Separation and characterization of two alpha 1,2-mannosyltransferase activities from Saccharomyces cerevisiae.

Abstract

Two GDP-mannose-dependent mannosyltransferase activities (designated M1MT-I and M2MT-I) from Triton X-100 extracts of Saccharomyces cerevisiae mnn1 microsomes were separated by concanavalin A lectin chromatography and partially purified. The two transferases were distinguished by differences in concanavalin A affinity and in carbohydrate acceptor specificity. Analyses of the reaction products indicate that both enzymes are alpha 1,2-mannosyltransferases. M1MT-I utilizes mannose or methyl-alpha-mannoside as acceptor while M2MT-I catalyzes the transfer of mannose from GDP-mannose to unsubstituted nonreducing alpha 1,6-linked mannose residues in the acceptor molecule. M2MT-I activity correlates with the presence of a single alpha 1,2-linked mannose residue at the nonreducing terminus of mnn2mnn9 and mnn2mnn10 outer chain oligosaccharides, and the enzyme may be involved in regulating outer chain elongation.

Full text

Page 1

THE JOURNAL
0 1991 by The American Society for Biochemistry and Molecular Biology, Inc
OF BIOLOGICAL
CHEMISTRY
Vol. 266, No. 13, Issue of May 5, pp. 8255-8261,1991
Printed in U.S.A.
Separation and Characterization of Two a1,2-Mannosyltransferase
Activities from Saccharomyces cerevisiae*
(Received for publication, October 25, 1990)
Mark S. Lewis and Clinton E. BallouS
From the Deoartment of Molecular and Cell Biology. Division of Biochemistry and Molecular Biology, University of California,
Berkeley, Caiifornia 94720
I_
.
Two GDP-mannose-dependent mannosyltransferase
activities (designated MIMT-I and M2MT-I) from Tri-
ton X-100 extracts of Saccharomyces cerevisiae mnnl
microsomes were separated by concanavalin A lectin
chromatography and partially purified. The two trans-
ferases were distinguished by differences in concana-
valin A affinity and in carbohydrate acceptor specific-
ity. Analyses of the reaction products indicate that both
enzymes are al,2-mannosyltransferaes. MIMT-I uti-
lizes mannose or methyl-a-mannoside as acceptor
while M2MT-I catalyzes the transfer of mannose from
GDP-mannose to unsubstituted nonreducing a1,6-
linked mannose residues in the acceptor molecule.
M2MT-I activity correlates with the presence of a sin-
gle al,a-linked mannose residue at the nonreducing
terminus of mnn2mnn9 and mnn2mnnlO outer chain
oligosaccharides, and the enzyme may be involved in
regulating outer chain elongation.
Glycosyltransferases responsible for glycoprotein
thesis in eukaryotes are localized to internal membranes,
particularly the endoplasmic reticulum and the Golgi appa-
ratus. In the yeast Saccharomyces cerevisiae, the endoplasmic
reticulum-localized transferases synthesize the N-linked
“core” oligosaccharide, and the Golgi-localized transferases
synthesize the oligosaccharide “outer chain” (Kukuruzinska
et ai, 1987).
The N-linked oligosaccharide structure of S. cereuisiae has
been determined in considerable detail (Fig. 1) (Hernandez et
al., 1989). The core portion is identical with the high mannose
oligosaccharides of higher eukaryotes, except that it is modi-
fied in yeast by the addition of several al,3-mannose units
(Alvarado et al., 1990). The outer chain portion consists
al,6-linked polymannose chain (the “backbone”) with
chains (“branches”) of one or two al,2-linked mannoses, some
of which are substituted with mannosylphosphate (Thieme
and Ballou, 1971; Rosenfeld and Ballou, 1974; Hernandez et
al., 1989; Ballou et al., 1990) or al,3-linked mannose (Lee and
Ballou, 1965; Nakajima and Ballou, 1974).
Although several of the glycosyltransferases involved in
core oligosaccharide synthesis have been characterized and a
few have been cloned (Albright and Robbins, 1990; Barnes et
al., 1984), none of the enzymes involved in outer chain bio-
biosyn-
of an
side
* This work was supported in part by National Science Foundation
Grants DCB-87-03141 and DCB-90-03409 and by a grant from the
L. P. Markey Charitable Trust. The costs of publication of this article
were defrayed in part by the payment of page charges. This article
must therefore be hereby marked “advertisement” in accordance with
18 U.S.C. Section 1734 solely to indicate this fact.
f To whom correspondence and reprint requests should
dressed.
