Expression of Pneumocystis jirovecii Major Surface Glycoprotein in Saccharomyces cerevisiae
Katherine J. England
Joseph A. Kovacs
Critical Care Medicine Department, National Institutes of Health (NIH) Clinical Center
The major surface glycoprotein (Msg), which is the most abundant protein expressed on the cell surface of Pneumocystis organisms, plays an important role in the attachment of this organism to epithelial cells and macrophages. In the present study, we expressed Pneumocystis jirovecii Msg in Saccharomyces cerevisiae, a phylogenetically related organism. Full-length P. jirovecii Msg was expressed with a DNA construct that used codons optimized for expression in yeast. Unlike in Pneumocystis organisms, recombinant Msg localized to the plasma membrane of yeast rather than to the cell wall. Msg expression was targeted to the yeast cell wall by replacing its signal peptide, serine-threonine-rich region, and glycophosphatidylinositol anchor signal region with the signal peptide of cell wall protein -agglutinin of S. cerevisiae, the serine-threonine-rich region of epithelial adhesin (Epa1) of Candida glabrata, and the carboxyl region of the cell wall protein (Cwp2) of S. cerevisiae, respectively. Immunofluorescence analysis and treatment with -1,3 glucanase demonstrated that the expressed Msg fusion protein localized to the yeast cell wall. Surface expression of Msg protein resulted in increased adherence of yeast to A549 alveolar epithelial cells. Heterologous expression of Msg in yeast will facilitate studies of the biologic properties of Pneumocystis Msg.
Pneumocystis is a genus of opportunistic pathogens
that cause pneumonia in human immunodeficiency
virus (HIV)infected patients, as well as other
immunocompromised patients [1, 2]. This genus is
genetically divergent: each host species is infected by a
genetically distinct strain that represents Pneumocystis
species, with Pneumocystis jirovecii infecting humans,
Pneumocystis carinii infecting rats, and Pneumocystis
murina infecting mice . Major surface
glycoprotein (Msg; also called gpA), is the most abundant
surface protein of Pneumocystis organisms and is
encoded by a multicopy gene family . Msg
expression is regulated by a single expression site
encoding the upstream conserved sequence (UCS), which
contains a signal peptide that likely directs the protein
to the endoplasmic reticulum for subsequent
processing, including cleavage of the UCS, glycosylation, and
surface export . Msg variants are related but
unique and potentially permit the organism to vary its
surface antigen as a means of evading host immune
In yeast, many cell surface proteins are
glycophosphatidylinositol (GPI)anchored proteins that are
covalently linked to cell wall glucans. Msg has features
characteristic of GPI-anchored proteins, including a
signal peptide at the UCS and a hydrophobic tail at the
C-terminus. The latter can function as a GPI anchor in
mammalian cell systems . Msg may play an
important role in host-parasite interactions by mediating
adherence of Pneumocystis organisms to host alveolar
epithelial cells and macrophages . In vitro
studies to better define the interaction of Pneumocystis
organisms with host lung alveolar epithelial cells are
limited because of the lack of an efficient culture system
for Pneumocystis species. Heterologous expression of
epithelial adhesion protein (Epa1) of Candida glabrata,
a GPI-anchored cell wall protein, has been used to
study adherence properties of this protein to cultured
human epithelial cells . In the present study, we undertook
to express Msg on the cell wall of the yeast Saccharomyces
cerevisiae, an organism that is phylogenetically related to
Pneumocystis organisms and to test its ability to mediate adherence to a
lung alveolar epithelial cell line.
MATERIALS AND METHODS
Yeast Strains, Culture, and Transformation
S. cerevisiae strain YPH 499 was obtained from Stratagene
(Santa Clara, CA). Yeasts were grown at 30C in yeast peptone
dextrose adenine medium. For plasmid maintenance, synthetic
dextrose medium without uracil (SD dropout medium) was
used. Yeast cultures were grown overnight in 2% raffinose
instead of dextrose, before inducing protein expression in 2%
galactose for 4 hours. Transformation was carried out using
lithium acetate as described in the pESC Yeast Epitope Tagging
Vectors instruction manual (Stratagene).
