|
|
|
|||||||
|
|||||||||
(Received for publication, March
27, 1995; and in revised form, May 31, 1995) From the
Glycosylphosphatidylinositol (GPI) anchors of the yeast Saccharomyces cerevisiae have been reported to contain three
different types of side chains attached to the
The addition of glycosylphosphatidylinositol (GPI) (
Figure 1:
Glycosylphosphatidylinositol anchor
structures of yeast proteins. The scheme outlines the structural
variants found in S. cerevisiae(3) . The mannoses are
annotated by M1-M5. The majority of GPI anchors of wild type
cells contain only M1 to M4. A fifth mannose (M5) is present
only on part of the anchors and is linked either
Figure 2:
Sec18 cells were radiolabeled with myo-[2-
Figure 3:
Anchor glycopeptides were prepared from
radiolabeled X2180 cells, were divided into 4 equal aliquots, and were
subjected to sequential treatments in the order indicated on top of
each panel. N-Ac, N-acetylation. Resulting fragments
were sized by paper chromatography. Recoveries of counts/min were close
to 100% in all treatments.
The slower migrating peak was assumed to represent a
fragment with a fifth mannose (M5, Fig. 1) which is
predicted by the previous analysis of yeast anchors(3) . Its
identity was confirmed by the following observations: (i) when isolated
from a preparative paper chromatogram and treated with jack bean
In
summary, ER forms of GPI-anchored proteins contain the same four
mannoses as the candidate precursor lipids CP1 and CP2. Moreover, the
absence of M5 in secretion mutants which block vesicular traffic
between the ER and Golgi strongly suggests that the addition of M5
occurs in the Golgi.
To probe the
distribution of transferases involved in the addition of M5, we used sec7, a secretion mutant which blocks between early and mid
Golgi compartments (24) (Table 1). The disappearance of
In this context it should be noted that there is no
increase in
Figure 4:
1.5
As for many other glycan structures, the exact role of the
mannose side chain on GPI anchors is presently not well understood.
Nevertheless, single mannose side chains (M4) have been found in
mammals, D. discoideum, P. falciparum, and S. cerevisiae and therefore, the possibility to add
M4 seems to have been maintained during evolution in several phyla. On
the other hand, the GPI biosynthesis machinery, at least of mammals and
trypanosomes, does not require the presence of M4, since most mammalian
and trypanosomal cells do not contain this residue either on the
complete precursor lipids or on the GPI proteins. In contrast to M4, M5
residues have only been described in S. cerevisiae. In various
wild type strains, M5 residues are present in 18-32% of anchors,
and the ratio of The identity of the mannosyltransferases involved in the addition of
M4 and M5 are unclear at the moment. The general experience in
glycobiosynthesis is that each kind of linkage is achieved by a
different glycosyltransferase. Since M4, as M3, is The GPI anchor peptides
analyzed here were purified over octyl-Sepharose and hence contain a
lipid moiety. We conclude that the M5-transferases get access to M4
without any need for previous removal of the lipid moiety. This is in
agreement with the idea that the glycan part of GPIs can assume a
relatively extended configuration and form a broad platform between
protein and lipid(40) . Yet, we routinely find that
10-20% of the anchor peptides eluted from ConA-Sepharose do not
bind to octyl-Sepharose. Indeed it has recently been reported that for
some proteins the GPI anchor represents a necessary and sufficient
signal for their incorporation into the cell wall and that upon arrival
at the plasma membrane part of the GPI anchor including the lipid
moiety is removed(39, 41, 42) . Having
restricted our analysis to lipid-containing anchors, it is obvious that
the possible further additions of glycans onto the GPI core structure
or onto the GPI side chain during this incorporation process would have
escaped detection. Through analysis of N-glycan structures
of glycoproteins accumulating in sec7, sec14, sec18, and sec23, of Base-sensitive anchors have a tendency
to receive The studies with
mutants such as mnn9, van1, anp1, and erd1 demonstrate that their deficiency only abolishes the capacity of N-glycan elongation but does not eliminate other glycosylation
events of the Golgi such as addition of M5 to GPI proteins.
Mannosylation of inositolphosphoceramides seems to be intact, since the
pattern of inositolphosphoceramides and the mannosylated forms thereof
are normal in all of these mutants (not shown). Further studies will be
required to delineate the primary event leading to this specific
deficiency in N-elongation.
