiochem) and rat
2,6-sialyltransferase (Roche Molecular Biochemicals), respectively.
The details of these syntheses will be published
elsewhere.2 Compounds 2 and 3 were characterized by 1H NMR spectroscopy and MALDI-TOF
mass spectrometry, as below.
Partial 1H NMR (300 MHz, D2O) compound 2 and
(500 MHz, D2O) compound 3 are: 2, 
4.87 (d,
J = 1.3 Hz, H-1'), 4.65 (s, H-1), 4.54, 4.56 (2d,
J = 7.8 Hz, 2H), 2.74 (dd, J = 4.6, 12.6 Hz, H-3eq), 2.04, 2.01 (2 s, 6H, NHAc), 1.78 (t, J = 12.3 Hz, H-3ax); 3, 4.89 (s, H-1'), 4.67 (s, H-1), 4.58-4.63 (m, 1H), 4.44 (d,
J = 8.2 Hz, 1H), 2.67 (m, H-3eq), 2.02, 2.08 (2 s, 6H, NHAc), 1.72 (t, J 
12 Hz,
H-3ax). m/z (MALDI-TOF) are: 2, 1133.6 [M+Na]+ (calculation, 1133.4); 3, 1133.5 [M+Na]+ (calculated, 1133.4).
Expression of NCAM Together with ST8Sia II, ST8Sia III, or ST8Sia
IV in HeLa Cells--
Since HeLa cells were negative for both
polysialic acid and NCAM, they were transfected with pcDNAI-ST8Sia
III and pSV2neo using LipofectAMINE (Life Technologies,
Inc.) as described before (23). After selection by G418, clonal cell
lines stably expressing ST8Sia III, HeLa-ST8Sia III, were selected by
staining with M6703 antibody, which reacts strongly with
NeuNAc28NeuNAc28NeuNAc23GalR (35). Similarly,
HeLa-ST8Sia I was established (28). HeLa-ST8Sia I and HeLa-ST8Sia III
were transiently transfected with pCDM8-NCAM as described (23). HeLa
cells expressing ST8Sia II or ST8Sia IV together with NCAM, called
HeLa-ST8Sia II or HeLa-ST8Sia IV, were established as described
previously (23). NCAM·IgG was purified from the spent medium of COS-1
cells that had been transfected with pIG-NCAM·IgG as described
previously (4). NCAM·IgG isolated was used for in vitro
assay of ST8Sia II, ST8Sia III, and ST8Sia IV.
Similarly, the parent HeLa cells were stably transfected with
pIG-NCAM·IgG and NCAM·IgG was purified from the spent medium (4).
NCAM·IgG was digested with trypsin and chymotrypsin, then heated in a
boiling water bath for 3 min to inactivate the proteases. N-Glycans were then isolated after N-glycanase
treatment followed by phenol-water partition and washing of the water
phase by chloroform (4). This sample of NCAM N-glycans was
used as acceptors.
Sialyltransferase Assays and Product Characterization--
HeLa
cells expressing a soluble chimeric form of ST8Sia II or ST8Sia III or
ST8Sia IV were established as described previously (4). The soluble
forms of these enzymes were purified using IgG-Sepharose, and the
enzymatic activities were measured as described (4, 20). The substrate
solution (50 µl) contained 10 pmol of NCAM·IgG containing 60 pmol
of N-glycans, 30 pmol of 2-HS-glycoprotein containing 60 pmol of N-glycans (Sigma) or 60 pmol or 1 nmol
of synthetic oligosaccharides, colominic acid, or NCAM
N-glycans, and 2.4 nmol (0.7 mCi) of
CMP-[14C]NeuNAc in the same reaction mixture described
(4).
To this substrate solution, 50 µl of the enzyme solution, prepared as
described above, was added, and the reaction mixture was incubated at
37 °C for 2 h or 18 h. After a brief centrifugation, the
supernatant was analyzed by SDS-polyacrylamide gel electrophoresis followed by fluorography as described previously (20).