Construction of msg Expression Plasmid
A DNA construct using S. cerevisiae codon preference was
synthesized (GenScript USA, Piscataway, NJ) to encode full-length
Msg, including the UCS (GenBank accession no. AF321555),
followed by a P. jirovecii Msg variant, Msg 32 (corresponding to
bp 213080 of GenBank accession no. AF033212), that was
randomly selected from previously cloned Msg variants . A
cloning artifact stop codon (TGA) present at 30273029 bp of
Msg 32 was modified to TGG. A hemagglutinin (HA) tag, a
histidine (His) tag, and an enterokinase site were added
between the UCS and Msg variable region.
The synthesized insert was cloned into S. cerevisiae vector
pESC-URA (Stratagene), which has a GAL1 promoter. The
insert with or without the UCS sequence was also cloned into
the destination vector, pBC542 (a kind gift from Dr Brendan
Cormack) (GenBank accession no. EU418980), using Gateway
technology with the Clonase II kit (Invitrogen, Grand Island,
NY). The 3 end of the construct (corresponding to 27723080
bp of msg 32) was deleted before cloning into pBC542, which
contains sequences coding for the serine-threoninerich region
(338580 amino acids; GenBank accession no. AAQ82691) of
epithelial adhesin (Epa1) of C. glabrata, followed by sequences
coding for 2192 amino acids of Cwp2 (GenBank accession
no. EEU05144), a cell wall protein of S. cerevisiae. The vector
has a tef1 promoter and encodes 3 HA epitopes upstream of
Epa1 . We added sequences encoding a FLAG epitope
immediately after the signal peptide.
A construct encoding 59 amino acids (corresponding to 917
975 amino acids of Msg 32) of the serine-threoninerich region
of Msg was synthesized by Genscript. Four repeats of the same
sequence were used to replace the Epa1 sequence in the vector.
In certain cases, the Msg signal peptide
(MRVALFALSAQVGCA) was replaced with the signal peptide (MFTFLK
IILWLFSLALA) of -agglutinin (GenBank accession no.
AAA34417.1), a cell wallanchored surface protein of S.
cerevisiae , using the QuikChange II site-directed mutagenesis
Polymerase Chain Reaction (PCR)
PCR was performed using AccuPrime Pfx (Invitrogen) or High
Fidelity PCR master mix (Roche Diagnostics, Indianapolis, IN).
For Gateway cloning, the amplification was performed using
primers with attB1 sites. General PCR conditions were as
described previously .
Site-Directed Mutagenesis and DNA Sequencing
All mutations and deletions were done using the QuikChange
II or QuikChange II XL site-directed mutagenesis kit
(Stratagene) according to the guidelines of the manufacturer. Plasmid
DNA preparations made using the QIAprep spin miniprep kit
(Qiagen, Valencia, CA) were sequenced using an ABI 3100
Genetic Analyzer (Applied Biosystems, Carlsbad, CA).
Preparation of Spheroplasts
To make spheroplasts, the CelLytic Y Plus Kit (Sigma-Aldrich,
St. Louis, MO) was used. Yeast cells transformed with the msg
constructs or vector alone were grown to mid-log phase, and
the cells were treated with -1,3 glucanase until spheroplast
formation was complete. Spheroplasts were centrifuged at 9300 g
for 5 minutes. The pellet containing spheroplasts was used for
immunofluorescent staining and Western blot analysis, and the
supernatant was used for Western blot analysis.
Preparation of Cell Membranes and Deglycosylation
Spheroplasts resuspended in extraction buffer (CelLytic Y Plus
Kit, Sigma-Aldrich) containing protease inhibitor (P8849,
Sigma-Aldrich) were disrupted using glass beads, followed by
centrifugation at 1020 g for 5 minutes at 4C. The pellet (cell
debris and nuclei) was analyzed by Western blot. The
supernatant was centrifuged at 150 000 g for 30 minutes at 4C, and
the resulting supernatant (cytosol fraction) and pellet (cell
membrane) were also analyzed by Western blot.
