Identification and Characterization of High Molecular-Mass Mucin-Like
Glycoproteins in the Plasma Membrane of Airway Epithelial Cells
Emmanuel Paul, Dong Ik Lee, Sang Won Hyun, Sandra Gendler, and K. Chul Kim
Department of Pharmaceutical Sciences, University of Maryland School of Pharmacy, Baltimore, Maryland;
and Mayo Clinic Scottsdale, Scottsdale, Arizona
A previous lectin binding study demonstrated the presence of high molecular-mass mucin-like glycoproteins (HMGP) on the surface of hamster tracheal surface epithelial (HTSE) secretory cells (Proc. Natl.
Acad. Sci. USA 1987;84:9304). In the present study, we intended to isolate and characterize these HMGP
from the plasma membrane of the primary HTSE cells and then to determine whether or not these membrane HMGP are Muc-1 mucins, a type of mucins originally discovered on the surface of some carcinomas. A subcellular fraction enriched with the plasma membrane was obtained using a sucrose density gradient centrifugation. This fraction contained high molecular-mass glycoconjugates which were excluded
from Sepharose CL-4B gel. Biochemical characterization of these glycoconjugates revealed the following
characteristics: (1) susceptibility to both pronase and mild alkaline treatments, but totally resistant to proteoglycan-digesting enzymes; (2) partitioning in the detergent phase of Triton X-114 and resistance to digestion by phosphatidylinositol phospholipase C or D; (3) a buoyant density of 1.5 g/ml based on CsCl
density gradient centrifugation; (4) polydispersity in terms of both size and charge density; and (5) lack of
immunoreactivity with an anti-Muc-1 mucin antibody. We conclude that the plasma membrane of HTSE
cells at confluence contains HMGP, which seem to be the integral membrane proteins but different from
Muc-1 mucins, and that these membrane HMGP appear to share some similarities with secreted mucins in
terms of size and charge. Paul, E., D. I. Lee, S. W. Hyun, S. Gendler, and K. C. Kim. 1998. Identification and characterization of high molecular-mass mucin-like glycoproteins in the plasma membrane
of airway epithelial cells. Am. J. Respir. Cell Mol. Biol. 19:681–690.
Airway mucus plays an important role in host defense
against airborne particles through its viscoelastic property.
This physicochemical property of the mucus, which is crucial for its proper mucociliary function, is contributed
mainly by mucus glycoproteins or mucins present in the
mucus. Abnormalities in either the quality or quantity of
these mucins therefore could lead to the pathology of the
airway. In the airway, mucins are secreted by two types of
cells: goblet cells on the surface epithelium and mucous
cells in the submucosal gland. Although a great deal of in-
(Received in original form January 27, 1997 and in revised form February
10, 1998)
Address correspondence to: K. Chul Kim, Ph.D., Dept. of Pharmaceutical
Sciences, University of Maryland School of Pharmacy, Baltimore, MD
21201.
Abbreviations: an anti-Muc-1 antiserum, CT-1; diethylaminoethyl, DEAE;
glycosylphosphatidylinositol-phospholipase D, GPI-PLD; N-2-hydroxyethylpiperazine-N9-ethane sulfonic acid, Hepes; high molecular-mass mucin-like glycoproteins, HMGP; hamster tracheal surface epithelial cells,
HTSE; polyacrylamide gel electrophoresis, PAGE; phosphate-buffered
saline, PBS; phosphatidylinositol-phospholipase C, PI-PLC; sodium dodecyl sulfate, SDS; void volume, Vo.
Am. J. Respir. Cell Mol. Biol. Vol. 19, pp. 681–690, 1998
Internet address: www.atsjournals.org
formation is now available regarding the regulation of mucin secretion by these cell types (1, 2), very little is known
about either the structure or the subcellular origin of these
mucins.
The presence of high molecular-mass mucin-like glycoproteins (HMGP) was demonstrated on the apical surface
of hamster tracheal surface epithelial (HTSE) cells in primary culture based on lectin binding studies (3, 4). These
cell-surface mucins had a slow turnover rate compared
with that of the mucins secreted spontaneously, and were
releasable by human neutrophil elastase by a proteolytic
cleavage (3). These cells have also been shown to express
messenger RNAs of Muc-1 (5, 6)—the cell-surface mucins
that were originally discovered in some carcinomas (7). In
the present study, we intended to isolate and characterize
these HMGP from the plasma membrane of the primary
HTSE cells and then to determine whether or not these
membrane HMGP are Muc-1 mucins. Here we report that
the plasma membrane of primary HTSE cells contain
HMGP that seem to be integral glycoproteins but different from the cell-surface Muc-1 mucins, and that both the
size and charge of these HMGP appear to be similar to
those of secreted mucins. Thus the present results indicate
that there are at least two different types of mucins on the
surface of HTSE cells.
682
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 19 1998
Materials and Methods
Materials
3
H-Hyaluronic acid was purified from spent media of primary rabbit tracheal epithelial cells grown on plastic dishes,
which had been metabolically radiolabeled with [6-3H]glucosamine as described previously (8). All other materials
were obtained from Sigma Chemical Company, Inc. (St.
Louis, MO) unless stated otherwise.
Cell Culture and Metabolic Radiolabeling
Tracheas were obtained from male Golden Syrian hamsters of 6–7 wk of age (Harlan Sprague-Dawley, Indianapolis, IN). Primary HTSE cells were prepared and cultured
as previously described (4). Briefly, cells were cultured on
a thick collagen gel and maintained in a medium supplemented with various factors and 5% fetal bovine serum
(4). The culture medium was changed on Days 1, 3, and 5
following the plating. Metabolic radiolabeling of cells was
performed by incubating confluent cultures in 100-mm
dishes for 24 h with the medium containing one or a combination of the following materials: [6-3H]glucosamine (10
mCi/ml), [1-14C]glucosamine (5 mCi/ml), and [3H]palmitic
acid (40 mCi/ml) (New England Nuclear, Boston, MA).
Following incubation, the spent media were collected and
the floating cells were removed by centrifugation at 200 3 g
at 48C for 5 min.
fonic acid (Hepes), pH 7.5, transferred to polyallomer
tubes, and centrifuged at 100,000 3 g for 1 h at 48C in an
ultracentrifuge (Beckman LS-80; Beckman Instruments,
Inc., Palo Alto, CA) to precipitate a second membrane
fraction (P3). The resulting supernatant (S3) was discarded.