Products obtained from 2-HS-glycoprotein and NCAM·IgG
were purified by microcon 30 (Amicon) to remove free
CMP-[14C]NeuNAc and digested with N-glycanase
as described above. After incubation of fetuin oligosaccharides, NCAM
oligosaccharides, colominic acid, and 2,3-sialylated or
2,6-sialylated oligosaccharide, NeuNAc23Gal14GlcNAc1octyl or
NeuNAc28NeuNAc23Gal14GlcNAc1octyl, the
reaction mixtures were directly analyzed by HPLC. Products obtained
from sialylparagloboside were purified by reverse-phase column
chromatography using TSK-GEL® ODS-80Ts (TosoHAAS) column
at room temperature. After elution with buffer A, 200 mM
ammonium acetate buffer, pH 4.0, for 10 min, the column was eluted with
linear gradient elution to 100% of acetonitrile in the subsequent 80 min. The oligosaccharides were released by endoglycoceramidase II (36, Calbiochem) from unbound (fractions 1-10) and bound (fractions 11-40)
fractions, and the resultant oligosaccharides were subjected to Mono-Q
chromatography (see below).
The products were separated by Mono-Q HPLC using the elution conditions
modified from those previously reported (2, 4) as follows. Mono-QHR 5/5
(0.5 × 5 cm, Amersham Pharmacia Biotech) was equilibrated with
the solvent A, 2 mM Tris-HCl, pH 7.5, and then eluted with
three-step linear gradient elution using the solvent B, 1 M
NaCl and 10% acetonitrile in 2 mM Tris-HCl, pH 7.5, at
room temperature at a flow rate of 1 ml/min using a Gilson 306 HPLC.
After elution with solvent A for 10 min, the linear gradient elution
was carried out from 0% to 33% of the solvent B in 80 min and then to
45% of solvent B in the next 50 min and, finally, to 100% of the
solvent B in the last 10 min. After eluting with 100% of solvent B for
an additional 10 min, the column was equilibrated again with solvent A. The sample was co-injected with partial hydrolysates of colominic acid
to determine the degree of polymerization. Sialic oligomers and
polymers were detected by the adsorption at
A214 nm, whereas the oligosaccharides were
detected by determining the radioactivity of the effluent by
scintillation counting as described previously (4).
| |
RESULTS |
Predicted Amino Acid Sequence of Human ST8Sia III--
To isolate
cDNA encoding human ST8Sia III, we cloned the cDNA by reverse
transcriptase-PCR using human brain poly(A)+ RNA based on
the mouse ST8Sia III sequence (22), which was then amended by the
5'-RACE, resulting in the full-length ST8Sia III.
Fig. 1 shows the nucleotide and deduced
amino acid sequences of human ST8Sia III. The amino acid sequence of
human ST8Sia III has 34.8% and 33.3% identity with that of human
ST8Sia IV (PST, 17) and ST8Sia II (STX, 19, 23), respectively
(identical amino acid residues are shown in capital letters
in Fig. 1). The homology between ST8Sia III and ST8Sia IV or ST8Sia II
is extensive in both sialyl motif L and S (see double and
single underlines). However, this homology extends to the
whole region of a presumed catalytic domain, in particular to the amino
acid sequence close to the carboxyl terminus. The position of the
introns was deduced by comparison of the cDNA sequence with the
genomic sequence deposited into the GenBankTM data base (accession
number AC003971). This genomic sequence was identified in chromosome
18. Consistent with this assignment, we determined the ST8Sia III locus
at chromosome 18, q21.2 using fluorescence in situ
hybridization (data not shown). Human ST8Sia III is less similar to
other 2,8-sialyltransferases (29.8% identity with ST8Sia I (28) and
26.9% identity with human ST8Sia V (37)).
View larger version (48K):
[in this window]
[in a new window]
|
Fig. 1.