be ad-
synthesis has been purified. Consequently, the pathway for
outer chain biosynthesis has been inferred from the known
N-linked oligosaccharide structure and from the effects of the
mnn mutations on that structure (Raschke et
et al., 1980; Ballou et al., 1980). Collectively, the mnnl, mnn2,
mnn5, and mnn9 strains define at least 6 steps that appear to
require distinct mannosyltransferases. The mnnl mutation
eliminates terminal al,3-linked mannose
N-linked oligosaccharide, while the mnn2 and mnn5 oligosac-
charide structures lack the first and second rows, respectively,
of al,2-linked mannoses in the outer chain, indicating that
outer chain branching requires two independent a1,2-man-
nosyltransferases. Since a1,2 branching still
nonreducing terminus of the backbone in the mnn2 mutant,
we infer the existence of an additional mnn2/mnn5-independ-
ent al,2-transferase (Gopal and Ballou, 1987; Hernandez et
al., 1989; Ballou et al., 1989). Finally, studies of the mnn9
mutant have led to the identification of two a1,G-mannosyl-
transferase activities, one that initiates outer chain synthesis
(Romero and Herscovics, 1989) and one that acts to elongate
the a1,G-backbone (Gopal and Ballou, 1987; Flores-Carreon
et al., 1990). Whereas extracts of the mnnl mutant show a
reduced a1,3-mannosyltransferase activity (Nakajima and
Ballou, 1975), extracts of the mnn2 and mnn9 strains are not
deficient in any activity that would explain the mutant phe-
notype (Gopal and Ballou, 1987).
In the present study, we describe the solubilization, sepa-
ration, and partial purification of two outer chain al,2-man-
nosyltransferases, MIMT-I and M,MT-I, from S. cereuisiae
microsomes. The two transferases are distinguished on the
basis of acceptor specificity and the structure of the products
they form. The methyl-a-mannoside-dependent transferase
activity (MIMT-I) has been described previously (Lehle and
Tanner, 1974; Parodi, 1979) and may be a general outer chain
branching enzyme. The M,MT-I activity
properties suggest a role in regulating outer chain elongation.
al., 1973; Cohen
from all parts of the
occurs at the
is novel, and its
EXPERIMENTAL PROCEDURES
Yeast Strains and Growth Conditions-S. cereuisiae MATa mutant
strains mnnl (LB1-22D), mnnlmnn.3 (LB202-8A), mnnlmnn2 (LB1-
33B), mnnlmnn5 (LB74-3Bj, and mnnlmnn9 (LB2008-4Aj were
obtained from laboratory stocks, courtesy of Lun Ballou. All strains
were grown in YPD media (1% yeast extract, 2% peptone, 2% glucose,
Difco Laboratories), except mnnlmnn9, which was grown in YPD
plus 1 M sorbitol to minimize lysis. Cultures were grown at 30 “C to
early or midlog phase (2-6 A,, units) and harvested
Materials and Methods-Triton
from RPI International, Dowex 1 C1- and Dowex 50 H’ resin (200-
400 mesh), acrylamide and bisacrylamide, Ultrapure Tween 20, and
the colloidal gold stain were from Bio-Rad, as was the phenylmercury
agarose (Hg-agarose) resin (Affi-Gel 501). Sodium borohydride (98%
pure) was from Aldrich. GDP and GDP-mannose were Sigma prod-
ucts, while GDP-[“C]mannose (236.9 mCi/mmol) was purchased
from Du Pont-New England Nuclear. Nitrocellulose sheets (0.4 Gm)
by centrifugation.
X-100 (scintillation grade) was
8255

Page 2

8256
Yeast a1 -+ 2-Mannosyltransferases
OUTER CHAIN
'.* CORE
"I I
ma3 : Reduces average chain lenth
&, d:
S i m i l a r t o m
m:
x = 1 - 4
Wild T V D ~
: x = -10
* : Additional phosphorylation sites
FIG. 1. Chemotypes of mannoprotein mutants. The revised
structure of S. cereuisiae X2180 mannoprotein carbohydrate chains
is shown with an indication of the effects of mnn mutations. In the
mnn2 or mnn2 mnnlO mutant mannoprotein, the italicized mannose
shown in bold font in the core is absent
nonreducing end of the outer chain. In the mnn9 or mnn2 mnn9
mannoproteins, the outer chain is shortened to one mannose, and the
bold italicized mannose is present. Thus, the addition of this a1,2-
linked mannose is not regulated by the MNN2 locus. Outer chain
phosphorylation is reduced by the mnnl and mnn6 mutations, and
additional sites of phosphorylation (indicated by asterisks) are ob-
served in mnnl mnn9 and mnnl mnn2 mnnlO cells and may be
present in wild-type mannoproteins.
OLIGOSACCHARIDES
BASE-LABILE
and is located at the
were obtained from Schleicher and Schuell. A.C.S. grade methanol
(Fisher Scientific) was dried with Molecular Sieve 3A (3:1, v/w;
Aldrich). al,2-Mannosidase was purified from Aspergillus satoi as
described (Ichishima et al., 1981).