Resolution of P2 and P3 fractions into purified membrane
fractions was achieved by the following steps. Both fractions were resuspended separately in 3 ml of 50 mM Hepes,
pH 7.5, and loaded over a 2-ml sucrose cushion (48% wt/
vol) in Ultra Clear tubes (Beckman Instruments) fitted onto
an SW 55Ti rotor (Beckman Instruments) and centrifuged
at 100,000 3 g for 1 h at 4 8C. Following centrifugation,
each tube displayed four regions: a soluble fraction at the
upper end of the tube (fraction I), a membrane band floating above the sucrose cushion (fraction II), the sucrose
cushion itself (fraction III), and a hard membrane pellet
(fraction IV). Fractions II, III, and IV were resuspended
separately with 50 mM Hepes, pH 7.5, to make 38.5 ml,
and centrifuged at 100,000 3 g to remove sucrose. Each
pellet fraction was diluted with 50 mM Hepes, pH 7.5, and
used immediately for marker enzyme assays.
Identification of the Plasma Membrane Fraction
(Marker Enzyme Assay)
Both the identity and purity of the plasma membrane fraction were monitored by analysis of various marker en-
Dissociation and Homogenization of HTSE Cells
Radiolabeled cells in 100-mm dishes were rinsed with
phosphate-buffered saline (PBS), pH 7.4, and then incubated in 5 ml of Medium 199, pH 7.2, containing 4,000 U
of collagenase (type IV; Worthington Biochemical Corp.,
Freehold, NJ) for 4 h inside a 378C CO2 incubator to dissociate the cells from the gel. The suspension was pelleted by
centrifugation at 200 3 g for 10 min at 48C. The pellet was
washed twice by centrifugation with 5 ml of a homogenization buffer before resuspending in 1 ml of the same buffer. The homogenization buffer contained 0.28 M sucrose,
10 mM MgCl2, 150 mM NaCl, 50 mM 2-(N-Morpholino)ethanesulfonic acid, and 1 mM ethyleneglycol-bis-(b-aminoethyl ether)-N,N9-tetraacetic acid, pH 6.0. Immediately
before homogenization, 0.1 mM phenylmethylsulfonyl fluoride (PMSF) and 0.1 mg/ml soybean trypsin inhibitor
were added to the cell suspension. Homogenization was
performed at 48C inside a 50-ml polypropylene tube by
sonication twice for 15 s at power setting 6 using a microultrasonic cell disruptor (Model W-10; Mansonics, Plainview, NY).
Subcellular Fractionation
Subcellular fractionation was carried out with a slight
modification of the protocol described by Langridge and
colleagues (9) (Figure 1). Briefly, the crude homogenate
diluted to 10 ml in the homogenization buffer was transferred to a 50-ml propylene tube and centrifuged at 700 3 g
for 15 min at 48C. The pellet (P1) was discarded and the resulting supernatant (S1) was then centrifuged at 10,000 3 g
for 30 min at 48C to sediment a crude membrane fraction
(P2). The resulting supernatant (S2) was diluted to 38.5 ml
with 50 mM N-2-hydroxyethylpiperazine-N9-ethane sul-
Figure 1. Flow chart of subcellular fractionation of hamster tracheal epithelial cells. Confluent HTSE cells were homogenized
and subjected to differential centrifugation as described in M ATERIALS AND METHODS. After the sucrose density gradient, four
fractions were collected: P2-I from the top supernatant, P2-II
from the interface between the top supernatant phase and the sucrose phase, P2-III from the sucrose phase, and P2-IV from the
pellet. These fractions were subjected to the marker enzyme assay as seen in Figure 2.
Paul, Lee, Hyun, et al.: Mucins on the Surface of Airway Epithelial Cells
zymes in the fraction. The marker enzymes used for different subcellular fractions and the analytical assays used for
the enzymes were: alkaline phosphatase (10) and 59-nucleotidase (11) for the plasma membrane, acid phosphatase
(12) for the lysosomal membrane, thiamine pyrophosphatase (13) for the Golgi, glucose-6-phosphatase (14) for
the rough endoplasmic reticulum, and succinate dehydrogenase (15) for mitochondria. Protein was quantitated by
the method of Bradford (16) with bovine serum albumin
as standard.
Enzymatic Digestions
All of the digestions, except with phosphatidylinositolphospholipase C (PI-PLC) and glycosylphosphatidylinositol-phospholipase D (GPI-PLD) were carried out as previously described (17). Briefly, the enzyme digestions were
performed as follows: hyaluronidase (Sigma type VI-S,
200 U/ml) at 378C for 16 h in PBS adjusted to pH 5.5 by
the addition of 0.1 M citric acid; chondroitinase ABC (0.4
U/ml) at 378C for 5 h in 0.1 M Tris acetate, pH 7.3; heparitinase (5 U/ml) at 428C for 5 h in 0.1 M sodium acetate,
1 mM CaCl2, pH 7.0; and pronase (Type XIV protease,
5 mg/ml; Sigma) at 378C for 16 h in PBS, pH 7.2. At the
end of each enzyme digestion, the reaction mixture was
boiled for 2 min to denature the enzyme and then centrifuged at 12,000 3 g for 2 min, and the supernatant was applied to a Sepharose CL-4B column.
Phase Transition Using Triton X-114 and
Digestion with PI-PLC and GPI-PLC
Triton X-114 was precondensed as described by Bordier
(18). The purified plasma membrane fraction was lyophilized and then mixed with a 1.5% Triton X-114 solution
containing 10 mM Tris-HCl, pH 7.5, and 0.5 M NaCl. The
mixture was equilibrated in ice for 30 min to solubilize the
membrane lipids, then incubated at 378C for 10 min to precipitate the detergent. The two phases were separated by
centrifugation at 2,000 3 g for 10 min at 378C. The precipitate was extracted twice with 100 ml of 0.006% Triton
X-114 and the supernatant was pooled with the original supernatant (the aqueous phase) and lyophilized. The “pellet” (the detergent phase) was solubilized with 1 ml cold
absolute acetone (2808C) to dissolve the detergent while
precipitating the proteins. Both the lyophilized aqueous
phase and the acetone pellet were dissolved in 50 mM sodium acetate, pH 7.0, containing 0.1% sodium dodecyl
sulfate (SDS), and boiled for 2 min before applying to
Sepharose CL-4B columns. Digestions with PI-PLC and
GPI-PLD were performed as previously reported (19–22).