Volume 270,
Number 34,
Issue of August 25, pp. 19709-19715, 1995
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
1,2-linked mannose
of the conserved
protein-ethanolamine-PO
-Man1,2Man1,6Man1,4GlcNH
-inositol
glycan core. The possible side chains are Man1,2- or
Man1,2Man1,2- or Man1,3Man1,2- (Fankhauser, C.,
Homan, S. W., Thomas Oates, J. E., McConville, M. J., Desponds, C.,
Conzelmann, A., and Ferguson, M. A.(1993) J. Biol. Chem. 268,
26365-26374). To determine in what subcellular compartment these
side chains are made, we metabolically labeled GPI-anchored membrane
proteins with myo-[2-
H]inositol in
secretion mutants blocked at various stages of the secretory pathway
and analyzed the anchor structure of the labeled glycoproteins. When
the exit of vesicles from the endoplasmic reticulum or entry into the cis-Golgi were blocked in sec12 or sec18 cells, all anchors contained a side chain consisting of a single
1,2-linked mannose. GPI proteins trapped in the cis-Golgi
of sec7 contained Man1,3Man1,2- but no
Man1,2Man1,2- side chains. Mutants blocked at later stages of
the secretory pathway made increased amounts of side chains containing
two mannoses. Man1,2Man1,2- and Man1,3Man1,2- side
chains were preferentially associated with ceramide- and
diacylglycerol-containing GPI anchors, respectively. Mnn1, mnn2,
mnn3, mnn5, and mnt1(=kre2),i.e. mutants
which lack or down-regulate 1,2- and 1,3-mannosyltransferases used in
the elongation of N- and O-glycans in the Golgi, add
the fifth mannose to GPI anchors normally. The same conclusion was
reached through the analysis of deletion mutants in KTR1, KTR2,
KTR3, KTR4, and YUR1 which all are open reading frames
with high homology to MNT1. Mutants deficient in the Golgi
elongation of N-glycans such as anp1, van1, mnn9 are
deficient in the maturation of the N-glycans of GPI-anchored
glycoproteins, but process the GPI anchor side chain normally. Data are
consistent with the idea that the fourth mannose is added to proteins
as part of the anchor precursor glycolipid in the endoplasmic
reticulum, whereas the fifth mannose is added by not yet identified
1,3- and 1,2-mannosyltransferases located in the Golgi
apparatus.
)anchors to the carboxyl terminus of newly synthesized
polypeptides occurs as an early post-translational modification of
proteins entering the secretory pathway(1, 2) . While
the core carbohydrate structure linking the protein to the lipid moiety (Fig. 1) is conserved throughout eukaryotic evolution, many
organisms attach additional sugars and/or other groups to this core.
All GPI anchors of Saccharomyces cerevisiae contain a fourth
mannose residue (M4, Fig. 1) attached to the
1,2-linked mannose of the glycan core and part of yeast GPI
anchors also contain a fifth mannose (M5) which is linked
either 1,2 or 1,3 to M4 (3) . It is unknown if a
given protein is made with several different kinds of side chains or if
each protein is made with only one kind. The presence of a fourth,
1,2-linked mannose has also been found as a species- and
tissue-specific modification in mammalian GPI
anchors(4, 5) . Recently, the same 1,2-linked
mannose has also been found in Dictyostelium discoideum(6) and on a GPI glycolipid made by merozoites of Plasmodium falciparum(7) . Here we undertook to
exploit the well defined secretion and glycosylation mutants of S. cerevisiae in order to investigate the subcellular
localization and identity of the mannosyltransferases involved in the
biosynthesis of the mannose side chains of yeast GPI anchors.
1,2 or 1,3
to M4 (3) . The sites of cleavage obtained by HF
dephosphorylation and phosphatidylinositol-specific phospholipase C (PI-PLC) are indicated. Arrows point toward the
linkages which can be hydrolyzed by the 1,2-linkage-specific
exomannosidase from ASAM. EtNH, ethanolamine; R,
alkyl chain.