All polysaccharide acceptors were from laboratory stocks, except
the methylglycosides, which were purchased from Sigma. Manlo-
GlcNAc was isolated from endo H-treated mnnlmnn9 mannan (Tsai
et al., 1984), and Mang-GlcNAc was obtained by combined endo H
and endo-al,6-mannanase digestion of mnnlmnn2 mannan (Cohen
et al., 1982). Unsubstituted al,g-linked oligosaccharides (M, to M4)
were obtained by endomannanase digestion or partial acetolysis of
mnn2 mannan (Nakajima et al., 1976). The dcM' tetrasaccharide
[aMan + '(aMan + ')aMan + 'apMan] was purified from an
endomannanase digest of mnnlmnn.3 cell mannan (Cohen et al.,
1980), and the nrM tetrasaccharide [aMan + 'aMan + 'aMan -+
fiapMan] was purified from the endomannanase digest of
mnnlmnn2mnnlO cell mannan (Fig. 2) (Ballou et al., 1989). a1,2-
Linked di- and trisaccharides were isolated from partial acetolysates
of mnn5 and mnnl mannan, respectively. D-Mannose was from
Sigma, and methyl a-D-mannoside was from Aldrich. All acceptors
were purified by gel filtration on Bio-Gel P-2 or P-4 columns and
were at least 90% pure by DIONEX anion exchange HPLC.
Gel filtration and aminospherisorb HPLC were performed with
two Waters model 510 extended flow pump heads, a Waters gradient
controller, and a Spectroflow 773 UV detector (AB1 Kratos). The gel
filtration column was a Bio-Rad TSK 20-XL model (7.5 mm X 300
mm), and the aminospherisorb column (4.6 mm X 250 mm) was from
Supelco. All HPLC solvents were degassed for 20 min under vacuum
(water aspirator) and sparged with helium before use. Glass-distilled
water and HPLC-grade acetonitrile (Fisher) were used for HPLC.
Anion exchange HPLC was performed on a DIONEX BioLC
system with a CarboPak PA1 column controlled by a Spectrophysics
SP4270 integrator. Carbohydrate was detected with a pulsed amper-
ometric detector using the electrode potential settings El = 0.00 V,
' The abbreviations used are: dcM, disubstituted central mannose;
nrM, nonreducing terminal mannose; reM, reducing mannose; HPLC,
high performance liquid chromatography; SDS, sodium dodecyl sul-
fate; ConA, concanavalin A.
E2 = 0.67 V, and E3 = -0.85 V. The elution buffers contained NaOH
and NaOAc in glass-distilled water and were sparged with helium
during the operation.
SDS-polyacrylamide gel electrophoresis was performed according
to Laemmli (1970). During enzyme purification, proteins were stained
in polyacrylamide gels by conventional silver stain (Morrissey, 1981)
or periodate-silver stain (Dubray and Brezard, 1981). Protein was
concentrated, and Triton X-100 was removed for
chloroform/methanol precipitation method of Wessel and Flugge
(1984). It was sometimes necessary to concentrate dilute samples
before precipitation by loading 2 ml of the protein solution into a
Centricon 30 microconcentrator (Amicon), centrifuging it in a refrig-
erated rotor at 4000 X g for 1-2 h, and collecting the material that
did not pass through the filter (30-kDa cutoff).
Mannosyltransferase Assay-The assay mixture contained solubi-
lized membrane protein, acceptor saccharide, and 1 PM GDP-["C]
mannose donor (0.05 PCi) in 200 pl of assay buffer (50 m M Tris-HC1,
5 m M Mn2+, 0.1-0.3% Triton X-100, pH 7.5), and it was incubated
for 60 min at 30 "C. Control tubes lacked acceptor saccharide. The
reaction was initiated by the addition of radiolabeled GDP-mannose
and terminated by the addition of 0.8 ml of
mixture was then passed over a Dowex 1 C1- column (0.5 g of resin)
to remove unreacted GDP-mannose, and the column was washed
once with 1 ml of 20 mM acetic acid. A portion of the effluent (0.5
ml) was added to 4 ml of Scint A fluid (Packard Technologies) and
counted for radioactivity in a Packard Tri-Carb 2000CA scintillation
counter.