Briefly, the detergent “pellet” obtained from the Triton
X-114 phase separation ( see below) was dissolved with
200 ml of the above Triton X-114 buffer, and kept at 158C
for 5 h in the presence of either 15 mU/ml of PI-PLC (Bacillus cereus; Sigma) or 50 mU/ml of GPI-PLD (Streptomyces species, Type VII; Sigma). In the case of GPI-PLD,
CaCl2 was added to 5 mM and the reaction mixture was
adjusted to pH 7.0 with Tris. After the reaction, the phases
were separated and subjected to Sepharose CL-4B column
chromatography as described previously.
683
Chemical Digestions
Reductive b-elimination was performed as previously described (17). Briefly, the sample was incubated in 50 mM
NaOH/1 M NaBH4 at 378C for 16 h, then neutralized with
glacial acetic acid at the end of the reaction, and the borate
was removed by repeated evaporation with methanol containing 1% acetic acid. The sample was then lyophilized
and redissolved in the elution buffer before being subjected to Sepharose CL-4B chromatography.
Chromatography
Column chromatography procedures used in this experiment have been described previously (17) and gel-filtration columns were eluted with a buffer containing 50 mM
sodium acetate, pH 7.2, and 0.1% SDS unless otherwise
specified. Briefly, for diethylaminoethyl (DEAE)-Sephacel
anion exchange chromatography, the 3H-void volume (Vo)
sample was first subjected to a buffer exchange by passing
through a Sephadex G-50 column that had been equilibrated with a buffer containing 8 M urea, 0.15 M NaCl,
0.5% (vol/vol) Triton X-100, and 0.05 M sodium acetate,
pH 6.0. The Vo fractions were pooled and then applied to a
DEAE-Sephacel column (1.0 3 2.5 cm) that had been
equilibrated with the same buffer. After washing with 5 ml
of this buffer, elution was continued by applying a gradient
of NaCl from 0.15 to 1.4 M over 46 ml. Fractions of 0.8 ml
were collected at a flow rate of 3.0 ml/h. The recoveries of
radioactivity after column chromatography were greater
than 80% in gel-filtration columns, and about 90% in the
anion-exchange column.
CsCl Equilibrium Density Gradient Centrifugation
CsCl density gradient centrifugation was performed as previously described (23). Briefly, 100 ml of each sample was
added to a CsCl solution prepared in PBS, pH 7.2, with a final density of 1.5 g/ml that contained 4 M guanidine HCl and
0.2% SDS. The CsCl sample solution was boiled for 5 min,
transferred into polyallomer tubes, loaded into a SW 55Ti
rotor, and centrifuged at 105,000 3 g for 48 h at 108C.
Immunoprecipitation with an Anti-Muc-1
Mucin Antibody
The immunoprecipitation was carried out with the polyclonal antibody CT-1 as previously described (24). Briefly,
confluent HTSE cells grown on 100-mm dishes were first
starved for 2 to 4 h in a glucose-free Dulbecco’s modified
Eagle’s medium, and then metabolically radiolabeled for
24 h by incubating the cells in the complete medium (2 ml/
dish) which contained a “low” level of glucose (0.4 g/liter)
and [6-3H]glucosamine (100 mCi/ml). At the end of the labeling, cells were washed with PBS and lysed using a
buffer containing 40 mM sodium phosphate (pH 7.2), 250
mM NaCl, 50 mM NaF, 5 mM ethylenediaminetetraacetic
acid, 1% Triton X-100, 1% deoxycholate, 10 mM benzamidine, 25 mg/ml leupeptin, 110 mg/ml PMSF, and 10 mg/ml
aprotinin. The cell lysate was incubated with rabbit preimmune sera for 1 h at 48C to minimize the nonspecific binding to the antisera. At the end of the incubation, 1/10 volume of protein A-agarose beads (10% in the lysing buffer,
vol/vol) was added and mixed for 1 h on a rotating wheel
684
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 19 1998
at 48C, followed by centrifugation to remove any nonspecific binding to the beads. The same procedures were applied to this “precleaned” lysate, namely, incubation first
with an anti-Muc-1 antiserum (CT-1) in the presence or
absence of a 17-mer blocking peptide (0.1 mg/ml) consisting of the C-terminal region of human Muc-1 mucin (24),
and then with the protein A-agarose beads. The immune
complex was precipitated by centrifugation at 10,000 3 g
for 2 min at 48C. The precipitate was washed 4 times, each
time using a different buffer: (1) the lysis buffer in 0.25 M
NaCl, (2) the lysis buffer in 0.5 M NaCl, (3) the lysis buffer
in 0.25 M NaCl, and (4) double-distilled water. The extensively washed precipitate was then mixed with 23 Laemmli’s sample buffer and boiled for 2 min, and its total volume was adjusted for gel-filtration using the elution buffer.
In the case of the membrane 3H-HMGP, the 3H-radioactivity of the final precipitate was too low (, 50 cpm) to run
the CL-4B column and therefore the supernatant samples
were applied to the column.
SDS-Polyacrylamide Gel Electrophoresis
SDS-polyacrylamide gel electrophoresis (PAGE) was carried out at 100 V through a 3% stacking gel and a 5% separating gel using a Minigel apparatus (Bio-Rad Laboratories, Hercules, CA). At the end of the electrophoresis, the
gel was dried and exposed to a sheet of X-ray film (Eastman Kodak, Rochester, NY) with intensifying screens for
10 d.