To deglycosylate N-linked oligosaccharides, proteins
extracted from the cell membrane fraction were denatured for 10
minutes, incubated with
N-linked-glycopeptide-(N-acetylbeta-D-glucosaminyl)-L-asparagine amidohydrolase (PNGase
F) for 1 hour at 37C (New England Biolabs, Ipswich, MA),
and analyzed by Western blot.
-1,3 Glucanase Digestion and Protein Extraction
Yeast cells expressing Msg on the surface were treated with
1,3 glucanase (lyticase, CelLytic Y Plus Kit, Sigma-Aldrich)
overnight at 37C. The supernatant obtained was analyzed by
Western blot, and the pellet was used for immunofluorescence
analysis. In some experiments, yeast cells expressing Msg were
broken with glass beads to prepare cell extracts for Western blot
formaldehyde, and analyzed by flow cytometry, using FACS
Calibur (BD Biosciences, San Jose CA).
Western Blot Analysis
Western blot analysis was performed as previously described
. The blots were probed with a 1:200 dilution of
peroxidase-conjugated anti-HA antibody (Santa Cruz Biotechnology,
Santa Cruz, CA). Immunoreactive bands were visualized using
BM Blue POD Substrate, precipitating (Roche Diagnostics).
Immunofluorescence Microscopic Analysis
Intact yeast cells or spheroplasts were added to individual wells
of 8-well slides and then were heat fixed. Samples were blocked
with 3% bovine serum albumin ( Jackson Immunoresearch
Laboratories, West Grove, PA), followed by incubation with a
1:100 dilution of fluorescein isothiocyanate (FITC)conjugated
anti-HA antibody (Santa Cruz Biotechnology) at 37C for 1
hour. The slides were washed with phosphate-buffered saline
(PBS), dried, and sealed with coverslips with anti-fade
mounting medium. Images were collected using a fluorescence
microscope (Nikon Eclipse E800, Barber Optics, Columbia, MD).
Yeast cells expressing Msg were incubated with 1:10 dilution of
FITC-conjugated anti-HA antibody (Santa Cruz
Biotechnology) at 4C for 60 minutes, washed with PBS, fixed with 2%
Yeast Cell Adherence Assay
Yeast cells transformed with msg-pESC-URA vector or empty
vector were grown overnight at 30C and then induced with
galactose for 4 hours. To fluorescently label the cells, 2 L of 5
mM carboxyfluorescein diacetate, succinimidyl ester (CFSE)
solution (CellTrace CFSE cell proliferation kit, Molecular
Probes) in 100 L of 0.5 M sodium acetate ( pH 4.0), was added
to 900 L of the cells (1.0 OD600/mL). The cells were incubated
at 30C for 1 hour, 500 L of cold culture medium was added,
and the cells were incubated on ice for 5 minutes. The cells
were centrifuged and resuspended in 500 L of PBS containing
2% dextrose. Labeling efficiency was checked under a
A549 epithelial cells (ATCC, CCL-185) were cultured in
6well plates in F12 medium containing 10% fetal bovine serum
(FBS). Before the adherence assay was performed, the medium
was changed to one without FBS. Labeled yeast cells were
added to the culture, and adherence was initiated by brief
centrifugation. After the cells were incubated for 1 hour, the
nonadherent cells were removed by washing 3 times with PBS. The
epithelial cells along with adherent yeast cells were scraped in
PBS, and the fluorescence was measured by a fluorometer
(Wallac Victor 3, PerkinElmer, Santa Clara, CA).
Characterization of P. jirovecii Msg Expression in S. cerevisiae
A DNA construct encoding the full-length Msg 32 open
reading frame, including the UCS, was synthesized using
codons optimized for expression in S. cerevisiae [11, 22]. The
schematic diagram of the construct is shown in Figure 1A. The
construct was cloned into pESC-URA vector and transformed
into yeast. Vector alone (no insert) served as a negative control.
After 4 hours of galactose induction, Western blot analysis of
cell extracts by using an anti-HA tag antibody identified 2
bands of the expected size, approximately 130 kDa (Figure 1B);
no reactivity was seen with the no insert control. To determine
whether recombinant Msg is expressed on the cell surface, yeast
cells were analyzed by immunofluorescence microscopy, using
FITC-conjugated anti-HA antibody; no fluorescence was seen
Additional studies were undertaken to localize Msg.