Mannosyltransferase Purification-S. cereuisiae mnnl cells were
grown to stationary phase in 200 ml of YPD media at 30 "C, and the
culture was used to inoculate 12 liters of YPD media in a 16-liter
fermentor. The culture was grown with moderate aeration and agi-
tation to early log phase (4.0 A,jw units), then harvested in a Sharples
centrifuge. A typical yield was 2.5-5.0 g wet weight of cells per liter
of culture. The harvested cells were washed by centrifugation in 100
ml of 1% KC1 and resuspended in ice-cold digestion buffer (30 m M
Tris-HC1,3 m M MgCl,, 0.5% glycerol, 1.0% P-mercaptoethanol, 1 m M
phenylmethylsulfonyl fluoride, pH 7.5). The cell suspension (30 ml)
was added to 40 g of acid-washed glass beads in an 80-ml Braun vessel
and homogenized 3 times for 1 min each. The vessel jacket was cooled
by 7-s bursts of liquid CO, for every 15 s of homogenization time,
with 1-min intervals between each 1-min homogenization to prevent
overheating. After homogenization, the material was centrifuged at
low speed (5000 X g) for 20 min in a Beckman J-21C centrifuge at
4 "C. The low speed supernatant was recovered and recentrifuged for
20 min at 15,000 X g. The supernatant was recovered and centrifuged
at 100,000 X g for 1 h at 6 "C in a Beckman ultracentrifuge (model
L8-70M) using a Beckman Type 40 rotor. The high speed supernatant
was discarded, and the tube was carefully rinsed four times with water
without disturbing the pellet, after which the pellet was resuspended
in cold solubilization buffer (50 m M Tris-HC1, 1% Triton X-100, 1
m M phenylmethylsulfonyl fluoride, pH 7.5). The membrane proteins
were solubilized in this buffer for 16 h at 4 "C, and then the solution
was centrifuged at 100,000 X g for 60 min at 6 "C. The supernatant
was recovered and assayed for protein (Bio-Rad assay) and manno-
syltransferase activity. This material (20-50 mg of
designated as the "crude" mannosyltransferase fraction.
The Triton X-100 solubilized membrane protein (10 ml) was di-
luted to 50 ml with ConA wash buffer (50 m M Tris-HC1,5% glycerol,
1 m M MnCl', 1 m M CaCl,, pH 7.5) and loaded onto a ConA-Sepharose
4B lectin column (5.0 ml of resin, 0.8 X 5.0-cm bed volume). Unbound
material was eluted by washing with 200 ml of wash buffer containing
0.2% Triton X-100, and glycosylated protein was eluted with 250 ml
of wash buffer containing 0.2% Triton X-100 and 0.75 M methyl-a-
mannoside. Fractions of 50 ml were collected. These and all subse-
quent chromatographic steps were performed in the cold room at
4 "C. During protein purification, the primary acceptors were man-
nose (11 mM), methyl-a-mannoside (65 mM), and al,6-mannobiose
(3.5 mM).
al,6-Mannobiose acceptor-dependent mannosyltransferase activ-
ity (M,MT-I) was eluted in the first three ConA wash fractions, while
methyl-a-mannoside acceptor-dependent mannosyltransferase activ-
ity (MIMT-I) was released from the ConA column with methyl-a-
mannoside. The most active ConA-bound and ConA flow-through
fractions were pooled and further purified independently in parallel
procedures starting with ion exchange chromatography on 3 ml of
DEAE-Sephacel resin (1.4- X 2.0-cm bed volume). After loading the
pooled active fractions, the DEAE-column was washed with 50 ml of
elution buffer (30 m M Tris-HCl, 5% glycerol, 1% Triton X-100, pH
analysis by the
ice-cold water. The
protein) was

Page 3

Yeast a1 + 2-Mannosyltramferases
8257
M
t'
M
M S M M S M L M M L M
1'
M
M L M
t'
M
a1,Z-Mannohiose al,6-Mannobiose al,6.Manaolriose nrM-Trisaccharide reM-Trisaccharide
FIG. 2. Abbreviated structures of
mannosyltransferase acceptors and
products. nrM, nonreducing mannose;
reM, reducing end mannose; dcM, disub-
stituted central mannose.
M
t'
M
MM
1'
M

MangGlcNAe M
= Mannose MnalOGieNAc
7.5), and the protein was eluted stepwise in a discontinuous KC1
gradient of elution buffer plus 50 mM, 100 mM, 150 mM, 200 mM, 300
mM, and 400 m M KCl. Aliquots of 50 pl were taken from each fraction
for assay. The active fractions from DEAE-chromatography were
applied independently to a Hg-agarose column (0.5 X 5 cm, 0.5 g of
resin), which was washed with 50 ml of KC1 buffer (50 m M Tris-HC1,
0.2 M KCl, 0.2% Triton X-100, pH 7.5) and eluted with 50 ml of wash
buffer containing 1% fl-mercaptoethanol.
Determination of Mannosyltransferase Substrate K,-The
rose-purified mannosyltransferase fractions were used for K,,, deter-
minations. The K,,, for GDP-mannose was determined by incubating
a fixed amount of 14C-radiolabeled GDP-mannose (0.1 pCi, 2.0 PM)
with variable amounts of unlabeled GDP-mannose (0-5 mM) in the
presence of 1 m M acceptor. After correcting for isotopic dilution, the
K,,, was estimated from a double-reciprocal plot of 1/V against l/(S).