Figure 2. Identification of subcellular fractions. The four different fractions obtained after the sucrose density gradient centrifugation as described in Figure 1 were analyzed for the presence of
the following enzymes as described in M ATERIALS AND METHODS: SDH (succinic dehydrogenase) as a marker for mitochondria; ALP (alkaline phosphatase) and 5-Nu (59-nucleotidase) for
the plasma membrane; ACP (acid phosphatase) for the lysosome;
TPP (thiamine pyrophosphatase) for the Golgi; and G6P (glucose-6-phosphatase) for the rough endoplasmic reticulum. Fractions 2, 3, and 4 represent fractions P2-II, P2-III, and P2-IV, respectively. Fraction P2-I was free of any of these enzyme activities,
whereas fraction P2-II was enriched with the plasma membrane
marker enzymes and will be referred to as the plasma membrane
preparation or fraction throughout the text.
Results
Subcellular Fractionation of HTSE Cells
Sequences involved in subcellular fractionation and activity profiles of various enzyme markers in individual fractions are indicated in Figures 1 and 2, respectively. The
plasma membrane markers alkaline phosphatase and 59nucleotidase were recovered mostly in fraction P2-II with
recoveries of 40 and 35% from the original crude homogenate, respectively. This fraction contained relatively low
activities of other marker enzymes: 0.9% in succinate dehydrogenase (mitochondria), 5.5% in acid phosphatase
(lysosome), 0.75% in thiamine pyrophosphatase (Golgi),
and 0% in glucose-6-phosphatase (rough endoplasmic reticulum). Therefore, the enzyme profile of the P2-II fraction
indicated that this fraction was highly enriched with the
plasma membrane. In contrast, the P3-II fraction was enriched with other microsomal enzymes (51.3% in acid
phosphatase, 53% in thiamine pyrophosphatase, and 45%
in glucose-6-phosphatase, calculated from the total activity
in the original homogenate) but had relatively low activities in the plasma membrane markers (16% in 59-nucleotidase and 5.5% alkaline phosphatase) compared with the
P2-II fraction. Fractions P2-I and P3-I were free of any of
these enzymes. On the basis of the total protein concentration, the P2-II fraction had a 61-fold enrichment in 59-nucleotidase activity and a 55-fold increase in alkaline phosphatase
activity over the original crude homogenate. On the other
hand, 3H-radioactivity in the P2-II fraction constituted 0.7%
of the total [3H]glucosamine-labeled glycoconjugates and
4.0% of the total crude homogenate.
Figure 3. Sepharose CL-4B gel-filtration column chromatography of the plasma membrane fraction. Confluent HTSE cells
were metabolically radiolabeled with [ 3H]glucosamine for 24 h,
and the membrane fraction (P2-II) was prepared from their homogenates as seen in Figures 1 and 2. The membrane fraction
was then subjected to Sepharose CL-4B gel-filtration column
chromatography eluting with 50 mM sodium acetate, pH 7.0, containing 0.1% SDS. Vo and Vt represent void volume and total volume, respectively. Three Vo fractions (#22–24) were pooled (Vo
sample) and subjected to further characterization.
Paul, Lee, Hyun, et al.: Mucins on the Surface of Airway Epithelial Cells
Purification of HMGP Synthesized by HTSE Cells
Both spent (conditioned) media and P2-II fractions that
were obtained from [3H]glucosamine-labeled HTSE cells
were subjected to Sepharose CL-4B column chromatography. The resulting 3H-Vo fractions (Figure 3) were pooled
and examined for the presence of proteoglycans by enzyme digestion. The 3H-Vo from the plasma membrane
fraction was totally resistant to chondroitinase ABC (Figure 4A), hyaluronidase (data not shown), or heparitinase
(data not shown), indicating the absence of hyaluronic
acid, dermatan sulfate, chondroitin sulfate, or heparan sulfate proteoglycans. However, the 3H-Vo was completely
digested with both pronase and alkaline borohydride
treatment, with the resulting digests included in the column with Kd of 0.59 (Figure 4B) and 0.74 (Figure 4C), respectively. (Kd is the distribution coefficient defined as the
fraction of the stationary phase which is available for diffusion of a given solute species, and calculated here using
the equation Kd 5 [Ve 2 Vo]/[Vt 2 Vo], where Ve, Vo, and
Vt represent elution volume, void volume, and total volume, respectively.) The results indicate that the 3H-Vo glycoconjugates are O-linked glycoproteins. The Vo from P2II will henceforth be referred to as HMGP of the plasma
membrane, or simply the membrane HMGP. Similarly,
the Vo obtained following digestion of the spent media
with hyaluronidase will be defined as HMGP of the spent
media, or the secreted HMGP.
685
tergent phase and virtually no 3H-activity detected in the
aqueous phase, suggesting that 3H-HMGP are not likely
anchored to the plasma membrane via the GPI anchor. A
similar result was obtained after treatment with 50 mU/ml
Extraction of HMGP with Triton X-114
Both the P2-II fraction and the spent media were extracted with Triton X-114, and the resulting aqueous and
detergent phases of Triton X-114 from each sample were
chromatographed separately on Sepharose CL-4B columns. Extraction of the plasma membrane fraction resulted in a distribution of 13 and 87% of the 3H-activity in
the aqueous and detergent phases, respectively. When
samples recovered from the detergent and aqueous phases
were chromatographed on Sepharose CL-4B columns, 72
and 28% of 3H-Vo were contributed by the detergent and
aqueous phases, respectively (Figure 5A). On the other
hand, after extraction of the spent media, 98 and 2% of the
3
H-activity were found in the aqueous and the detergent
phases, respectively. Sepharose CL-4B column chromatography of these samples revealed that 90% of 3H-Vo in the
spent media was from the aqueous phase and the remaining 10% from the detergent phase (Figure 5B). Thus the
result indicates that most of the HMGP in the plasma
membrane are integral membrane glycoproteins.
Digestion of the Plasma Membrane with
PI-PLC and GPI-PLD
To see if the membrane HMGP are anchored to the
plasma membrane via a GPI linkage, the detergent phase
obtained following the Triton X-114 phase separation was
treated with either PI-PLC or GPI-PLD, then subjected
again to the Triton X-114 phase partitioning. If digested,
HMGP are expected to be released into the aqueous
phase because of their loss of the diacylglycerol moiety.