Nucleotide and translated amino acid
sequences of human ST8Sia III. The signal/membrane-anchoring
domain is denoted by a bold line. Sialyl motifs L and S are
denoted by double underlines and a single
underline, respectively. In the consensus sequence
(Cons.), amino acid residues identical to ST8Sia III, ST8Sia
IV, and ST8Sia II are shown in capital letters, whereas
those identical to ST8Sia IV and ST8Sia II are shown in small
letters. I-1, I-2, and I-3
indicate the positions of predicted introns. Potential
N-glycosylation sites are shown by
asterisks.
|
|
ST8Sia III Forms Polysialic Acid in Glycoproteins Other than
NCAM--
As a part of systematic studies on sialic acid polymers
synthesized by ST8Sia III, HeLa cells stably expressing ST8Sia III cDNA, HeLa-ST8Sia III, were examined. As shown in Fig.
2, permeabilized HeLa cells expressing
ST8Sia III and NCAM were heavily stained by 12F8 (38) or 735 antibody
(39), both of which are specific for polysialic acid. The staining
pattern showed a typical perinuclear Golgi staining. This staining
disappeared after endo-N treatment, which cleaves polysialic acid (data
not shown). The same cells without permeabilization were, however,
negative for anti-polysialic acid antibody staining (Fig. 2). The
results suggest that ST8Sia III forms polysialic acid on glycoproteins
other than NCAM, which was expressed on the cell surface. HeLa cells
stably transfected with ST8Sia IV or ST8Sia II cDNA were positive
before or after permeabilization, whereas HeLa cells expressing ST8Sia
I were negative for both intracellular and cell surface staining (Fig. 2). More than 50% of all these transfected cells expressed similar amounts of cell surface NCAM detected by immunofluorescent staining as
described (23).
View larger version (55K):
[in this window]
[in a new window]
|
Fig. 2.
Expression of polysialic acid directed by
ST8Sia IV (PST), ST8Sia II (STX), or ST8Sia III. HeLa cells stably
expressing ST8Sia IV, ST8Sia II, ST8Sia III, or ST8Sia I were stably
(for HeLa-ST8Sia IV and HeLa-ST8Sia II) or transiently (for HeLa-ST8Sia
III and HeLa-ST8Sia I) transfected with NCAM cDNA and examined
without (surface) or after permeabilization (cytoplasm) as described
(52). For cytoplasm (intracellular) staining for ST8Sia IV- and ST8Sia
II-transfected cells, the cells were treated with endo-N before
staining. 12F8 anti-polysialic acid antibody was used in these
stainings. The same results were obtained using 735 antibody. The
magnification for surface staining of ST8Sia IV - or ST8Sia
II-transfected cells was 400× and 630× (bar = 20 µm) for the rest.
|
|
ST8Sia III Catalyzes Autopolysialylation--
To determine if NCAM
is a substrate for ST8Sia III, NCAM·IgG chimeric protein prepared
from COS-1 cells was incubated in vitro with a soluble form
of ST8Sia III, ST8Sia IV or ST8Sia II, as described previously (4). The
results shown in Fig. 3A
indicate that ST8Sia III barely added [14C]sialic acid to
NCAM·IgG. In contrast, ST8Sia IV and ST8Sia II added a large amount
of [14C]sialic acid, resulting in a broad band indicative
of polysialic acid.
View larger version (35K):
[in this window]
[in a new window]
|
Fig. 3.
Incorporation of sialic acid into NCAM
and sialyltransferase themselves by a soluble form of ST8Sia
III, ST8Sia IV (PST), and ST8Sia II (STX). A,
NCAM·IgG was incubated with CMP-[14C]NeuNAc and a
soluble form of ST8Sia III, ST8Sia IV, or ST8Sia II or protein A only
(MOCK) for 18 h. B, a soluble chimeric form
of ST8Sia III, ST8Sia IV, or ST8Sia II, adsorbed into IgG-Sepharose
beads, was incubated with CMP-[14C]NeuNAc for 18 h.