To determine acceptor K, values, the GDP-mannose concentration
was held constant at 1 p ~ , and the carbohydrate concentration was
varied from 0 to 100 m M (mannose or methyl-a-mannoside) or 0 to
10 m M (a1,G-mannobiose). Again, the K,,, was estimated from the
double reciprocal plot of substrate concentration uersus reaction rate.
All assays were done in duplicate, and the experiment was repeated
if the values varied by more than 20%.
Determination of Mannosyltransferase Substrate Specificio-
M,MT-I and M,MT-I fractions, purified by ConA, DEAE, and Hg-
agarose chromatography, were assayed with each of the following
acceptors: mannose, methyl-a-mannoside, al,2-mannobiose, a1,6-
mannobiose, methyl-al,2-mannobioside,
methyl-al,4-mannobioside, methyl-al,6-mannobioside, al,2-man-
notriose, al,6-mannotriose, al,6-mannotetraose, nrM tetrasaccha-
ride, dcM tetrasaccharide, Man9-GlcNAc, and Manlo-GlcNAc (Fig.
2). All acceptors were present at 2 mM, except mannose and methyl-
ru-mannoside, which were at 12 mM.
HPLC Carbohydrate Analysis-The
brated at a flow rate of 1 ml/min by injecting 50-100 pg of appropriate
saccharide standards and monitoring the absorbance of the effluent
at 190 nm with a sensitivity of 0.1 unit full scale. Oligosaccharides
containing N-acetylglucosamine could be detected at a much lower
concentration (1-5 pg) due to the UV absorbance of the amide bond
at 200 nm. The reduced, unmodified, and methylated forms of mono-
and disaccharides were each eluted as distinct peaks using an 85:15
(v/v) acetonitrile/water solvent, with the methylated derivatives elut-
ing faster than the unmodified saccharides and the reduced forms
eluting slower. Mannose, mannobiose, mannotriose, and mannote-
traose separated well in an 80:20 (v/v) acetonitrile/water solvent.
The carbohydrate content of each peak was verified by the phenol-
sulfuric acid method (Dubois et al., 1956). Under controlled condi-
tions, the elution volumes varied by less than 10%. Radiolabeled
saccharides were collected in 0.5-ml fractions, diluted with 4.0 ml of
Scint AX fluid (Packard), and counted.
The DIONEX BioLC system was used to identify the tetrasaccha-
ride product formed from the al,6-mannotriose acceptor by purified
M,MT-I enzyme and GDP-[I4C]mannose. Both the 14C-labeled prod-
Hg-aga-
methyl-al,3-mannobioside,
HPLC columns were cali-
uct and the unlabeled oligosaccharide standards (Fig. 2) were reduced
with sodium borohydride (see below) and separated by isocratic
elution at 1 ml/min in 5 m M NaOH on a CarboPak PA-1 column.
Fractions of 0.1 ml were collected, neutralized with 0.4 ml of 0.1
NaHP04, and counted for radioactivity with 4.0 ml of Scint AX fluid.
Linkage Assignment of Mannosyltransferase Reaction Produc-
tions-Linkage analysis of the mannosyltransferase reaction products
formed by the ConA, DEAE, Hg-agarose-purified MIMT-I, and
M,MT-I enzymes was accomplished by a combination of partial
acetolysis (Kobayashi et al., 1986), al,2-mannosidase digestion (Ich-
ishima et al., 1981), and DIONEX HPLC co-elution of the I4C-
radiolabeled polysaccharides. Because the transferases were isolated
from S. cereuisiae mnnl microsomes, al,3-mannosyltransferase activ-
ity was expected to be very low or absent (Raschke et al., 1973), and
the scheme was designed to distinguish a1,2- from al,6-mannosyl-
transferase activities.
To generate enough material for product analysis, the standard
mannosyltransferase reaction was incubated for 90 min at 30 "C using
0.1-0.2 pCi of GDP-[14C]mannose to increase radiolabel incorporation
into the acceptor to at least 1 nCi of [14C]mannose. If the radiolabeled
mannose from nonspecific hydrolysis of GDP-[14C]mannose repre-
sented more than 20% of the counts in the Dowex 1 C1- effluent, the
reaction product was dried, extracted with chloroform/methanol/
water, and separated from free [14C]mannose by HPLC. The column
used depended on the difference in size between free mannose and
the transferase reaction product. When mono-, di-, or trisaccharide
acceptors were used, the products could be purified on the amino-
spherisorb column by isocratic elution at 1 ml/min with acetonitrile/
water (85:15, v/v to separate mono- from disaccharides, 8020, v/v to
separate mono- from trisaccharides, and 75:25, v/v to separate mono-
from tetrasaccharides). For larger acceptors, ['4C]mannose was sep-
arated from the oligosaccharide product by gel filtration HPLC (Bio-
Gel TSK-20 XL column) with distilled water as the solvent.