Figure 6 shows that all 3H-radioactivity is found in the de-
Figure 4. Digestion of Vo fractions from the plasma membrane
with (A) chondroitinase ABC, (B) pronase, and (C) alkaline
borohydride. The Vo sample obtained from the plasma membrane preparation (Figure 3) was treated with either chondroitinase ABC, pronase, or alkaline borohydride as described in MATERIALS AND METHODS. At the end of the reaction, the reaction
mixture was applied to a Sepharose CL-4B column under the
conditions described in Figure 3. Open circles: before treatment;
closed circles: after treatment. Vo and Vt represent void volume
and total volume, respectively.
686
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 19 1998
Figure 6. Treatment of the plasma membrane fraction with PIPLC. Confluent HTSE cells were metabolically radiolabeled with
[3H]glucosamine for 24 h, and the P2-II fraction was purified as
described in Figure 1. The plasma membrane was then subjected
to phase separation using Triton X-114 as described in Figure 7.
The pellet recovered from the detergent phase was dissolved in
the same Triton X-114 solution and incubated at 158C for 5 h in
the presence of 15 mU/ml of PI-PLC as described in MATERIALS
AND METHODS. At the end of incubation the solution was phaseseparated at 378C, and individual phases were adjusted to 0.1%
SDS, heat-boiled, and finally applied to a Sepharose CL-4B column pre-equilibrated with 50 mM sodium acetate containing
0.1% SDS. Open circles: aqueous phase; closed circles: detergent
phase.
Figure 5. Sepharose CL-4B column chromatography following
partitioning in Triton X-114 of (A) the plasma membrane fraction and (B) the spent medium. Confluent HTSE cells were metabolically radiolabeled with [3H]glucosamine for 24 h, and both
the P2-II fraction and the spent media were subjected to phase
separation by Triton X-114 as described in M ATERIALS AND
METHODS. The resulting detergent and aqueous phases were individually applied to Sepharose CL-4B eluting with 50 mM sodium acetate, pH 7.0, containing 0.1% SDS. Open circles: aqueous phase; closed circles: detergent phase.
of GPI-PLD, as described in MATERIALS
(data not shown).
AND
METHODS
CsCl Equilibrium Density-Gradient Centrifugation
After CsCl equilibrium density-gradient centrifugation of
the membrane HMGP (double-radiolabeled with [3H]palmitic acid and [14C]glucosamine), radioactivity was found
in two major peaks, one with a density of 1.5 g/ml and another with 1.3 g/ml, and the profile of 3H-radioactivity was
virtually superimposable on that of 14C-radioactivity (Figure 7).
Sepharose CL-2B and DEAE-Sephacel
Column Chromatography
When 3H-HMGP from spent media and the plasma membrane were subjected to chromatography on a Sepharose
CL-2B column, 47% of the secreted and 14% of the membrane 3H-HMGP were excluded (Figure 8). Thus, HMGP
in the plasma membrane are also heterogeneous in size
and their average size seems to be smaller than that of secreted HMGP. The elution profile of DEAE-Sephacel anion exchange column chromatography (Figure 9) shows
that 59% of the total 3H was a pass-through (peak I),
whereas 41% was eluted with NaCl concentrations ranging from 0.15 to 0.5 M (peak II) with a peak fraction at
0.3 M.
Desialylation
Treatment of 3H-HMGP with 0.1 N sulfuric acid released
3
H-radioactivity, which accounted for 44% of the total 3H
radioactivity.
Immunoprecipitation with an Anti-Muc-1 Antibody
When the cell lysate was immunoprecipitated with an antiMuc-1 mucin antiserum (CT-1) and the precipitate subjected to Sepharose CL-4B column chromatography, most
of the 3H-radioactivity was included in the column (Figure
10A). The presence of the blocking peptide (0.1 mg/ml)
during the immunoprecipitation, however, resulted in disappearance of a major peak with Kd of 0.22 with no significant change in the size of 3H-Vo, which constitutes less than
4% of the total counts/min of all the eluted fractions (Figure 10A); this profile remained the same with higher concentrations (0.2 mg/ml and 0.4 mg/ml) of the blocking pep-
Paul, Lee, Hyun, et al.: Mucins on the Surface of Airway Epithelial Cells
687
Figure 7. CsCl equilibrium density gradient centrifugation
of the membrane HMGP. The membrane HMGP or Vo
metabolically double-radiolabeled with [ 3H]palmitic acid
and [14C]glucosamine were obtained as described in Figure
3. The radiolabeled HMGP preparation was mixed with
CsCl (a starting density of 1.5 g/ml) and centrifuged as described in MATERIALS AND METHODS. Note that both 3H
(closed circles) and 14C (open circles) radioactivities comigrate, suggesting the presence of covalent binding of lipids
to the membrane HMGP. Open triangles represent the
density of CsCl.
tide. On the other hand, when the membrane 3H-HMGP
was immunoprecipitated in the presence or absence of the
blocking peptide, most of the 3H activity was found in the
supernatants (Figure 10B) with little or no 3H in the precipitate of either sample (data not shown). This result was
reproduced in two additional experiments.
SDS-PAGE
As seen in Figure 11A, the glycoconjugates immunoprecipitated with the CT-1 antibody showed a smear starting
from the near origin down to about 170 kD. On the other
hand, the membrane HMGP remained at the origin. Extension of PAGE from 2 h (Figure 11A) to 6 h (Figure
11B) revealed very little, if any, overlap between the im-
Figure 8. Sepharose CL-2B column chromatography of HMGP.
The 3H-Vo sample obtained from Figure 3 was applied to a
Sepharose CL-2B column eluting with 50 mM sodium acetate,
pH 7.2, containing 0.1% SDS. Open circles: spent medium Vo
sample; closed circles: membrane Vo sample.
munoprecipitated sample and the HMGP, indicating that
most of the Muc-1 mucins, if not all, are smaller in size
than the HMGP of the plasma membrane.
Discussion
To identify the HMGP on the cell surface, we first purified
the plasma membrane by sequential and differential centrifugations (Figure 1). The purity of the plasma membrane was assessed by three criteria: (1) the presence of
marker enzymes, (2) their specific activity, and (3) the degree of cross-contamination with other membrane markers. Most of the activities of the plasma membrane enzymes
(alkaline phosphatase and 59-nucleotidase), the markers
used in this experiment, were found in the P2-II fraction
(Figure 2), indicating that this fraction is highly enriched
Figure 9. DEAE-Sephacel ion exchange chromatography of the
membrane HMGP. The membrane 3H-HMGP obtained from
Figure 3 was applied to the anion exchange column as described
in MATERIALS AND METHODS. The second peak appeared with
NaCl concentrations of 0.15 to 0.5 M, centering at 0.3 M. The dotted line represents the concentrations of NaCl.