After incubation, the beads were separated from the supernatant
(Sup.). [14C]NeuNAc incorporated into each
enzyme in the beads was analyzed before () and after (+) endo-N
treatment. The samples were separated by SDS-polyacrylamide gel
electrophoresis and subjected to fluorography.
|
|
To search for a favorable acceptor for ST8Sia III, we examined
autopolysialylation of ST8Sia III. It was demonstrated previously that
ST8Sia IV and ST8Sia II form polysialic acid on N-glycans attached to themselves (40, 41). This assay involves the incubation of
ST8Sia IV- or ST8Sia II-adsorbed beads with radiolabeled CMP-NeuNAc in
the absence of NCAM or other exogenously added acceptors. The results
shown in Fig. 3B demonstrate that ST8Sia III, adsorbed to
beads, formed polysialic acid that can be cleaved by endo-N. No
radioactive band in the beads was formed in the absence of ST8Sia III.
This radioactive polysialylated molecule was not derived from soluble
glycoproteins, which may be present in the incubation mixture since no
radioactive band was detected in the supernatant (Fig. 3B).
ST8Sia III is almost as good an acceptor as ST8Sia IV and ST8Sia II for
their respective autopolysialylation (Fig. 3B). Moreover,
the addition of increasing amounts of NCAM·IgG did not reduce
autopolysialylation by ST8Sia III, ST8Sia IV, or ST8Sia II (data not
shown), indicating that the enzymes themselves are efficient
substrates. This result was obtained probably because the catalytic
domain and acceptor sites are close to each other in
autopolysialylation. It is possible that one of the intracellular acceptors for ST8Sia III shown in Fig. 2 is ST8Sia III itself.
Polysialyltransferases Can Synthesize Polysialic Acid on
Oligosaccharides--
The results described so far indicate that
ST8Sia III shares properties with ST8Sia IV and ST8Sia II in forming
polysialic acid on themselves. ST8Sia III was shown to add sialic acid
residues to GM3 (NeuNAc23Gal14Glc1ceramide) and
GD3 (NeuNAc28NeuNAc23Gal14Glc1ceramide) (29). This result suggested that low molecular weight oligosaccharides might be utilized as acceptors by ST8Sia II and ST8Sia IV as well as by
ST8Sia III. Fig. 4 illustrates that
ST8Sia III, ST8Sia II, and ST8Sia IV efficiently form polysialic acid
on N-glycans prepared from fetuin (b,
f, and j). Compared with ST8Sia III and ST8Sia IV, ST8Sia II adds sialic acid much less efficiently to these low
molecular weight acceptors (Fig. 4, f, g, h). For these
experiments comparing three sialyltransferases, the amount of the
enzymes was first measured by Western blot analysis using antibodies
reacting with the protein A portion of each enzyme, as described
previously (4). The same amount of enzyme was then used for all
experiments described hereafter. Since oligosaccharides with more than
two sialic acid residues are not bound to a reverse-phase cartridge column (Sep-Pak), reaction mixtures were directly applied to Mono-Q column in these experiments.
View larger version (35K):
[in this window]
[in a new window]
|
Fig. 4.
Polysialic acid and oligosialic acid
formation by ST8Sia III, ST8Sia II (STX), and ST8Sia IV (PST) on
trisialyl tri-antennary N-glycan from fetuin and
synthetic oligosaccharides. N-Glycans from fetuin
(Fetuin triSA N), synthetic
NeuNAc26Gal14GlcNAc16Man16Man1octyl
(2,6N), or
NeuNAc23Gal14GlcNAc16Man16Man1octyl
(2,3N) (1 nmol each) were incubated for 18 h with
CMP-[14C]NeuNAc and the soluble form of ST8Sia III,
ST8Sia II, or ST8Sia IV, absorbed to beads. The supernatant containing
the reaction products was directly subjected to Mono-Q anion exchange
chromatography. At the top (a), the elution profile of
oligosialic and polysialic acid under the same chromatographic
conditions is shown. Polysialic acid eluted after fraction 130 contain
more than 60 sialic acid residues. CMP-[14C]NeuNAc
produced three radioactive peaks eluting at fractions 3-21, 22-33,
and 34-39 under these conditions. In e, i, and
m, the same reaction products in d, h,
and l were purified by a reverse-phase cartridge column
(Sep-Pak C18) and analyzed by the same Mono-Q column chromatography.