Yeast Strains and Materials
The secretion
mutants developed by Peter Novick and Randy Schekman (8, 9) were: HMSF1 MATa sec1-1, SF294-2B MATa sec7-1, SF226-1C MATa
sec12-4, HMSF169 MATa sec14-3, HMSF176 MATa sec18-1, with all of them being derived from
X2180-1A MATa SUC2 mal gal2 CUP1. Glycosylation mutants
were the ones described by C. Ballou (10) and P. Robbins et
al. (11): LB1-22D MAT mnn1 SUC2 mal gal2 CUP1,
LB1-16A MAT mnn2 SUC2 mal gal2 CUP1, LB54-3A MATa mnn3 and LB-65-5D MATa mnn5 are derived
from X2180. YAH 116 ura3-52 lys 2-801 ade2-101
trp1-
1 his3-200 mnt1::TRP1 is derived from YAH-115 which
is identical except that it has an intact MNT1 gene. Strains
containing disruptions in open reading frames which are highly
homologous to MNT1(=KRE2) were developed by M.
Jaquet and M. Lussier in SEY6210 leu2-3,112 ura3-52
his3200 lys2-801 trp1901 suc29. A quadruple
deletion contained leu2-3,112 ura3-52 his3200
lys2-801 trp1901 suc29 kre2::TRP1 yur1::HIS3 ktr1::LYS2
ktr2::URA3; strains containing single deletions were YBR1445 leu2-3,112 ura3-52 his3200 lys2-801
trp1901 suc29 ktr3::HIS3 and YBR1411 leu2-3,112 ura3-52 his3-200 lys2-801
trp1901 ktr4::HIS3(12) . Disruptants of ANP1 (RCY1 MATa leu2-3, 112 ura3-52 his3200
lys2-801 trp1901 suc29 anp1::LEU2) and VAN1 (RCY2 MATa leu2-3, 112 his3200 lys2-801
trp1901 suc29 van1::TRP1) constructed in SEY6210 were
obtained from R. Chapman and S. Munro(13) . ZY100 (MATa
ade2-101 leu2-3, 112 ura3-52 suc29 gal2
pep4::CAT) and ZY400 (MATa ade2-101 leu2-3, 112
ura3-52 suc29 gal2 pep4::CAT mnn9::URA3)(14) were
obtained from Vivian MacKay, ZYMOGENETICS. SEY2102 (MAT suc29, ura3-52 leu2-3, 112 his4) and a strain
with a deletion in ERD1 in this same background were obtained
from H. Pelham(15) . Cells were kept on YPD agar plates and
cultured on minimal SDC media containing salts, vitamins, and trace
elements but no myo-inositol(16) , 2% glucose as a
carbon source, 1% of casein hydrolysate, adenin and uracil (40
µg/ml). SDC medium of the same composition was used for
radiolabeling of cells. myo-[2-H]Inositol (20Ci/mmol) was
purchased from Amersham Corp. (Buckinghamshire, United Kingdom). All
glycosidases used (-mannosidases from jack bean and Aspergillus saito&ıuml;) were
purchased from the Oxford GlycoSystems, (Oxford, UK). ConA-Sepharose
and octyl-Sepharose CL-4B were from Pharmacia (Uppsala, Sweden).Preparation and Analysis of Radiolabeled GPI
Anchors
Unless indicated otherwise, all radiolabelings were done
under identical conditions at 24 °C except for temperature
sensitive secretion mutants which were concomitantly labeled at 37
°C. Exponentially growing cells were resuspended to give an
OD of 10 and labeled with myo-[2-
H]inositol (30 µCi/ml) for 40
min. At this stage the labeling medium was diluted with 4 volumes of
fresh SDC medium, and incubation was continued for another 90 min.
Cells were lysed and the glycoproteins were delipidated and adsorbed
onto ConA-Sepharose as described (17) (procedure C followed by
A). Control experiments showed that all labeled GPI proteins were
quantitatively adsorbed onto ConA-Sepharose. The Sepharose beads were
incubated with Pronase and resulting anchor peptides were purified over
octyl-Sepharose as described(18) . Anchor glycopeptides from up
to 150 OD units of cells were taken up in 100 µl of
10 mM Tris-HCl, pH 7.5, 0.2 mM EDTA, 20% propanol,
and treated with 0.1 units of PI-PLC for 3 h at 37 °C. Lipids were
removed by extraction with n-butanol. The lipid-free anchor
head groups were dephosphorylated by HF, treated with
-mannosidases, and analyzed by paper chromatography in
methylketone/pyridine/HO (20:12:11) as
described(19) . Before chromatography samples were N-acetylated and desalted (20) . Mild base-sensitive
lipid moieties were removed from anchor peptides by a 5-h incubation in
8 M NH in HO/CHOH (1:1) at
37 °C.