The transferase reaction product was divided into three equal parts.
Each fraction (about 0.8 ml) was dried by lyophilization, resuspended
in 100 pl of water, and extracted with chloroform/methanol/water by
the method of Wessel and Flugge (1984). After centrifugation to
separate the phases, the water phase was recovered and dried by
lyophilization. One part was subjected to acetolysis, another was
subjected to al,2-mannosidase digestion, and a third was left un-
treated. After treatment, the radiolabeled saccharides were analyzed
by HPLC, and their identities were assessed by co-elution with
standards.
For partial acetolysis (Kobayashi et al., 1986), the dry, extracted
carbohydrate product was resuspended by sonic agitation in 0.5 ml of
an acetic acid/acetic anhydride/sulfuric acid mixture (5050:1, v/v/
v), and the sealed tube was placed in a 40 "C water bath for 15 h. The
acetolysis reaction was stopped by adding 0.1 ml of dry pyridine to
neutralize the acid, the mixture was evaporated with a stream of Nl,
and the residue was resuspended in 1 ml of chloroform, which was
extracted four times with 1.5 ml of water. The upper water layer was
removed and discarded, and the chloroform phase was evaporated
M

Page 4

8258
Yeast a1 + 2-Mannosyltransferases
under a stream of nitrogen. The residue was dried under vacuum, and
the dry acetylated carbohydrate was deacetylated in 0.3 ml of dry
methanol with 30 pl of 1 M methanolic sodium methoxide for 1 h at
room temperature. The reaction mixture was diluted with 0.5 ml of
methanol and neutralized on a Dowex 50 Hf column (0.5 g), which
was washed 3 times with 0.75-ml portions
effluent was dried by evaporation and redissolved in water for HPLC
or scintillation counting.
For a-mannosidase digestion, the dried carbohydrate was resus-
pended in 100 pl of 0.1 M NaOAc buffer (pH 5.0) with 10 milliunits
of al,2-mannosidase, and the sample was incubated at 32 "C for 24
h. After terminating the reaction by boiling the mixture for 2 min,
the solution was injected directly into the HPLC or counted in the
scintillation counter. The control fraction was dried, extracted with
chloroform/methanol/water, and dried again by lyophilization. The
residue was resuspended in 100 pl of water, and one-half of the
material was analyzed by HPLC, while one-half was used for scintil-
lation counting.
Reduction of Mannosyltransferase Reaction Products-The Dowex
1 C1- effluent (neutral mannosyltransferase product) was dried, ex-
tracted with chloroform/methanol/water, dried again, and resus-
pended in 100 pl of 0.1 M NaHC03 buffer, pH 9.0. Sodium borohydride
(NaBH4) was added to 1 M concentration, and the solution was
incubated 24 h at room temperature. The reaction was terminated by
adding 0.4 ml of water and enough Dowex 50 H+ resin (200-400
mesh) to bring the pH below 5.0. The liquid was filtered from the
resin, dried by lyophilization, and free
methyl borate. The sample was dissolved in 1 ml of methanol, heated
to boiling on a steam bath, then cooled in an ice bath and evaporated
in a Rotovap apparatus. This procedure was repeated three times
until the white precipitate was removed.
Comparison of Mannosyltransferase Activities from mnn Mutant
Extracts-S. cerevisiae mnnl, mnnlmnn2, mnnlmnn.3, mnnlmnn5,
and mnnlmnn9 strains were grown for 24 h at 30 "C in 100 ml of
YPD medium (or, for mnn9 strains, YPD plus 1 M sorbitol), and then
inoculated into 1 liter of the same medium and grown to midlog phase
(6.0 AGW units). Detergent-solubilized microsomal proteins were pre-
pared from the cells as described above, separated from insoluble
material by ultracentrifugation, and assayed for mannosyltransferase
activity with 12 mM methyl-a-mannoside, 5.0 mM al,6-mannobiose,
and 3.5 mM al,2-mannobiose. Each tube contained 90 pg of protein
and 1 p~ GDP-['4C]mannose, and the assays were incubated for 45
min at 30 "C. Control tubes contained radiolabeled GDP-mannose
and membrane protein but no acceptor carbohydrate.