Galactose-induced yeast cells were treated with -1,3 glucanase to
digest the cell wall, followed by centrifugation to separate
spheroplasts from the digested cell wall products. Western blot
analysis using anti-HA antibody showed that Msg is expressed
within the spheroplasts and not in the cell wall (Figure 2A).
To further localize Msg, spheroplasts were fractionated by
vortexing with glass beads, followed by low-speed
centrifugation to eliminate nuclei and cell debris and then by high-speed
centrifugation to separate cell membrane from cytosol. By
Western blot analysis using anti-HA antibody, Msg was
localized to the cell membrane fraction (Figure 2A); no reactivity
was seen with the cytosol fraction. Immunofluorescence
microscopy using FITC-conjugated anti-HA antibody verified
that Msg was expressed on the cell membrane of spheroplasts
(Figure 2B). Thus, full-length Msg expressed in yeast is
localized to the plasma membrane and not the cell wall.
Because recombinant Msg showed 2 bands in Western blot
analysis and because native Msg is known to be glycosylated
, we explored the possibility that differential glycosylation
accounted for the 2 bands. Deglycosylation with pNGase F
resulted in the disappearance of the higher-sized band, indicating
that recombinant Msg is glycosylated in yeast (Figure 2C).
Mutation of the GPI Cleavage Site Does Not Change the
Localization of Expressed Msg
In Pneumocystis organisms, Msg is a cell surface protein that
can be released by glucanase treatment . In yeast, most
proteins extracted by glucanase treatment are attached to the
glucan of the cell wall through a modified GPI moiety [26, 27].
The C-terminal of the precursor to GPI-anchored proteins has
an anchor attachment signal with an anchor attachment site
(-site), a hydrophilic region followed by a hydrophobic region
of 1015 amino acid residues . In yeast, amino acids
flanking the -site help determine whether protein is localized to
the cell wall or cell membrane [29, 30]. We thus mutated the
amino acids near or at the putative -site of Msg to determine
whether that would be sufficient to allow localization to the
cell wall, using as a template the sequence in yjr151c, a
GPIdependent cell wall protein of S. cerevisiae  (Figure 3A).
Two sets of mutations were performed. In the first, amino acid
residues EDVKP of Msg at -minus site were replaced by
VSINT. In the second, the VSINT replacement was maintained,
the putative -site was changed from S to N, and the 2 amino
acid residues downstream of the -site, EG, were replaced by
GA (Figure 3A), to more closely resemble the -site and
flanking regions of yjr151c, given that these may be important for
GPI anchoring . By Western blot analysis using anti-HA
antibody, both mutated proteins were expressed in yeast
(Figure 3B). However, no immunoreactivity was detected on
the yeast surface (data not shown). To see whether the native
signal peptide prevented cell wall localization, we replaced the
Msg signal peptide in constructs with mutation 1 and 2 with
the signal peptide of -agglutinin, a cell wallanchored protein
of S. cerevisiae. Again there was no localization to the cell wall
(data not shown).
Surface Expression of Msg in Yeast
To target the expression of Msg to the cell wall of yeast, we used
pBC542, which contains sequences that code for the
serinethreoninerich region of epithelial adhesin (Epa1) of C.
glabrata followed by the C-terminal domain of cell wall protein
(Cwp2) of S. cerevisiae. Fusion proteins with inserts upstream
of these regions are expressed on the cell surface of S. cerevisiae
. We generated 3 different constructs to express the Msg
fusion protein (Figure 4A). All constructs included a near
fulllength (12771 bp of msg 32) codonbiased msg sequence; the
deleted region (27723080 bp) encodes a serine-threoninerich
region and a GPI anchor signal. Construct 1 also included a
region encoding the Msg signal peptide and UCS. In construct
2 the Msg signal peptide was replaced by a yeast signal peptide,
and in construct 3 the UCS sequence was deleted from
construct 2. By immunofluorescence, cell surface expression of
Msg was observed only for yeast cells transformed with
constructs 2 and 3, which included the yeast signal peptide
(Figure 4B). To confirm the cell surface expression of the
protein, we analyzed the cells by flow cytometry. When cells
were transformed with constructs 2 and 3, approximately 70%
showed cell surface immunofluorescence, while construct 1
showed no fluorescence (Figure 4C).