Those containing three or more sialic acids were not bound to a Sep-Pak
column under these conditions.
|
|
Strikingly, almost identical results were obtained when
NeuNAc23(or
6)Gal14GlcNAc16Man16Man1octyl, which mimics common N-glycan structures, was used as an acceptor (Fig. 4,
c, d, g, h, k,
and l). These radioactive peaks were susceptible to endo-N
treatment (data not shown), confirming that they are polysialylated products. The major peak of ST8Sia IV products contained as many as
50-60 polysialic acids. To determine if ST8Sia III, ST8Sia II, and
ST8Sia IV contribute to oligosialic acid synthesis, the products
derived from
NeuNAc23Gal14GlcNAc16Man16Man1octyl shown in Fig. 4 (d, h, and l)
were first purified by Sep-Pak as described previously (42) and
analyzed by HPLC. The results (Fig. 4, e, i, and
m) showed that all of these enzymes can form disialyl
products and that ST8Sia III produced about 10 times more disialyl
products (~1000 cpm) than ST8Sia II or ST8Sia IV. The results also
indicate that ST8Sia II is less efficient than ST8Sia IV in utilizing
oligosaccharides as acceptors (Fig. 4). In these experiments, we
noticed that polysialic acid products were also eluted with a high
concentration (1 M) of NaCl (fractions 145-160). It is
possible that some of these components are filamentous bundles of
polysialic acid reported recently (43).
We then tested if N-glycans derived from NCAM and colominic
acid serve as acceptors. The results shown in Fig.
5 demonstrated that ST8Sia IV utilized
these acceptors, although colominic acid was a much less efficient
acceptor than NCAM N-glycans. On the other hand, ST8Sia III
and ST8Sia II utilized those acceptors less efficiently than ST8Sia IV
(Fig. 5, a, b, c, and
d).
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 5.
Addition of polysialic acid to NCAM
N-glycans and colominic acid by ST8Sia III, ST8Sia II
(STX), and ST8Sia IV (PST). N-Glycans from NCAM and
colominic acid (average degree of polymerization = 30) (1 nmol each)
were incubated with CMP-[14C]NeuNAc and the soluble form
of the enzymes for 18 h at 37 °C. The reaction products were
directly subjected to Mono-Q anion exchange chromatography as described
in Fig. 4.
|
|
Monosialyl and Disialyl N-Acetyllactosamine Serve as Acceptors for
ST8Sia III, ST8Sia II, and ST8Sia IV--
It has been proposed
previously that polysialyltransferases may add polysialic acid more
efficiently on disialosyl structure, NeuNAc28NeuNAc23Gal14GlcNAcR, preformed by an
initiation enzyme on NeuNAc23Gal14GlcNAcR (44). If that
is the case, disialosyl structure should be a better acceptor than
monosialyl structure. To test this hypothesis, we took advantage of the
present finding that oligosaccharides can be utilized as an acceptor
for ST8Sia II and ST8Sia IV. The results shown in Fig.
6 demonstrate that both monosialyl and
disialyl N-acetyllactosamines worked well as acceptors for
ST8Sia III, ST8Sia II, and ST8Sia IV, although monosialyl
N-acetyllactosamine was a slightly better acceptor than
disialyl N-acetyllactosamine. The results clearly indicate that the formation of disialyl structures is not necessary for the
actions of these sialyltransferases.
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 6.