of methanol. The combined
boric acid was removed as
RESULTS
Separation and Partial Purification
and M2MT-I Mannosyltransferase Actiuities-Over
of the methyl-a-mannoside-dependent mannosyltransferase
(MIMT-I) activity, which was solubilized from S. cereuisiue
microsomes, bound to ConA-Sepharose. The ConA flow-
through contained less than 10% of the applied MIMT-I
activity, but this fraction also contained an activity that
catalyzed the transfer of [14C]mannose from GDP-[14C]man-
nose to al,G-mannobiose, which was designated M2MT-I. The
two activities were partially purified separately on ion ex-
change and Hg-agarose columns. MIMT-I activity was eluted
from DEAE-Sephacel at a lower salt concentration (65% in
the 100 mM KC1 fraction) than M2MT-I activity (75% in the
200 mM KCl), which allowed the removal of cross-contami-
nating mannosyltransferase activity from the
Both transferase activities bound to
could be eluted with 1% P-mercaptoethanol but not with 0.2
M KCl. After the Hg-agarose step, the specific activity of
MIMT-I was increased between 100- and 200-fold, while the
specific activity of M2MT-I was increased about
(Table I). There was a substantial loss of M2MT-I activity
during the purification procedure (70-80%), which was not
due to adsorption on the ConA column because
activity was not detectable in any DEAE-Sephacel column
fraction in the MIMT-I preparation. The low yields of both
transferases during purification apparently were due to de-
naturation on contact with the resins, which may have been
of the MIMT-I
90%
ConA step.
Hg-agarose resin and
10-fold
M2MT-I
the result of nonspecific hydrophobic interactions. This hy-
pothesis is supported by the observation that
transferase activities was extremely low from hydrophobic
resins such as ethyl-agarose.
The MIMT-I and M2MT-I activities
pletely (>go%) solubilized from S. cerevisiae mnnl microso-
mal pellets by Triton X-100. In contrast, little activity
either type was released by high salt (0.5 M KC1) or 2 M urea,
which indicates that these activities
membrane proteins. Silver-stained
electrophoresis of the Hg-agarose-purified enzymes revealed
several protein bands in each preparation (data not shown).
Since no activity could be recovered from SDS gels, we were
unable to identify the apparent
MIMT-I and M2MT-I Have Different Substrate Specifici-
ties-Hg-agarose-purified MIMT-I and M2MT-I preparations
were tested for activity with various acceptors (Table
Notable is the absence of overlap in the substrate profiles, a
good indication that the two preparations contain different
enzymes and not modified versions of the same enzyme.
MIMT-I utilizes mannose (K, = 11.5 mM), methyl-a-man-
noside, and al,2-mannobiose as primary acceptors. Con-
versely, M2MT-I utilizes al,6-mannobiose (K, = 3.5 mM),
mannotriose, and mannotetraose, as well as the dcM tetra-
saccharide and Man9-GlcNAc (see
M2MT-I have similar affinity for the GDP-mannose donor
(K, = 25 PM and 33 PM, respectively). Partially purified
M2MT-I does not utilize any monosaccharide or al,2-linked
mannose polymer as a substrate, whereas partially purified
MIMT-I does not utilize any al,6-linked mannose polymer or
core oligosaccharide as a substrate. The lack of substrate
overlap suggests that each is responsible for the biosynthesis
of a different part of the glycoprotein outer chain.
MIMT-I and M2MT-I Are al,2-Mannosyltransfermes-
Previous investigators have found that the mannose/methyl-
a-mannoside-specific mannosyltransferase in crude yeast ex-
tracts catalyzes the formation of a1,2-linkages (Lehle and
Tanner, 1974; Parodi, 1979) and, as expected, MIMT-I is an
al,2-mannosyltransferase. DIONEX HPLC showed that the
MIMT-I mannobiose transferase product was co-eluted with
an a1,Z-mannobiose standard, which was well separated from
a1,3- and al,6-mannobiose standards (data not
Mannosidase is highly specific for a1,2 bonds, and 10 milli-
units of this enzyme does not appreciably hydrolyze a1,3- or
a1,6-mannobioses in a 24-h incubation at 32 "C. al,P-Man-
nosidase digestion released more than 95% of the [I4C]man-
nose from the radiolabeled MIMT-I disaccharide transferase
products (mannobiose or methyl-a-mannobioside); and par-
tial acetolysis, which specifically cleaves al,6-glycosidic
bonds, did not change the HPLC elution
radiolabeled product (Fig. 3). These data provide convincing
proof that only a1,2 bonds are formed
MIMT-I preparation, although it is possible that this fraction
contains more than one c~1,2-mannosyltransferase.
A more unexpected result was the identification of MzMT-
I as an al,2-mannosyltransferase. To confirm this conclusion,
the reaction products formed with 3 different acceptor mole-
cules, al,6-mannobiose, al,6-mannotriose, and Man9-
GlcNAc, were analyzed. In each example, al,2-mannosidase
digestion released more than 95% of the [14C]mannose from
the radiolabeled mannotetraose, mannopentaose, and Manlo-
GlcNAc reaction products (Figs. 4 and 5). In addition, partial
acetolysis quantitatively released a radiolabeled disaccharide
from each of these products (Figs. 4 and 5). These data also
show that the partially purified M,MT-I preparation does not
have significant al,6-mannosyltransferase activity. If there
recovery of both
were almost com-
of
derive from
SDS-polyacrylamide gel
integral
M, of the transferase(s).