GPI Anchors of the Endoplasmic Reticulum Contain Four
Mannoses
Recently two very polar GPIs named CP1 and CP2 have
been identified in S. cerevisiae. CP1 and CP2 are present in
only very low amounts and are rapidly turning over; they contain a
single 1,2-linked mannose (M4) added onto the conserved core plus
a phosphoethanolamine group on M3 (18) (Fig. 1). The
presence of phosphoethanolamine suggests that they potentially might
represent the complete precursor GPIs used for GPI anchoring of newly
made proteins, a process which takes place in the
ER(21, 22) . If CP1 and CP2 are the donor lipids for
GPI anchoring, we predict that GPI proteins residing in the ER contain
four mannoses. To show this, temperature-sensitive secretion mutants
deficient in the vesicular transport from the ER to the Golgi were
metabolically labeled with myo-[H]inositol, the anchor peptides of
GPI proteins were isolated, treated with HF, and sized by paper
chromatography. As shown in Fig. 2, when the transport of
vesicles from ER to Golgi was blocked at 37 °C, the anchor peptides
of sec18 indeed contained four mannoses of which only one
could be removed by JBAM (Fig. 2, panels C and D). However, when sec18 cells were labeled at 24
°C, an additional, slower migrating peak was observed (Fig. 2, panel A). The additional, slower migrating
peak was also found in wild type cells, both at 24 °C as well as 37
°C Fig. 3, panel B). Results similar to the ones
shown for sec18 were also obtained in other mutants blocking
between ER and Golgi, namely sec12 (Table 1) and sec16.
H]inositol at 24 °C (panels A and B) or 37 °C (panels C and D). Glycopeptides were generated and either left untreated (panels A and C) or treated with jack bean
-mannosidase (JBAM) (panels B and D).
Subsequently all glycopeptides were dephosphorylated with hydrofluoric
acid (HF), resulting fragments were N-acetylated and sized by
paper chromatography and scintillation counting of 1-cm wide paper
strips. M3-M5 indicate the migrations of radiolabeled
Man-GlcNAc-Ins to Man
-GlcNAc-Ins standard
oligosaccharides (18) run in parallel on the same paper.
Labeling temperatures and the sequential order of treatments are
summarized at the top of each panel.
-mannosidase (JBAM), the slower migrating peak yielded a fragment
comigrating in thin layer chromatography with GlcNAc-Ins, not Ins (not
shown). (ii) When JBAM treatment of total anchor peptides was done before HF treatment, counts were quantitatively recovered in a
peak comigrating with the Man-GlcNAc-Ins standard
indicating that three mannoses were protected by an HF-sensitive group
also in this slower migrating peak (Fig. 3, panel C).
(iii) The HF fragment migrated much slower if N-acetylation
was omitted, thus indicating the presence of an amino group which is
typically found on the glucosamine of GPIs (Fig. 3, panel
A). (iv) Treatment of the total of anchor peptides with
-mannosidase from A. saitoi (ASAM) produced a
fragment comigrating with Man-GlcNAc-Ins whereby part of
the material proved to be resistant (Fig. 3, panel D).
These findings strongly argue that this additional HF fragment
represents a GPI structure. On the basis of the previous analysis of
the GPI anchor of mature GPI proteins from the same strain
(X2180)(3) , it seems safe to assume that the slower migrating
peak represents a mixture of
Man1,2Man1,2Man1,2Man1,6Man1,4GlcNAc1,6Ins
and
Man1,3Man1,2Man1,2Man1,6Man1,4GlcNAc1,6Ins,
the latter being resistant to ASAM. There is some variability among
different wild type strains with regard to the fraction of GPI anchors
containing a fifth mannose (M5, Fig. 1, Tables II and
IV). It should be noted that the percentages obtained for wild type
cells represent the status of GPI anchors in the Golgi and/or in
post-Golgi compartments. This can be stated because, when analyzed by
SDS-polyacrylamide gel electrophoresis, ER forms of GPI proteins are no
more detectable after a 2-h pulse labeling with myo-[2-H]inositol as was done in the
experiments described in here(23) . (ER forms of GPI proteins
have much lower molecular masses than more mature GPI proteins.)Addition of M5 Occurs in Early and Late Golgi
Compartments
The ASAM treatments of the type shown in Fig. 3appear to be exhaustive since they quantitatively
eliminated the Man-GlcNAc-Ins peak which can be considered
as an internal control. Therefore, performing an experiment as
described in Fig. 3, panels B and D, we could
obtain the percentage of 1,3-linked M5 from the percentage of
ASAM-resistant Man-GlcNAc-Ins and could calculate the
percentage of 1,2-linked M5 by subtracting the percentage of
ASAM-resistant Man-GlcNAc-Ins from the percentage of total
Man-GlcNAc-Ins present before ASAM treatment (Tables I-IV).