Addition of polysialic acid to monosialyl or
disialyl acceptor by ST8Sia III, ST8Sia II (STX), and ST8Sia IV
(PST). NeuNAc23Gal14GlcNAc1octyl
(Monosialyl LacNAc) and
NeuNAc28NeuNAc23Gal14GlcNAc1octyl
(Disialyl LacNAc) (1 nmol each) were incubated with
CMP-[14C]NeuNAc and the soluble form of the enzymes for
18 h at 37 °C. The reaction products were directly subjected to
Mono-Q anion exchange chromatography as described in Fig. 4.
|
|
ST8Sia III, ST8Sia II, and ST8Sia IV Differ Significantly in the
Efficiency of NCAM Polysialylation--
Taken together, the above
results indicate that polysialyltransferases and ST8Sia III share many
properties in transferring multiple 2,8-linked sialic acid residues
to oligosaccharide acceptors. On the other hand, in contrast to ST8Sia
II or ST8Sia IV, the results described earlier indicate that ST8Sia III
did not form polysialic acid on glycoproteins on the cell membrane
(Fig. 2) or on NCAM (Fig. 3). To compare the efficiency of
polysialylation between ST8Sia III, ST8Sia II, and ST8Sia IV, the
enzymes were incubated with the acceptors at a low concentration (60 pmol/100 µl) in a shorter period (2 h).
The results shown in Fig. 7 illustrate
that ST8Sia III formed very little oligosialic and polysialic acid on
NCAM, whereas both ST8Sia II and ST8Sia IV formed polysialic acid on
NCAM (Fig. 7, a, e, and i). ST8Sia IV
is more efficient in NCAM polysialylation than ST8Sia II, consistent
with the previous reports (4, 22). In contrast, ST8Sia III formed a
small amount of polysialic acid on 2-HS-glycoprotein, a
human counterpart of fetuin (45) (Fig. 7c, open
squares). After longer incubation, ST8Sia III formed more
polysialic acid on 2-HS-glycoprotein (data not shown).
ST8Sia III formed oligosialic acid on sialylparagloboside, whereas
ST8Sia IV and ST8Sia II were less efficient in forming oligosialic acid on the same acceptor (Fig. 7, d, h, and
l). These results combined together indicate that ST8Sia II
and ST8Sia IV are much more efficient in forming polysialic acid on
NCAM than ST8Sia III. On the other hand, ST8Sia II and ST8Sia IV are
slightly less efficient than ST8Sia III in utilizing
2-HS-glycoprotein and sialylparagloboside as
acceptors.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 7.
Comparison of polysialylation by ST8Sia III,
ST8Sia II (STX), and ST8Sia IV (PST). NCAM or
2-HS-glycoprotein (AHSG) containing 60 pmol
of N-glycans was incubated for 2 h with
CMP-[14C]NeuNAc and the soluble form of ST8Sia III,
ST8Sia II, or ST8Sia IV. After purification by membrane filtration, the
glycoproteins were digested with N-glycanase and analyzed by
Mono-Q anion exchange chromatography as described in Fig. 4. In
c, g, and k, the radioactivity of the same
samples was shown in two different scales. Similarly, 60 pmol of
N-glycans isolated from NCAM was incubated with the soluble
form of ST8Sia III, ST8Sia II, or ST8Sia IV under the same conditions
(NCAM N-glycan). The reaction mixture was directly subjected
to Mono-Q anion exchange chromatography as described in Fig. 4.
Sialylparagloboside (SPG, 60 pmol) was also incubated under
the same conditions. The products were first subjected to reverse-phase
column, those bound (open circles) and those unbound
(closed circles) were digested with endoglycoceramidase II,
and released oligosaccharides were analyzed by Mono-Q.
|
|
Finally, we tested N-glycans isolated from NCAM as acceptors
under the same conditions. The results demonstrated that ST8Sia IV
added only a small amount of polysialic acid, whereas no incorporation was found for ST8Sia III and ST8Sia II (Fig. 7, b,
f, and j). Only after a longer incubation of NCAM
N-glycans with ST8Sia IV, polysialic acid was formed on
N-glycans isolated from NCAM (Fig. 5e). These
results combined together indicate that ST8Sia II and ST8Sia IV add
polysialic acid to N-glycans attached to NCAM with a much
better efficiency than N-glycans isolated from NCAM.
| |
DISCUSSION |
The present study demonstrated that ST8Sia III, which is
moderately related to ST8Sia IV (PST) and ST8Sia II (STX), can also form polysialic acid in certain acceptors such as the glycans attached
to the enzyme itself (Figs. 2 and 3). This unexpected finding led us to
determine if ST8Sia III is also involved in polysialylation of NCAM.