688
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 19 1998
Figure 11. Comparison of the electrophoretic mobility of 3HHMGP and 3H-Muc-1 mucins on SDS-polyacrylamide gels. Both
the immunoprecipitated 3H sample from Figure 10 (lane 1) and
the membrane 3H-HMGP obtained from Figure 3 (lane 2) were
subjected to 5% SDS-PAGE with a 3% stacking gel at 100 V for
2 h (A) and 6 h (B) as described in MATERIALS AND METHODS.
3
H activities of the samples applied in A were 28,000 and 15,000
cpm for lanes 1 and 2, respectively, whereas those in B were
16,000 and 15,000 cpm. Myosin of rabbit muscles was used as a
prestained size marker for 205 kD.
Figure 10. Sepharose CL-4B gel-filtration column chromatography of the cell lysate or the plasma membrane HMGP following
immunoprecipitation with an anti-Muc-1 mucin antibody. Both
the cell lysate and the membrane HMGP of confluent HTSE cells
were incubated with an anti-Muc-1 mucin antibody (CT-1) in the
presence or absence of a blocking peptide followed by immunoprecipitation with Protein A-agarose beads as described in M ATERIALS AND METHODS. The resulting immunoprecipitate sample from the cell lysate (A) and the supernatant sample from the
membrane HMGP (B) were applied to Sepharose CL-4B columns. The amount of 3H activity immunoprecipitated from the
3
H-HMGP was negligible. Open circles: with the blocking peptide; closed circles: without the blocking peptide.
with the plasma membrane. Because 59-nucleotidase is
present in both the plasma and the nuclear membranes,
whereas alkaline phosphatase is found only in the plasma,
the latter has been chosen to be a better marker for the
plasma membrane (11). The specific activity of alkaline
phosphatase (848.5 nM 3 min21 3 mg protein21) in the
P2-II fraction of HTSE cells was within the range of those
reported for the apical membrane of bovine tracheal epithelial cells (9). This value was 55 times greater than that
of the crude homogenate, which is already 10- to 20-fold
higher than the values reported for renal and intestinal luminal membranes (25, 26). The P2-II fraction contained
very little or no enzyme activities of the Golgi, the rough
endoplasmic reticulum, and the mitochondria, but about
5% of the lysosomal enzyme activity (Figure 2). On the basis of this purity data, we conclude that the P2-II fraction is
highly enriched with the plasma membrane and contains
relatively small degrees of contamination with other membranes.
We then examined whether P2-II fraction contains
HMGP. Mucins are generally defined as HMGP with O-glycosidic linkage, which are excluded from the Sepharose
CL-4B gel filtration column (17). Therefore, we first applied the P2-II fraction to a Sepharose CL-4B column eluting with a detergent-containing buffer; the presence of a
detergent in the buffer has been shown to be crucial for
dissociation of hydrophobic macromolecules from mucins
(23). Figure 3 shows the profile of Sepharose CL-4B column chromatography of the P2-II fraction. To determine
whether Vo fractions contain some proteoglycans, we treated
the Vo with various proteoglycan-digesting enzymes using
the same protocol we used previously for secreted Vo fractions (17): chondroitinase ABC for hyaluronic acid, dermatan sulfate, and chondroitin sulfate proteoglycans; heparitinase for heparan sulfate proteoglycan. The Vo fractions
were totally resistant to either chondroitinase ABC (Figure 4A) or heparitinase (data not shown). Thus, total resistance to these enzymes indicates that they are not those
proteoglycans. On the other hand, complete degradation
Paul, Lee, Hyun, et al.: Mucins on the Surface of Airway Epithelial Cells
of these glycoconjugates by pronase (Figure 4B) indicates
that these are glycoproteins.
To see whether the Vo fractions contain O-linked glycoproteins, the Vo sample was subjected to the reductive
mild alkaline hydrolysis, which was shown to deglycosylate
O-linked oligosaccharides of mucins (27). Figure 4C shows
that 3H-glycoconjugates in the Vo fractions were completely
digested by the alkaline borohydride treatment. Thus the
result indicates that the plasma membrane contains HMGP.
The plasma membrane 3H-HMGP were further characterized as follows. We first determined whether these
3
H-HMGP are integral membrane glycoproteins. Triton
X-114 is frequently used to distinguish various membrane
glycoproteins between integral and external glycoproteins
(18). Figure 5A shows that most of the membrane HMGP
are partitioned in the detergent phase, indicating that they
are mostly integral glycoproteins. In contrast, secreted
mucins were found mostly in the aqueous phase (Figure
5B). We then examined whether they are anchored to the
membrane through GPI linkage. The GPI anchor has been
shown in a number of macromolecules and is generally
susceptible to enzymes such as PI-PLC or GPI-PLD (19).
If the membrane HMGP is bound to the plasma membrane through the GPI linkage, treatments with these enzymes should result in a decrease in the 3H-Vo of the detergent phase with a concomitant increase in the 3H-Vo of
the aqueous phase. Figure 6 shows that the enzyme treatment did not affect the partitioning of the 3H-Vo fractions,
indicating that 3H-HMGP in the plasma membrane are
not anchored via GPI. We thus concluded that the plasma
membrane HMGP are very likely integral or transmembrane glycoproteins.
The plasma membrane 3H-HMGP were then compared
with secreted mucins in some of the characteristics previously reported on secreted mucins (17, 23)—the distributions of the buoyant density, the size, and the charge. The
distribution of the buoyant density of the plasma membrane 3H-HMGP (Figure 7) was almost identical to that of
secreted mucins (23). The low-density HMGP is likely due
to the association of lipids and therefore might be similar
to the ones which are releasable by dirhamnolipid, a toxin
from Pseudomonas aeruginosa, as recently reported by
Fung and coworkers (28). On the other hand, both the size
(Figure 8) and charge distribution (Figure 9) seem somewhat different between these two mucins. The overall
greater size of secreted mucins, however, does not seem to
be a result of aggregation because a highly dissociative
condition was used for the column elution (23). The
slightly greater acidity of the membrane HMGP is likely
due to a difference in 3H-sialic acid contents between the
membrane HMGP and secreted mucins—44 and 20–30%
(17) of the total 3H activity, respectively—and/or a difference in sulfation (data not available). These biochemical
characteristics of the plasma membrane and secreted mucins are summarized in Table 1.