In a control experiment we preparatively isolated the
Man-GlcNAc-Ins peak from a paper chromatogram, treated it
with ASAM and reanalyzed the products by a further paper
chromatography. By this procedure we obtained the same ratio of
ASAM-resistant to ASAM-sensitive Man-GlcNAc-Ins as was
calculated from the ASAM digest of the total anchor peptide
fraction (containing Man-GlcNAc-Ins and
Man-GlcNAc-Ins as in Fig. 3, panels B and D). The relative contributions of 1,2- and
1,3-linked M5 seem to be quite variable among different wild type
strains, with most of the variability stemming from variations in the
amount of 1,2 linkages (Table 2, IV).
1,2-linked but not 1,3-linked M5 in sec7 upon shift
to 37 °C strongly suggests that an 1,3-mannosyltransferase is
encountered by newly made GPI proteins already in the earliest Golgi
compartment lying proximal to the sec7 block whereas the
1,2-mannosyltransferase is localized in later Golgi compartments
lying beyond the sec7 block. While interpreting this result
one should keep in mind that significant amounts of enzymes destined
for a distal compartment will accumulate proximally to a secretory
block that is maintained for some time. This implies that during a
prolonged secretory arrest in sec7, the cis-Golgi
compartment might actually take on characteristics of a mid or trans-Golgi compartment and carry out distal modifications.
Thus, the absence of 1,2-linked M5 in sec7 allows to
formally conclude that in wild type cells, 1,2-linked M5 is added
in the mid or trans-Golgi. On the other hand, in spite of the
persistence of 1,3-linked M5 in sec7, the corresponding
1,3-mannosyltransferase cannot be assigned unequivocally to the cis-Golgi since 1,3-mannosyltransferase might be a trans-Golgi enzyme which, if artificially retained proximal to
the sec7 block, assumes an active conformation. The trans-Golgi localization of the 1,3-mannosyltransferase
may appear less likely since the enzyme is not active if retained in
the ER proximal to a sec18 block and because the related
1,2-mannosyltransferase is not active if retained in the cis-Golgi.Late Sec Mutants Accumulate GPI Anchors Containing Five
Mannoses
Two alternative possibilities may be considered to
explain the fact that not all GPI anchors receive M5 residues in wild
type strains. (i) Only part of the GPI proteins might be substrates for
the M5-transferases, and thus the limiting factor might be some steric
or topological problem which restricts access of proteins to these
enzymes. (ii) Alternatively, the rapid transit of GPI proteins through
the Golgi might render the time for interaction with processing enzymes
limiting. In this latter case, the prolonged retention of proteins
proximal to a secretion block might increase the frequency of M5
addition.1,3-linked M5 in sec7 as compared to wild
type. Assuming that the corresponding 1,3-transferase is located
in the cis-Golgi also in wild type cells, this result suggests
that in wild type the exposure time of GPI proteins to
1,3-mannosyltransferase in the cis-Golgi is not limiting
or else that Sec7p is required to maintain the
1,3-mannosyltransferase of the cis-Golgi in a functional
state. Later secretion mutants (sec14 and sec1) seem
to enhance addition of M5 (Table 1), possibly because GPI
proteins remain for prolonged periods in contact with 1,3- and the
1,2-mannosyltransferases present in the later Golgi. It is
noteworthy that a block beyond the Golgi (sec1) induces an
increase in both 1,2- and 1,3-linked M5, whereas the
intra-Golgi block of sec14 results in a decrease of 1,2-
but a compensatory increase in 1,3-linked M5 (Table 1).