In vitro incubation of ST8Sia III with NCAM·IgG chimeric
protein (Fig. 3) and metabolic labeling of NCAM chimeric protein in
ST8Sia III-expressing HeLa cells, as done in the same way described
previously (4), provided evidence that NCAM is a very poor, if at all,
acceptor for ST8Sia III. Fetuin was shown to be the best acceptor among
glycoproteins tested for mouse ST8Sia III (24), and
2-HS-glycoprotein, the human counterpart of fetuin, is
also a good acceptor for human ST8Sia III (Fig. 7). These results
combined together indicate that ST8Sia III adds oligosialic acid and
small amounts of polysialic acid to 2-HS-glycoprotein
and its related glycoproteins.
Our initial striking finding that ST8Sia III forms polysialic acid also
directed us to reexamine the acceptor specificity of ST8Sia II and
ST8Sia IV. Since ST8Sia III was shown previously to add 2,8-linked
sialic acids to low molecular weight acceptors (24), we tested various
oligosaccharides as acceptors for ST8Sia II, ST8Sia III, and ST8Sia IV.
Unexpectedly, ST8Sia II and ST8Sia IV formed polysialic acid on
synthetic 2,3- or 2,6-linked oligosaccharides, which mimics
common N-glycan structures, and N-glycans
isolated from NCAM and fetuin (Figs. 4 and 5). We demonstrated
previously that 2,6-linked sialic acid attached to
N-glycans in NCAM served as an acceptor, although it was
utilized less efficiently than 2,3-linked sialic acid in the same
acceptor (4). The present study extended these findings and further
demonstrated that even 2,3-sialyl, 2,6-sialyl, or disialyl
(NeuNAc28NeuNAc23) N-acetyllactosamine can
serve as an acceptor.
Previously, Kojima et al. (22) reported that
N-glycans derived from NCAM did not serve as acceptors for
ST8Sia II or ST8Sia IV, although the N-glycans were neither
isolated nor characterized in their studies. It is possible that those
previous results were obtained because the amount of
N-glycans used was too low to analyze. Similarly, it was
previously reported that colominic acid does not serve as an acceptor
for polysialyltransferases (30). The present study demonstrated that
ST8Sia IV can add polysialic acid to colominic acid, although it is a
poor acceptor (Fig. 5f). It is likely that this very low
incorporation of polysialic acid to colominic acid can be detected only
because the assay established in the present study is highly sensitive.
It has been demonstrated recently that disialyl structure
NeuNAc28NeuNAc23Gal14GlcNAcR is widely present in
N-glycans attached to various glycoproteins in porcine
embryonic brains (46). It has been also reported that -subunit of
sodium ion channel contains 5-10 residues of 2,8-linked sialic acid
per chain (47, 48). Our studies indicate that ST8Sia III, ST8Sia II,
and ST8Sia IV can add disialyl and oligosialic residues on acceptors
such as 2-HS-glycoprotein, oligosaccharides, and
glycolipids (Figs. 4 and 7). These results suggest that disialyl and
oligosialic acid structures in nature are synthesized by ST8Sia III as
well as by ST8Sia II and ST8Sia IV. However, further studies are
necessary to determine if there is a correlation between the presence
of disialic and oligosialic acid structures and the presence of these enzymes in various tissues. During these experiments, we noticed that
monosialyl N-acetyllactosamine was also formed (Fig. 4,
e, i, and m). This was probably
produced from a minor contamination of non-sialylated acceptor. This
finding is consistent with a previous report that ST8Sia IV can form
2,8-linked sialic acid on N-acetyllactosamine during
autopolysialylation (40).