11).
Fig. 2). MIMT-I and
shown). a1,2-
position of the
by the partially purified

Page 5

Yeast al + 2-Mannosyltransferases
8259
Transferase
TABLE I
from Purification of mannosyltransferases S. cerevisiae microsomes
step
Total protein
mg
40
2
0.6
0.2
40
38
3
Volume Total activity Specific
pmollhlmg
2,800
50,500
140,000
322,000
activity
Purification Yield
M,MT-I
MnMT-I
Triton
ConA-Sepharose
DEAE-Sephacel
Hg-agarose
Triton
ConA-Sepharose
DEAE-Sephacel
Hg-agarose
X-100 extract”
X-100 extract*
ml pmol/h
112,000
101,000
84,000
64,000
6,000
4,300
2,900
1,800
-fold %
8
1
100
90
75
58
100
72
49
30
50
50
20
18
50
115
8 150
110
970
1
0.7
6.5
8
50
50
20 1.5 1,200
a Assayed
b Assayed
with
with
65 mM methyl-n-mannoside
2.5 mM al,6-mannobiose
as the acceptor.
as the acceptor.
TABLE
II
AcceDtor sDecificitv of Hg-agarose-ourified M,MT-I and MZMT-Z
Acceptor (concentration)
Mannose
Methyl-n-mannoside
nl,2-Mannobiose
cul,G-Mannobiose
cul,G-Mannotriose
oll,6-Mannotetraose
nrM-Tetrasaccharide
dcM-Tetrasaccharide
Man,GlcNAc
Man,,,GlcNAc
(20 mM)
(20 mM)
(2 mM)
(2 mM)
(2 mM)
(2 mM)
(2 mM)
(2 mM)
(2 mM)
(2 mM)
M,MT-I
activity
M,MT-I
activity
pmol mannose.
200
1320
560
0
0
0
-0
-a
0
0
h-’
ml-’
20
10
0
270
110
190
10
710
250
10
” Not determined.
Fraction Number
Mannabiose
Fraction Number
FIG.
3. Aminospherisorb
product
HPLC elution
as the acceptor.
profiles of M,MT-I
+, untreated
product;
transferase
[“Clmannobiose
l , acetolyzed
(85:15,
positions
with mannose
A, cul,2-mannosidase-digested
column was eluted
fractions. The
and crl,2-mannobiose
product; and
product.
in
The with acetonitrile/water
indicate
standards.
v/v) 0.5-ml arrotus the elution
of mannose
had been substitution
[‘C]mannose
tolysate.
The
[“‘C]ManlOGlcNAc
aI,2-addition
linked mannose
cause al,2-linked
would have been released
[Wlmannotriose
detected.
Analysis
of M,MT-I
tolysis and oc-mannosidase
M,MT-I is an crl,2-mannosyltransferase,
at position
have been present
release of radiolabeled
product
must have been on the one unsubstituted
residue in the Man,-GlcNAc
[i4C]mannose added at any other
upon
or as [‘4C]Man5-GlcNAc,
6 of the acceptor molecules,
would in the partial
disaccharide
furthermore,
ace-
from the
that
a1,6-
revealed, the
acceptor, be-
position
either
were not
partial acetolysis
which
as
Transferase
digestion
Products-Although
demonstrated
ace-
that the
techniques these
0
FIG. 4. Aminospherisorb
product
[“Clmannotriose
and +, acetolyzed
HPLC
nl,6-mannobiose
product;
product.
elution profiles of MzMT-I
transferase
treated
product;
Fig. 3.
with acceptor. +, un-
A, cul,2-mannosidase-digested
elution was performed The as in
200
Mannose
,
Mannobiose
4f
t
J-
O
10 20 30 40
Fraction
Number
i
4
5
i
0
FIG. 5. Aminospherisorb
product
[“C]ManloGlcNAc
and +, acetolyzed
but the solvent
HPLC
Man*GlcNAc
product;
product.
at A to acetonitrile/water
elution profiles
as
of M2MT-I
transferase
treated
product;
with acceptor. +,
un-
A, al,2-mannosidase-digested
elution Initial was as in Fig.
(50:50,
3,
was changed v/v).
alone could not identify
mannooligosaccharide
2. For the Lul,G-mannobiose
relatively easy because the molecule
sible acceptor sites.
DEAE-, Hg-agarose-purified
to partial acetolysis,
which
acceptors
mannose
was substituted
acceptor,
contained
[‘4C]Mannotriose
M2MT-I
and compared
residue in the (u1,6-
at position
this assignment was
only two pos-
by ConA-,
subjected
formed
was reduced,
by chromatography with