Since the sec14 block is supposed to be distal to the sec7 block, this might indicate that the sec14 block renders
the access of GPI proteins to the 1,2-mannosyltransferase
difficult and that part of the 1,2-transferase is located in a
very late Golgi compartment. Alternatively, the sec14 block
might delay transition of GPI proteins through the early and middle
Golgi, thus increasing the time of exposure of the anchors to
1,3-mannosyltransferase(s) so that less substrate would be left
for the 1,2-mannosyltransferase.Involvement of Known Mannosyltransferases in Side Chain
Addition
MNN1 encodes for a Golgi
1,3-mannosyltransferase which adds terminal mannoses onto O- and N-glycans (14, 28) whereas MNN2, MNN3, and MNN5 control the addition of
1,2-linked mannoses onto the outer chains of N-glycans in
the Golgi(29, 30, 31) . MNT1 (=KRE2) encodes for a Golgi
1,2-mannosyltransferase which adds the third mannose of O-glycans. YUR1, KTR1, KTR2, KTR3, and KTR4 were identified as open reading frames with considerable homology
to MNT1(12, 32, 33) . As shown in Table 2, normal amounts of 1,3-linked M5 are added in mnn1 and similarly, no reduction in 1,2-linked M5 was
observed in mnn2, mnn3, mnn5, and mnt1 relative to
the corresponding wild types. Also, single or combined deletions of YUR and KTR sequences produced at most a moderate
(1.5-fold) reduction in 1,3-linked M5 (ktr3). Thus, none
of these genes seems to be essential for addition of either
1,2-linked or 1,3-linked M5 residues. YUR and KTR deletion mutants also made normal amounts of
mannosylinositolphosphoceramide(34) , thus indicating that
these open reading frames are not essential for the GDP-Man-dependent
mannosylation of inositolphosphoceramide in the Golgi (data not shown) (35) . The 1,3-linked M5 is significantly reduced in mnn3, a mutant with a general shortening of N- and O-linked glycans(10) . The fact that only addition of
1,3- but not 1,2-linked M5 is reduced may indicate that the
pleiotropic mutation of this strain affects mainly the early Golgi
compartments.Correlation between the Type of Side Chain and the Lipid
Moiety of GPI-anchored Proteins
GPI-anchored proteins are made
with two different lipid moieties, mild base-resistant ceramides, and
mild base-sensitive diacylglycerols(3, 17) . To see
whether both types of anchors were associated with all types of side
chains, mild base-sensitive and mild base-resistant anchor peptides
were analyzed separately as described in Table 3. The results
indicate that all types of side chains can be associated with both
lipid moieties. Yet, the ratio of 1,3-linked M5 over
1,2-linked M5 anchors is different in base-sensitive and
base-resistant anchors, namely 2.8 (X2180), 5.1 (SEY6210), or infinite
(ZY100) in base-sensitive but only 0.5 (X2180), 0.23 (SEY6210), or 0.07
(ZY100) in base-resistant anchors ( Table 3and Table 4).
Thus, base-sensitive anchors have a tendency to receive 1,3-linked
M5 whereas base-resistant anchors preferentially are provided with an
1,2-linked M5 in all strains.
Mnn9 Abolishes N-Glycan Elongation but Not Addition
of M5 on GPI-anchored Proteins
There exist several mutants (mnn9, anp1, van1, erd1, and pmr1) which appear to be
unable to elongate N-glycans in the Golgi because they do not
add 1,6-linked mannose onto the N-glycans of secretory
proteins such as invertase and
mannoproteins(10, 13, 15, 36) . It
is unlikely that these mutants are deficient in some
glycosyltransferase, but it is rather suspected that their problem
might originate from some abnormality of protein trafficking, from an
anomalous Golgi organization, from an inability to retain
glycosyltransferases in the Golgi, or from a disturbed ionic
environment in the Golgi(14, 37, 38) . MNN9, VAN1, and ANP1 are homologous, and Anp1p has
been shown to reside in the ER(13) . In studies concerning the
functional organization of the Golgi of such mutants, it obviously is
not possible to rely on the N-glycan maturation as a criterion
for the normal progression of proteins through the Golgi apparatus. We
thus exploited the above established Golgi localization of
M5-transferases as a means to assay the traffic of GPI proteins through
the Golgi in these mutants. To demonstrate that their defects affect
not only soluble but also GPI proteins, mnn9, anp1, and van1 were labeled with myo-[2-H]inositol and GPI proteins were
visualized by SDS-polyacrylamide gel electrophoresis/fluorography. As
shown in Fig. 4, these mutants make large amounts of GPI
proteins which are very much smaller than in wild type cells. This is
consistent with the view that the N-glycans of GPI proteins,
as the ones of invertase(13) , are not elongated in mnn9 and van1 and are only partially elongated in anp1. It also is evident that the labeling of GPI proteins in
mutant cells is significantly enhanced, possibly because the transit of
GPI proteins through the Golgi and the ensuing removal of the anchor at
the cell surface is slowed down in these mutants(39) .