In light of the fact that these membrane HMGP are not
only the cell-surface mucins but also highly sialylated, it
was suspected that they might be the product of the Muc-1
mucin gene which was originally identified in breast and
pancreatic cancer cells (27, 29). The gene is expressed not
only in the cancer cells but also in various secretory epi-
689
TABLE 1
Comparison between the plasma membrane
and secreted mucins
Characteristic
1. Size
CL-2B profile
2. Charge
DEAE-Sephacel profile
Sialylation
3. Density
CsCl density gradient
4. Proteoglycan-digestion
enzymes
Membrane
Secreted**
14% as Vo*
47% as Vo*
Pass-through: 59%†
Retained: 41%†
centered at 0.3 M NaCl
44% of total 3H‡
65%
35%
0.26 M
20–45%
Two peaks§
(1.5 g/ml and 1.3 g/ml)
Negativei
Same
Negative
Sources of the data:
* Figure 8.
†
Figure 9.
‡
See Desialylation section of RESULTS.
§
Figure 7.
i
Figure 4.
** See Kim and associates (17).
thelial cells, including the airway (5, 24). We have recently
shown that the present HTSE cells also express the Muc-1
gene at confluence (6), and the deduced amino-acid sequence of the Muc-1 complementary DNA revealed the
presence of a transmembrane domain (6, 30, 31). To identify Muc-1 mucins from HTSE cells, the whole cell lysate
was immunoprecipitated with an anti-Muc-1 mucin antibody (CT-1) (24) in the presence or absence of a blocking
peptide (24), and the immunoprecipitated sample was subjected to Sepharose CL-4B column chromatography. This
antibody recognizes the cytoplasmic tail of Muc-1 mucins
and has been shown to cross-react with Muc-1 mucins from
various species (24). Figure 10A shows that Muc-1 mucins
are included in the column but not excluded. To confirm
that the Vo fractions or 3H-HMGP are not Muc-1 mucins,
we performed two additional experiments as follows. First,
we examined whether the membrane 3H-HMGP bind specifically to CT-1. When the plasma membrane 3H-HMGP
sample (3,000 cpm) was immunoprecipitated with CT-1 in
the presence and absence of the blocking peptide, the
3
H-radioactivity of the final precipitates were almost negligible (less than 50 cpm) in both cases. Lack of 3H-activity
in the precipitates was not due to the final washing step
with water because the result was exactly the same without
the final wash (data not shown). On the other hand, Vo
profiles of the resulting supernatant samples were virtually
superimposable (Figure 10B). Therefore, it appears that
the small amount of counts per minute present in the Vo in
Figure 10A might be a result of nonspecific binding. Second, we compared the sizes of these two samples using
SDS-PAGE. Figure 11 shows that, while HMGP are at the
origin of the running gel, Muc-1 mucins are distributed
from about 170 kD to near the origin. The extended PAGE
seems to show little or no overlap between the HMGP and
Muc-1 mucins. These results, however, do not conclusively
rule out the presence of some Muc-1 mucins in the HMGP
690
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 19 1998
because “high molecular-mass” Muc-1 mucins as previously reported in a human breast cancer cell line (139H2)
(32) that are not immunoprecipitable with the CT-1 antibody might be present. Nevertheless, the present data seem
to indicate that most, if not all, of the membrane HMGP
are not Muc-1 mucins.
In view of all of the biochemical and physical characteristics examined here, the plasma membrane HMGP appear to be somewhat similar to the secreted mucins (17),
albeit with some differences in both the size and the
charge of the molecules. It is possible that the membrane
HMGP are released from the cell surface, perhaps by
some unknown proteolytic mechanisms (3, 33), to become
part of the secreted mucins. It is also possible that they derive from the secretory granules as previously proposed
(1). In this model, mucins are packaged inside the secretory granules or vesicles in either free or membranebound forms. During exocytosis, mucins in the free form
are released into the extracellular fluid, whereas mucins in
the membrane-bound form remain bound to the apical
surface of the cells. Further work is necessary, however, to
confirm this possibility. Given that these are extremely
large glycoproteins on the cell surface that are not only
highly glycosylated but also negatively charged, it might be
possible that they play an important role in the epithelial
cell physiology and pathology. Possible functions of these
cell-surface HMGP are currently under investigation.
Acknowledgments: This work was supported by a grant from National Institutes
of Health, RO1 HL-47125.
References
1. Kim, K. C. 1994. Regulation of airway goblet cell mucin secretion. In Airway Secretion: Physiological Bases for the Control of Mucus Hypersecretion. Lung Biology of Health and Disease Series. T. Takishima and S.
Shimura, editors. Marcel Dekker, Inc., New York. 433–449.
2. Shimura, S., and T. Takishima. 1994. Airway submucosal gland secretion. In
Airway Secretion: Physiological Bases for the Control of Mucus Hypersecretion. Lung Biology of Health and Disease Series. T. Takishima and S.
Shimura, editors. Marcel Dekker, Inc., New York. 325–398.
3. Kim, K. C., K. Wasano, R. M. Niles, J. E. Schuster, P. J. Stone, and J. S.
Brody. 1987. Human neutrophil elastase releases cell surface mucins from
primary cultures of hamster tracheal epithelial cells. Proc. Natl. Acad. Sci.
USA 84:9304–9308.
4. Wasano, K., K. C. Kim, R. M. Niles, and J. S. Brody. 1988. Membrane differentiation markers of airway epithelial secretory cells. J. Histochem. Cytochem. 36:167–178.
5. Hollingsworth, M. A., S. K. Batra, W. N. Qi, and J. R. Yankaskas. 1992.
MUC-1 mucin mRNA expression in cultured human nasal and bronchial
epithelial cells. Am. J. Respir. Cell Mol. Biol. 6:516–520.