By Western blot, immunoreactivity to an approximately
150kDa band was seen with protein extracts prepared from cells
transformed with construct 1, while protein extracts prepared
from cells transformed with constructs 2 and 3 showed higher
molecular weight smears (Figure 4D), indicating that the
recombinant protein is still attached to remnants of the cell wall. Thus,
construct 1 is expressed but not transferred to the surface, while
the other 2 constructs are expressed on the surface of yeast.
Yeast cells expressing Msg were resuspended in buffer with
or without -1,3 glucanase and incubated overnight at 37C.
The pellet (residual yeast cells) obtained after centrifugation
was analyzed by immunofluorescence microscopy (Figure 4E).
Untreated cells showed fluorescence, while glucanase-treated
cells showed little fluorescence. By Western blot, an
approximately 150-kDa band was detected in the glucanase-treated
supernatant after centrifugation, demonstrating that Msg was
released from the cell wall.
Msg Serine-ThreonineRich Region Does Not Target the Yeast
A long serine-threoninerich region (200300 amino acids) appears
to play an important role in localizing proteins to the yeast cell
wall . To see whether sequences encoding the
serine-threoninerich region of P. jirovecii Msg could function in this manner,
we replaced the serine-threoninerich region of Epa1 encoded in
construct 2 with 4 repeats (59 amino acids each) of the Msg
serine-threoninerich region. No reactivity was seen by
immunofluorescence microscopy or flow cytometry in yeast transformed
with this construct (Figure 5A and 5B). Western blot analysis
demonstrated that the Msg fusion protein is expressed, with
immunoreactivity to an approximately 150-kDa band (Figure 5C).
Thus, the serine-threoninerich region of Msg was unable to
target localization of the expressed fusion protein to the cell wall.
Yeast Expressing Msg Show Increased Adherence
To improve expression, the insert from construct 3 was cloned
into pESCURA vector downstream of the GAL1 promoter.
Following induction with galactose, the cell surface expression was
analyzed by immunofluorescence microscopy (Figure 6A). Cells
transformed with the msg construct showed fluorescence when
stained with FITC-conjugated anti-HA antibody, while cells
transformed with vector alone showed no immunoreactivity.
Approximately 84% of the cells showed immunofluorescence
when analyzed by flow cytometry (Figure 6B) as compared to
70% with the tef1 promoter.
This construct was used to determine whether Msg could
mediate adherence. Yeast cells expressing Msg were labeled with
CFSE to facilitate quantitation of adherent organisms;
fluorescent microscopy verified labeling of most cells (Figure 7A).
Labeled cells were added to cultured A549 epithelial cells grown
in 6-well plates. After incubation for 1 hour, yeast cells
expressing Msg showed an approximately 9-fold increase in
adherence as compared to the negative control (vector with no insert;
In the present study, we expressed a full-length P. jirovecii Msg
for the first time, using a yeast expression system. We found
that in this heterologous system the native protein sequence
targeted Msg to the cell membrane rather than to the cell wall;
we also demonstrated that modification of the signal peptide, as
well as the serine-threoninerich region plus the GPI anchor,
resulted in cell wall expression. While the mechanism of
expression in yeast is likely different from mechanisms used by
Figure 6. Improved cell surface expression of Msg fusion protein under
the control of gal1 promoter. Construct 3 (Figure 4A) was cloned into
pESC-URA vector downstream of the gal1 promoter. After 4 hours of
galactose induction, the yeast transformants were analyzed for the cell surface
expression of Msg by immunofluorescence, using fluorescein
isothiocyanateconjugated anti-hemagglutinin (HA) antibody. A,
Immunofluorescence microscopic analysis. Cells transformed with the msg construct
showed immunofluorescence, whereas cells transformed with vector
alone showed no fluorescence. B, Flow cytometric analysis. The histogram
shows the cell counts on the y-axis and fluorescence intensity on the
x-axis. Solid histogram shows nonspecific staining with the negative
control (mouse immunoglobulin G), and the open histogram shows specific
staining with the anti-HA antibody. Approximately 84% of the cells
showed surface expression of the Msg fusion protein; vector alone
showed no staining.