The results obtained in the present study demonstrated that monosialyl
and disialyl N-acetyllactosamines serve almost equally as
acceptors (Fig. 6). Previously, it has been postulated that polysialic
acid may be synthesized on the first 2,8-linked sialic acid
preformed by an initiation enzyme on NeuNAc23Gal1
4GlcNAcR. This hypothesis, however, was based on the biosynthetic
mechanisms for polysialic acid synthesis in bacteria (44). A similar
conclusion was made based on structural determination of polysialic and
oligosialic acid present in different stages of trout egg development,
although the "initiation enzyme" was not assayed (49). Disialo and
oligosialic acid can be formed when the amount of ST8Sia III, ST8Sia
II, or ST8Sia IV is low (Figs. 4 and 7). Moreover, polysialic acid
synthases in vertebrates and bacteria share no resemblance in protein
structures and are judged to be independently evolved. If the
initiation enzyme plays an important role in polysialic acid synthesis
of vertebrates, a disialyl oligosaccharide should be a much better acceptor than a monosialyl acceptor, but no remarkable difference was
observed in the present study. The results obtained on ST8Sia II and
ST8Sia IV from various laboratories (4, 20-23) all indicate that
ST8Sia II or ST8Sia IV alone can sufficiently form polysialic acid from
2,3-linked or 2,6-linked sialic acid attached to
N-acetyllactosamine in vitro. In combination with
the results obtained in the present study, these results indicate that
the initiation enzyme, if present, may not play a critical role in
polysialylation of vertebrates in vivo.
We previously reported that ST8Sia I utilizes GM3 to produce GD3 and
GT3 rather than utilizes GD3 for GT3 generation (28). In addition, we
demonstrated here that ST8Sia III as well as ST8Sia II and ST8Sia IV
can form polysialic acid in vitro directly on 2,3-,
2,6-, or 2,8-linked sialic acid by a single enzyme, indicating that polymerization of 2,8-linked sialic acid is a unique and common
activity of the family of these enzymes. In contrast to low elongation
activity by ST8Sia I and ST8Sia III, polysialyltransferases have
evolved to synthesize long polymer of sialic acid. It is tempting to
speculate that each 2,8-sialyltransferase has a different capacity
to hold on oligosialic/polysialic acid, and once de novo synthesized sialic acid polymer reach a certain size, the enzyme may be
released from the enzyme-acceptor complex. ST8Sia IV may have the best
capacity to hold on to longer polysialic acid.
The present study also demonstrated that ST8Sia II and ST8Sia IV are
much more efficient in NCAM polysialylation than ST8Sia III. On the
other hand, ST8Sia IV can form oligosialic acid on oligosaccharide
acceptors as much as does ST8Sia III. We also found that in a short
incubation time, the addition of polysialic acid on oligosaccharide
acceptors including NCAM N-glycans is negligible compared
with polysialylation of NCAM by ST8Sia II or ST8Sia IV (Fig. 7). This
finding is consistent with the fact that NCAM is the almost exclusively
polysialylated glycoprotein in the brain (30). In invertebrates such as
Drosophila and Aplaysia, NCAM function is
attenuated by either expressing an anti-adhesive protein, Beat (50), or
removal of NCAM-like molecules from the cell surface by endocytosis
(51). In vertebrates, on the other hand, attenuation of NCAM adhesive
activity is achieved by polysialylation, thereby allowing complex
neural development such as defasiculation. Such a single universal
modification can be finely tuned by spatial and temporal-specific
expression of ST8Sia II and ST8Sia IV. These results, combined
together, strongly suggest that ST8Sia II and ST8Sia IV evolved in
vertebrates to form polysialic acid on NCAM, probably from a primordial
2,8-sialyltransferase that adds oligomers of 2,8-linked sialic
acid residues. Future studies will be of significance to determine how
the protein structure of NCAM influences the actions by ST8Sia II and
ST8Sia IV.
2,8-Sialyltransferases, ST8Sia IV (PST), ST8Sia II
(STX), and ST8Sia III*
,
TOP
ABSTRACT
INTRODUCTION
These authors contributed equally to this work.