6. Park, H. R., S. W. Hyun, and K. C. Kim. 1996. Expression of MUC-1 mucin
gene by confluent hamster tracheal epithelial cells in primary culture. Am.
J. Respir. Cell Mol. Biol. 15:237–244.
7. Hilkens, J., M. J. L. Lightenberg, H. L. Vos, and S. V. Litvinov. 1992. Cell
membrane-associated mucins and their adhesion-modulating property.
Trends Biol. Sci. 17:359–363.
8. Kim, K. C. 1985. Possible requirement of collagen gel substratum for production of mucin-like glycoproteins by primary rabbit tracheal epithelial
cells in culture. In Vitro 21:617–621.
9. Langridge, J. E., M. Field, and W. P. Dubinsky. 1983. Isolation of transporting plasma membrane vesicles from bovine tracheal epithelium. Biochim.
Biophys. Acta 731:318–328.
10. Pekarthy, J. M., J. Short, A. Lansing, and I. Lieberman. 1972. Function and
control of liver alkaline phosphatase. J. Biol. Chem. 247:1767–1774.
11. Howard, E. W. 1979. Preparation and characterization of mammalian
plasma membranes. In Laboratory Techniques in Biochemistry and Molecular Biology. T. S. Work and E. Work, editors. North-Holland Publishing Company, New York. 102–259.
12. Gianetto, R., and C. De Duve. 1955. Comparative study of the binding of
acid phosphatase, b-glucuronidase and cathepsin by rat-liver particles.
Biochem. J. 59:433–438.
13. Cooper, J. R. 1970. Thiamine pyrophosphatase from rabbit brain. In Methods in Enzymology, Vol. XVIII, Part A. S. P. Colowick and N. O. Kaplan,
editors. Academic Press, New York. 227–229.
14. Swanson, M. A. 1955. Glucose-6-phosphatase from liver. In Methods in Enzymology, Vol. II, Part A. S. P. Colowick and N. O. Kaplan, editors. Academic Press, New York. 541–543.
15. Earl, D. C., and A. Korner. 1965. The isolation and properties of cardiac ribosomes and polysomes. Biochem. J. 94:721–734.
16. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248–254.
17. Kim, K. C., J. I. Rearick, P. Nettesheim, and A. M. Jetten. 1985. Biochemical characterization of mucous glycoproteins synthesized and secreted by
hamster tracheal epithelial cells in primary culture. J. Biol. Chem. 260:
4021–4027.
18. Bordier, C. 1981. Phase separation of integral membrane proteins in Triton
X-114 solution. J. Biol. Chem. 256:1604–1607.
19. Conzelman, A., A. Spiazzi, and C. Bron. 1987. Glycolipid anchors are attached to Thy-1 glycoprotein rapidly after translation. Biochem. J. 246:
605–610.
20. Low, M. G., and A. R. Saltiel. 1988. Structural and functional roles of glycosylphosphatidylinositol in membranes. Science 239:268–275.
21. Low, M. G., and A. R. S. Prasad. 1988. A phospholipase D specific for the
phosphatidylinositol anchor of cell-surface proteins is abundant in plasma.
Proc. Natl. Acad. Sci. USA 85:980–984.
22. Paul, E., F. A. Leblond, and D. Lebel. 1991. In resting conditions, the pancreatic granule membrane protein GP-2 is secreted by cleavage of its glycosylphosphatidylinositol anchor. Biochem. J. 277:879–881.
23. Kim, K. C. 1991. Mucin-like glycoproteins secreted from cultured hamster
tracheal surface epithelial cells: their hydrophobic nature and amino acid
composition. Exp. Lung Res. 17:65–76.
24. Pemberton, L., J. Taylor-Papadimitriou, and S. J. Gendler. 1992. Antibodies
to the cytoplasmic domain of the MUC-1 mucin show conservation
throughout mammals. Biochem. Biophys. Res. Commun. 185:167–175.
25. Kessler, M., O. Acuto, C. Storelli, H. Murer, M. Martin, and G. Semenza.
1978. A modified procedure for the rapid preparation of efficiently transporting vesicles from small intestinal brush border membranes. Biochim.
Biophys. Acta 506:136–154.
26. Turner, R. J., and M. Silverman. 1978. Sugar uptake into brush border vesicles from dog kidney. Biochim. Biophys. Acta 507:305–321.
27. Carlson, D. M. 1968. Structures and immunochemical properties of oligosaccharides isolated from pig submaxillary mucins. J. Biol. Chem. 243:
616–626.
28. Fung, D. C. K., M. Somerville, P. S. Richardson, and J. K. Sheehan. 1995.
Mucus glycoconjugate complexes released from feline trachea by a bacterial toxin. Am. J. Respir. Cell Mol. Biol. 12:296–306.
29. Gendler, S. J., and A. P. Spicer. 1995. Epithelial mucin genes. Annu. Rev.
Physiol. 57:607–634.
30. Gendler, S. J., C. A. Lancaster, J. Taylor-Papadimitriou, T. Duhig, N. Peat,
J. Burchell, L. Pemberton, E. N. Lalani, and D. Wilson. 1990. Molecular
cloning and expression of human tumor-associated polymorphic epithelial
mucin. J. Biol. Chem. 265:15286–15293.
31. Lan, M. S., S. K. Batra, W. N. Qi, R. S. Metzgar, and M. A. Hollingsworth.
1990. Cloning and sequencing of a human pancreatic tumor mucin cDNA.
J. Biol. Chem. 265:15294–15299.
32. Ligtenberg, M. J. L., L. Kruijshaar, F. Buijs, M. van Meijer, S. V. Litvinov,
and J. Hilkens. 1992. Cell-associated episialin is a complex containing two
proteins derived from a common precursor. J. Biol. Chem. 267:6171–6177.
33. Boshell, M., E. N. Lalani, L. Pemberton, J. Burchell, S. Gendler, and J. Taylor-Papadimitriou. 1992. The product of the human MUC-1 gene when secreted by mouse cells transfected with the full-length cDNA lacks the cytoplasmic tail. Biochem. Biophys. Res. Commun. 185:1–8.