Pneumocystis organisms, expression in yeast provides a tool to
facilitate studies of the biological properties of Msg.
We were previously unsuccessful in expressing full-length
Msg in different expression systems, including bacteria, yeast,
and baculovirus, and thus used fragments of the gene to
produce recombinant proteins. Possible reasons for the failure
may be the large size of Msg, the lack of codon optimization, or
the potential toxicity to the cells because of the presence of
multiple cysteine residues. In the current study, the msg DNA
sequence was optimized to accommodate the S. cerevisiae
The goal of our studies was to express Msg in yeast to
ultimately allow us to study host cellMsg interactions. Given the
close phylogenetic relationship between the organisms , we
hypothesized that Msg expressed in S. cerevisiae would be
processed as in Pneumocystis organisms, where it is targeted to an
external location in the cell wall. However, we found that
recombinant Msg using the native protein sequence was not
expressed on the surface but localized to the plasma membrane.
Msg has a signal peptide at the N-terminal and a hydrophobic
tail at the C-terminal, which are characteristics of
GPIanchored proteins. The ferret Pneumocystis Msg hydrophobic
tail has been shown to function as a GPI signal in Cos cells .
While the GPI anchor targets a plasma membrane location
in many organisms, in yeast most GPI proteins are attached to
the cell wall, using a GPI remnant that links to the -glucan of
the cell wall . Based on localization of recombinant Msg to
the plasma membrane, the native hydrophobic leader and tail
sequences appear to serve as true GPI signals in yeast but are
inadequate to lead to cell wall localization. In yeast, amino acid
residues V, I, or L at the -4/5 site and Y or N at the -2 site
may act as a signal for cell wall incorporation . However,
modification of both the leader and the putative -site and
flanking amino acids to yeast sequences that, in other studies,
target a cell wall location in yeast, was insufficient to allow Msg
to localize to the cell wall.
We were ultimately able to target the expression of
recombinant Msg protein, excluding the serine-threoninerich region
and GPI anchor signal, to the yeast cell wall by using a
vector ( pBC542) that contains sequences encoding the
serinethreoninerich region of Epa1, an epithelial adhesin of C.
glabrata, and the carboxyl region of the cell wall protein (Cwp2) of
S. cerevisiae, which contains a yeast GPI-anchor sequence .
Cell wall expression also required a yeast signal peptide. The
cell wall localization was confirmed by immunofluorescence
analyses and by solubilization of the expressed Msg protein
with -1,3 glucanase treatment.
Most GPI cell wall proteins appear to have a long
serinethreoninerich region in the carboxy half of the protein that
allows extension of the protein to the cell surface . In C. glabrata,
proteins containing <100 amino acids in a serine-threoninerich
region are not expressed on the cell surface . Msg has a
serine-threoninerich region that contains only 59 amino acid
residues. To see whether a longer Pneumocystis-specific region
could function to target the cell wall in yeast, we replaced the
serine-threoninerich region of Epa1 with 4 tandem repeats of
the Msg serine-threoninerich region; however, this again was
insufficient to localize Msg to the cell wall. It thus appears that
the domains that presumably allow Msg to be expressed on the
surface of Pneumocystis organisms, including the signal peptide,
the serine-threoninerich region, and the carboxy terminal GPI
signal, are not processed by yeast in a similar manner.
The recombinant Msg protein that was expressed on the
yeast cell wall increased adhesion of yeast to A549 alveolar
epithelial cells, as measured by a binding assay using CFSE-labeled
yeast cells. Msg has previously been reported to mediate the
binding of Pneumocystis organisms to alveolar epithelial cells
[18, 20]. This adherence assay could be used to further
investigate the role of Msg in host-organism interactions.
Financial support. This work was supported by the Intramural
Research Program of the National Institutes of Health Clinical Center.
Potential conflicts of interest. All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential
Conflicts of Interest. Conflicts that the editors consider relevant to the
content of the manuscript have been disclosed.