(Invertase secretion by mnn9 has been reported to be slowed
down(37) .) Analysis of the GPI anchors shows that all mutants
can add 1,2-linked as well as 1,3-linked M5 (Table 4).
In all elongation mutants the proportion of base-sensitive anchors is
diminished, and the percentage of 1,2-linked M5 is reduced. Beyond
these similarities two mutants show quite specific changes which are
not to be found in others: Mnn9 shows a drastic increase in
1,3-linked M5 in the base-resistant anchor fraction, thus
accumulating a normally quite uncommon kind of anchor; anp1 has a severely reduced amount of M5 with losses occurring in both
the mild base-sensitive 1,3-linked M5 as well as the mild
base-resistant 1,2-linked M5; in van1 and erd1 the changes are less extensive. While the alterations in mnn9 are difficult to interpret it is likely that the drastic reduction
of M5 in anp1 results from an inability to efficiently retain
M5-transferases in the Golgi apparatus(13) . Whatever the cause
of the deficiency in N-glycan elongation of mnn9,
van1, anp1, and erd1 may be, the GPI proteins do
get into contact with the M5-transferases normally found in the Golgi
and their GPI anchor can be matured, although the N-glycans on
the same GPI proteins remain immature. Data however cannot tell whether
these M5-transferases really reside in their normal location in these
mutants.
10
cells from
different insertional mutants in N-glycan elongation and
corresponding parental cells were labeled for 2 h with 20 µCi of myo-[2-H]inositol at 24 °C. Proteins
were extracted and analyzed by SDS-polyacrylamide gel
electrophoresis/fluorography. Exposure was for 1 week. The GPI anchor
maturation of these mutants is summarized in Table 4.
1,2- versus 1,3-linked mannose
varies from 1:4 to 2.5:1, depending on the strain (Table 2). 1,2-linked, it
is conceivable that the same mannosyltransferase is responsible for the
addition of both M3 and M4. On the other hand, the M5 adding
mannosyltransferases are definitely different from the ones that add M4
since, according to our data with secretion mutants, they reside in the
Golgi whereas the M4 addition must occur in the ER. Our data show that
none of a panel of cloned Golgi mannosyltransferases or genes
regulating such transferases is essential for the addition of
1,2-linked or 1,3-linked M5. The slight reduction in
1,3-linked M5 observed in mnn3, ktr3, and ktr4 mutants might be taken as an indication that all of these enzymes
are involved in the addition of 1,3-linked M5, but this would be
against the general ``one linkage-one enzyme'' rule mentioned
before. Also, this seems unlikely in view of the fact that KTR3 shows much less homology with KTR4 than with KTR1 which latter is without influence on the addition of M5 (Table 2)(12) . Thus, although we cannot exclude that
redundant enzymes are responsible for the addition of M5, it seems
more likely that M5 addition is due to the presence of some other, yet
unknown Golgi mannosyltransferases.-factor maturation events in these secretion
mutants and through subcellular fractionation studies, the yeast Golgi
could be divided into two distinct early (cis) compartments containing
1,6-mannosyltransferases for the elongation of N-glycans,
a later (mid) compartment containing the N- and O-glycan elongating 1,3-mannosyltransferase coded for by MNN1, and an even later (trans) compartment containing the
processing protease coded for by KEX2(24, 25, 26, 27) . The
presence of 1,3-linked or 1,2-linked M5 residues on GPI
anchors can now serve as an alternative means for tracking GPI proteins
and their vesicular flow through the Golgi. The presence of
1,2-linked M5 indicates that a GPI protein has reached or passed
the later Golgi compartments whereas the addition of 1,3-linked M5
in sec7 cells indicates that a GPI protein has reached the cis-Golgi. Subcellular fractionation studies will be required
to decide whether the presence of an 1,3-mannosyltransferase is a
distinguishing feature of the cis-Golgi also in wild type
cells. It should be noted that early and late Golgi modifications of
GPI anchors are independent or even mutually exclusive events, whereas
in the N-glycan elongation, addition of 1,6-mannose in
the early Golgi is a prerequisite for the addition of 1,3-linked
mannose in the later Golgi.1,3-linked M5 whereas base-resistant anchors
preferentially are provided with an 1,2-linked M5. The bias of
1,3-linked M5 for base-sensitive anchors can be interpreted in
several alternative ways which at present are not easy to distinguish
experimentally. (i) 1,3-Mannosyltransferases might prefer
diacylglycerol-based anchors over ceramide-based ones, and the inverse
might be true for the 1,2-mannosyltransferase. (ii) Since lipid
moieties of GPI anchors might still get exchanged in the
Golgi(17) , the Golgi lipid exchange enzyme might introduce
ceramides preferentially on anchors with 1,2-linked M5. (iii) GPI
proteins might get sorted according to their lipid domain, and
diacylglycerol-based GPI proteins might have more access to the
1,3-mannosyltransferase of the early Golgi whereas ceramide-based
ones have better access to the 1,2-mannosyltransferase of the late
Golgi. (iv) Diacylglycerol-based GPI proteins might transit more slowly
through the Golgi than ceramide-based ones so that the cis- or
mid Golgi 1,3-mannosyltransferases have more time to attach a M5
and that the 1,2-mannosyltransferase of the late Golgi cannot find
any suitable substrate any more. Enzyme specificities invoked in i and
ii are rendered less likely by the high amounts of 1,3-linked M5
found on ceramide-based anchors of mnn9. This latter finding
clearly shows that either 1,3-mannosyltransferase can act on
ceramide-based anchors or that the lipid exchange enzyme can act on
anchors with 1,3-linked M5. Sorting of GPI proteins according to
their lipid moiety (possibilities iii and iv) would represent a novel
mechanism. However, it has previously been demonstrated that the
vesicular transport of yeast GPI proteins from ER to Golgi is dependent
on ceramide biosynthesis whereas the vesicular flow of membrane
proteins containing a hydrophobic, membrane spanning sequence is not
dependent on ceramide biosynthesis(43) .
)-mannosidase; CP, complete
precursor; ConA, concanavalin A; HF, hydrofluoric acid; JBAM, jack bean
-mannosidase; PI, phosphatidylinositol; PI-PLC,
phosphatidylinositol-specific phospholipase C; ER, endoplasmic
reticulum.
We thank Clinton Ballou, Rowan Chapman, Michel Jaquet,
Marc Lussier, Vivian MacKay, Sean Munro, Hugh Pelham, Phil Robbins,
and Randy Schekman for yeast strains.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Orlean and A. K. Menon Thematic review series: Lipid Posttranslational Modifications. GPI anchoring of protein in yeast and mammalian cells, or: how we learned to stop worrying and love glycophospholipids J. Lipid Res., May 1, 2007; 48(5): 993 - 1011. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Nishikawa, J. B. Poster, Y. Jigami, and N. Dean Molecular and Phenotypic Analysis of CaVRG4, Encoding an Essential Golgi Apparatus GDP-Mannose Transporter J. Bacteriol., January 1, 2002; 184(1): 29 - 42. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Reggiori and A. Conzelmann Biosynthesis of Inositol Phosphoceramides and Remodeling of Glycosylphosphatidylinositol Anchors in Saccharomyces cerevisiae Are Mediated by Different Enzymes J. Biol. Chem., November 13, 1998; 273(46): 30550 - 30559. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. C. Rayner and S. Munro Identification of the MNN2 and MNN5 Mannosyltransferases Required for Forming and Extending the Mannose Branches of the Outer Chain Mannans of Saccharomyces cerevisiae J. Biol. Chem., October 9, 1998; 273(41): 26836 - 26843. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Sekiya-Kawasaki, D. Botstein, and Y. Ohya Identification of Functional Connections Between Calmodulin and the Yeast Actin Cytoskeleton Genetics, September 1, 1998; 150(1): 43 - 58. [Abstract] [Full Text]
|
|
|
|
|
|
R. Kollar, B. B. Reinhold, E. Petrakova, H. J. C. Yeh, G. Ashwell, J. Drgonova, J. C. Kapteyn, F. M. Klis, and E. Cabib Architecture of the Yeast Cell Wall. beta (1right-arrow 6)-GLUCAN INTERCONNECTS MANNOPROTEIN, beta (1right-arrow 3)-GLUCAN, AND CHITIN J. Biol. Chem., July 11, 1997; 272(28): 17762 - 17775. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Lussier, A.-M. Sdicu, A. Camirand, and H. Bussey Functional Characterization of the YUR1, KTR1, and KTR2 Genes as Members of the Yeast KRE2/MNT1 Mannosyltransferase Gene Family J. Biol. Chem., May 3, 1996; 271(18): 11001 - 11008. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |