Protein Expression and Purification 68 (2009) 183–189
Contents lists available at ScienceDirect
Protein Expression and Purification
journal homepage: www.elsevier.com/locate/yprep
Expression of bovine follicle-stimulating hormone subunits in a Hansenula
polymorpha expression system increases the secretion and bioactivity in vivo
Weidong Qian a,1, Yueyong Liu b,1, Chaozheng Zhang a, Zhendong Niu a, Haolei Song a, Bingsheng Qiu a,*
a
b
Center for Agricultural Biotechnology, Institute of Microbiology, Chinese Academy of Sciences, Beijing 100101, China
California Pacific Medical Center Research Institute, San Francisco, CA 94107, USA
a r t i c l e
i n f o
Article history:
Received 4 May 2009
and in revised form 21 July 2009
Available online 25 July 2009
a b s t r a c t
Bovine follicle-stimulating hormone (bFSH), a pituitary gonadotropin, is a heterodimer hormone that consists of a common a-subunit non-covalently associated with the hormone-specific b-subunit. Unfortunately, expression levels of recombinant bFSH or its subunits are invariably low. We report here the
secretory expression of biologically active bFSHa and bFSHb subunit in the methylotrophic yeast Hansenula
polymorpha. A slightly higher level of expression of recombinant bFSH subunits was achieved by using the
Sacchromyces cerevisiae-derived calnexin (ScCne1) as a chaperone in engineered H. polymorpha strains. The
preliminary data also suggested that bFSH subunits expressed in H. polymorpha appeared to be less-glycosylated. This isoform had been shown to be 80% increase in in vivo bioactivity compared with the hyperglycosylated Pichia pastoris-derived recombinant bFSHa/b. More sophisticated applications of bFSH would
profit from the assembled less-glycosylated heterodimer.
Ó 2009 Elsevier Inc. All rights reserved.
Bovine follicle-stimulating hormone (bFSH)2, which belongs to
the family of glycoprotein hormones produced either in the pituitary
or in the placenta, is a heterodimer composed of two non-covalently
associated a- and b-subunits. There are five intrachain disulfide
bridges in a-subunit and six in b-subunit, each containing two Nlinked carbohydrate moieties [1]. Purified FSH has been used to improve reproductive efficiency in both human and economically
important animals, but the results with the hormone or with pregnant mare serum gonadotropin, which contains intrinsic FSH and
luteinizing hormone activity, have been variable. The use of recombinant bovine FSH (bFSH), which is homologous to the species in which
it is applied most frequently, may improve superovulation results.
Nevertheless, bFSH has been difficult to purify and to obtain in sufficient quantities from bovine pituitaries [2]. Therefore, a standardized
source of recombinant bFSH would provide an attractive alternative
to produce large quantities of highly purified bFSH protein or its subunits to allow their use in many fertility related applications. Recombinant bFSH or its subunit has been expressed in the milk of
transgenic mice [2], in insect cells [3], in plants [4], in the milk of
transgenic rabbits [5], and in Pichia pastoris [6,7]. However, secretory
* Corresponding author.
E-mail address: qiubs@sun.im.ac.cn (B. Qiu).
1
Both authors contributed equally to this work.
2
Abbreviations used: bFSH, bovine follicle-stimulating hormone; Cne1, calnexin;
ScCne1, Saccharomyces cerevisiae calnexin; BiP, binding protein; MALDI-TOF MS,
matrix-assisted laser desorption/ionization time-of-flight mass spectrometry; hCG,
human Choriogonadotropin.
1046-5928/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.pep.2009.07.008
expression of recombinant bFSH is invariably low, which could be
attributed to inefficient gene expression due to a hairpin structure
at the 50 end of mature FSHb mRNA [8] and poor protein folding
capacity due to the complex arrangement of disulfide bonds of FSH
subunits [6]. In eukaryotes, protein folding and disulfide bond formation are well-known bottlenecks in the secreted heterologous proteins [9]. The specific folding of these proteins occurs in the
endoplasmic reticulum (ER), where several proteins are involved in
this process. Notable members of this group are the lectin chaperone
calnexin and the Hsp70-like chaperone BiP [10,11]. Calnexin is involved in the folding and quality control of nascent glycoproteins
[9]. Cne1, a homologue of calnexin in Saccharomyces cerevisiae, exhibits chaperone activities in a manner similar to mammalian calnexin,
and participates in protein folding, assembly, and post-translational
modification through an oligosaccharide moiety, Glc1Man9GlcNAc2
[10]. Modulation of these chaperones has been successfully used to
improve heterologous protein folding and secretion [12,13]. However, the effect of heterologous chaperone gene on the secretion of
heterologous protein has not been investigated in Hansenula
polymorpha.
The N-linked carbohydrate chains of natural pituitary FSH exhibit considerable variation in both size and structure [14–16]. Moreover, oligosaccharides modulated glycoprotein hormone efficacy
through an influence on hormone conformation [17]. Biologically
active recombinant bFSH has been expressed in different systems
[2–7], which appeared to be hyperglycosylated. However, to date,
there is no report about expression of less-glycosylated FSH protein
or its subunits and investigation of the effect this isoform on its
184
W. Qian et al. / Protein Expression and Purification 68 (2009) 183–189
biological activity. For the above reasons, high production at low
cost and less-glycosylation are essential when choosing engineered
organisms to generate and assemble the recombinant bFSH subunits. Consequently, we chose H. polymorpha expression system
for expression of bFSH as it offers some unique advantages over
other yeast expression systems. Hansenula polymorpha is a methylotrophic yeast that has been used as a host organism for industrial
scale production of eukaryotic proteins with clinical and therapeutic values, such as hirudin, hepatitis B surface antigen, aprotinin,
and human urokinase [18], because of its capability of faster growth
in simple defined media and the low antigenicity of proteins produced in H. polymorpha [19]. In addition, the prevention of hyperglycosylation in the H. polymorpha system constitutes an
additional advantage over other yeast expression systems such as
S. cerevisiae and P. pastoris for recombinant glycosylated proteins
production [20].
In this study, we report a slightly higher expression level of recombinant bFSH subunits by codon usage optimization scheme
and co-expression with S. cerevisiae-derived calnexin (ScCne1) in
H. polymorpha. Thus, it adds another genetic tool to the growing
toolbox for efficient secretion of heterologous proteins in H. polymorpha. Furthermore, the preliminary data suggested that bFSHa
and bFSHb expressed in H. polymorpha existed as a low molecular
weight isoform, which appeared to be less-glycosylated. This isoform exhibited 80% increase in in vivo bioactivity compared with
recombinant Pichia-derived hyperglycosylated bFSHa/b. This information can further be exploited for the structural-based functional
study of glycoprotein hormone.
of pGEX-6P-1. Expression and purification of GST-fused Cne1p proteins were performed as described by Xu et al. [22].
Construction of the vectors
Materials and methods
The native and codon-optimized genes of bFSH a- and b-subunits without their cognate signal peptide sequences were assembled from synthetic oligonucleotides by overlap-extension PCR,
respectively. A 6 His-tag was added at the C-terminus of each codon-optimized gene for further purification. As shown in Supplementary Fig. S1, the PCR amplified codon-optimized products are
315 nucleotides for bFSHa and 363 nucleotides for bFSHb. The synthetic genes bFSHa and bFSHb were assembled, respectively, from
oligonucleotides. Briefly, the oligonucleotides were synthesized
on a 50 nmol scale with no purification and dissolved in water to
a final concentration of 100 lM each. To assemble the oligonucleotides, PCR reactions were performed as described [23] with minor
modifications. To construct the specific pHMOXG-alpha-bFSHa
and pHMOXG-alpha-bFSHb, the synthesized bFSHa and bFSHb
genes were cloned into pHMOXG-alpha-A using the restriction sites
XhoI and NotI, respectively. Then, the pHMOXG-alpha-bFSHa vector was digested with BglII and BamHI, and ligated into BglII-digested pHMOXG-alpha-bFSHb. The resulting vector contains bFSH
subunits a and b, and was designated pHMOXG-alpha-bFSHabFSHb. To construct the specific pHFMDZ-ScCne1, the chaperone
ScCne1 fragment was amplified from genomic DNA of S. cerevisiae
with primers F2 (50 -CCATGGTGAAGTTTTCTGCGTATTTA-30 , the NcoI
site is underlined) and R2 (50 -TCTAGACTATGTAAATACTACACA-30 ,
the XbaI site is underlined). The resulting ScCne1 fragment was
digested by appropriate restriction enzymes and subcloned into
similarly digested pHFMDZ-A, yielding pHFMDZ-ScCne1.
Strains, plasmids and reagents
Yeast transformation and selection of recombinant clones
Hansenula polymorpha NCYC495 (leu 1.1) was a gift from Dr. Jan
A.K.W. Kiel, University of Groningen, the Netherlands, and served
as the standard host for protein expression. pHMOXG-alpha-A, a
G418-selectable plasmid, contains a methanol oxidase (MOX)
promoter from H. polymorpha fused to the a-mating factor from
S. cerevisiae for directing the protein to the secretory pathway.
pHFMDHZ-A, a zeocin-selectable plasmid, contains a formate
dehydrogenase (FMD) promoter from H. polymorpha. The two plasmids were constructed in our laboratory [21] and used for cloning
and expression in H. polymorpha. Methanol oxidase (MOX) and formate dehydrogenase (FMD) promoters derived from key genes of
the methanol utilization pathway are strong inducible promoters
[19]. Strong induction does not exclusively depend on the presence
of toxic methanol but can be elicited by limited addition of glycerol
or glucose [19]. Saccharomyces cerevisiae strain AH109 was purchased from Clontech. Escherichia coli DH5a and BL21 (DE3) (Novagen), kept in our laboratory, were used as a cloning strain and an
expression strain, respectively. Taq DNA polymerase and endonucleases were purchased from Takara Biotechnology Co., Ltd. (Dalian, China). All other chemicals were of analytical grade and
obtained from the local commercial resources.
To express separately the bFSHa and bFSHb genes, H. polymorpha competent cells were transformed with the DraI-linearized
pHMOXG-alpha-bFSHa and pHMOXG-alpha-bFSHb by electroporation [24], respectively, and screened at 37 °C on YPD selective
plates (1% (w/v) difco yeast extract, 2% (w/v) bacto-tryptone and
2% (w/v) dextrose) supplemented with 200 lg/ml G418. Similarly,
to co-express the bFSHa and bFSHb genes, H. polymorpha competent cells were transformed with the BglII-linearied pHMOXG-alpha-bFSHa-bFSHb by electroporation. Selection of transformants
containing a single copy of bFSHa or bFSHb gene was carried out
by PCR amplification of the promoter fragment described by Brito
et al. [25]. The positive transformants were designated H. polymorpha/bFSHa, H. polymorpha/bFSHb and H. polymorpha/bFSHa/b,
respectively. To co-express ScCne1 and either bFSH a- or b-subunit, H. polymorpha/bFSHa and H. polymorpha/bFSHb competent
cells were transformed, respectively, with the ApaI-linearized
pHFMDZ-ScCne1 by electroporation and screened at 37 °C on
YPD selective plates containing 100 lg/ml zeocin. The selected
co-expression recombinants were validated by PCR. The resulting
positive transformants were designated H. polymorpha/bFSHa/
ScCne1 and H. polymorpha/bFSHb/ScCne1, accordingly.
Expression of recombinant bFSHa and bFSHb in shaking flask
Polyclonal antibody production
The polyclonal antisera against calnexin were obtained by
injecting rabbits with GST-fused Cne1p fusion proteins expressed
in E. coli, according to the standard immunization protocol. The
chaperone ScCne1 fragment was amplified with primers F1
(50 -GAATCCATGAAGTTTTCTGCGTATTTA-30 , the EcoRI site is underlined) and R1 (50 -GCGGCCGCCTATGTAAATACTACACA-30 , the NotII
site is underlined) and cloned between the EcoRI and NotI sites
Pre-culture from a single colony was grown with shaking at
37 °C for 12 h in 3 ml of YPD selective medium in a test tube supplemented with 100 lg/ml zeocin. The first pre-culture was collected by centrifugation (2000g, 15 min) and resuspended in
30 ml of YPD medium in a 500 ml flask and inoculated at 37 °C
for 24 h (to OD600 of 2–6). The cells were harvested and resuspended in 300 ml of YPD medium in a 2-L flask to an initial
OD600 of approximately 1.0. For induction, methanol (100%) was
W. Qian et al. / Protein Expression and Purification 68 (2009) 183–189
added every 12 h to a final concentration of 0.5% (v/v) throughout
the induction phase (3 days). The supernatants of the cultures containing bFSHa and bFSHb were harvested by centrifugation (4000g,
15 min).
Purification of recombinant bFSHa and bFSHb
The supernatant containing bFSHa protein was centrifuged,
concentrated and dialyzed against binding buffer (50 mM Na2HPO4, 300 mM NaCl, pH 7.5) by tangential flow filtration using a
5,000 MWCO PES membrane (Vivaflow 200, Sartorius Stedim Biotech, Goettingen, Germany) and later loaded onto Ni–NTA resin
column (BD Biosciences Clontech) equilibrated with the same buffer. The bound proteins were eluted using a gravity-flow column
starting with extraction/wash buffer (50 mM sodium phosphate,
300 mM NaCl, and 20 mM imidazole, pH 7.0) to remove all unbound impurities and ending with the same buffer containing
150 mM imidazole. Fractions were pooled and desalted with a
desalting column (Sephadex G-25, Amersham Pharmacia) in Tris/
HCl buffer (20 mM Tris, pH 7.5) at 4 ml/min. A similar protocol
was followed for purification of bFSHb protein from the culture
supernatant. Protein concentration was determined by the method
of Bradford using BSA as a standard.
SDS–PAGE and Western blotting
For SDS–PAGE analysis, the samples were mixed with 5 reducing sample buffer and separated by 15% SDS–PAGE gels. The gels
were stained with Coomassie blue R-250. For Western blot analysis, the gels were allowed to equilibrate for 5 min in transfer buffer
(25 mM Tris, 192 mM glycine and 20% (v/v) methanol). Proteins
were then transferred onto Hybond-P PVDF membranes (Amersham Pharmacia) at 250 mA for 3 h using a minitransblot electrophoresis cell (Bio-Rad Laboratories, Richmond, CA). The
membranes were blocked with 2.5% casein in PBST (80 mM Na2HPO4, 20 mM NaH2PO4, 100 mM NaCl, and 0.1% Tween 20) at 37 °C
for 2 h. The membranes were incubated with either the rabbit
polyclonal antibody raised against the bovine FSH (AbD Serotec)
(diluted 1:3000 in PBST with 1% BSA) or the rabbit polyclonal antibody raised against Cne1p (diluted 1:1000 in PBST with 1% BSA) at
37 °C for 1 h followed by three consecutive washes in PBS containing 0.1% Tween 20 (15 min/wash). The membranes were further
incubated with HRP-labeled goat anti-rabbit IgG (diluted 1:4000
in PBST with 1% BSA) at 37 °C for 1 h followed by three washes. Finally, the membranes were revealed with ECL Western blotting reagent (Amersham Biosciences). And the results were analyzed with
software Bandscan 5.0.
Quantitative ELISA
The amount of intracellular or extracellular recombinant
bFSHa, bFSHb and bFSHa/b proteins was calculated by quantitative
ELISA using pituitary porcine FSH (Sigma) as a standard as described by Huo et al. [7].
Mass spectrometry
For matrix-assisted laser desorption/ionization time-of-flight
mass spectrometry (MALDI-TOF MS) analysis, the purified proteins
by Ni–NTA affinity chromatography were loaded, respectively,
onto a Hi-load Superdex 75 HR 10/30 column equilibrated with sodium phosphate buffer (50 mM sodium phosphate, 100 mM NaCl,
pH 7.0) using an AKTA Explorer FPLC system (Amersham Pharmacia). The fractions under the peak were collected and dialyzed
against Tris/HCl buffer (20 mM Tris, 0.02% NaN3, pH 8.0), and then
185
analyzed by Voyager-DE-STR MALDI/TOF mass spectrometer
(PerSeptive BioSystems) as described previously [26].
Deglycosylation analysis
The recombinant Pichia-derived bFSH was prepared and purified as described before [7]. The purified recombinant Pichia-derived bFSH, Hansenula-derived bFSHa and bFSHb subunits were
digested, respectively, with PNGaseF (New England Biolabs) following the manufacturer’s instructions. After digestion, the proteins were precipitated using acetone and subjected to SDS–PAGE
analysis. As a control, the same procedure was performed in the
absence of PNGaseF.
In vivo bioassay of recombinant bFSHa and bFSHb
In vivo bioactivity was assessed by measuring ovarian weight
gain 3 days post-injection [27]. In brief, 50 female mice (21-dold) were randomized into five groups (10 mice in each group),
and inoculated subcutaneously 3 times by once-daily injections
of: group 1, the recombinant bFSHa and bFSHb with a 1:1 ratio (total dose, 0.4 mg) and human Choriogonadotropin (hCG) (total dose,
40 IU); group 2, the recombinant Pichia-derived bovine FSH (total
dose, 0.4 mg) and hCG (total dose, 40 IU); group 3, pituitary porcine FSH (total dose, 4 IU) and hCG (total dose, 40 IU); group 4,
hCG only as a control (total dose, 40 IU); group 5, PBS only as a negative control, respectively, according to the method described by
Steelman. At 24 h after the last injection, all mice were administered diethylether anesthesia and then exsanguinated. The ovarian
were removed from each mouse and weighed. The significance of
the differences between group means was determined by oneway analysis of variance (ANOVA) followed by Student–Newman–Keuls test with the level of significant difference at p < 0.05.
Results
Expression of native bFSHa/b, bFSHa- and b- subunit genes
At first, H. polymorpha cells were co-transformed with the
bFSHa and bFSHb expression constructs to obtain cells expressing
heterodimeric bFSH. However, ELISA analysis showed that the
extracellular and intracellular levels of the best performing H. polymorpha strain in producing heterodimeric bFSH were as low as 0.3
and 0.7 mg/L culture, at the shake culture level, respectively. Then,
bFSHa- and b-subunits were separately produced by the H. polymorpha cells transformed with the vectors containing each of
bFSHa- and b-subunit genes. Unfortunately, ELISA analysis showed
that the yields of extracellular and intracellular bFSHa protein correspond to 1.2 and 1.5 mg/L culture, while those of bFSHb correspond to 0.3 and 0.5 mg/L culture, respectively.
Expression of codon-optimized bFSHa/b, bFSHa- and b- subunit genes
Due to the hairpin structure present in the first 12 codons of mature FSHb mRNA and the resultant downregulation of gene expression, it is difficult to optimally express FSHb gene in heterologous
systems. In addition, expression levels of the genes are frequently
correlated with codon usage in yeast expression systems [28].
Therefore, to improve expression levels of bFSH subunits in H. polymorpha, we replaced some key codons with those most frequently
used in H. polymorpha genes without altering the protein sequences
according to synonymous codons preferred by H. polymorpha based
on multivariate statistical analysis software (http://bioweb.pasteur.fr/seqanal/interfaces/codonw.html). In order to analyze the
performance of the co-expression and independent expression of
186
W. Qian et al. / Protein Expression and Purification 68 (2009) 183–189
codon-optimized bFSHa- and b- subunits genes as described above,
similar experiments were carried out. Accordingly, H. polymorpha
cells were transformed with the vectors containing both codonoptimized bFSHa- and b- subunits. ELISA analysis showed that
the levels of secretion of extracellular and intracellular heterodimeric bFSHa/b correspond to 0.5 and 1.1 mg/L culture, respectively.
However, of the two subunits expressed independently, two subunits were successfully produced in abundance as a single subunit
as compared to co-expression of two subunits (see below). Expression data at 72 h for one clone of each of the two synthetic genes are
shown in Fig. 1. The H. polymorpha/bFSHa and H. polymorpha/bFSHb
clones expressed proteins that were detectable by Western blotting
(Fig. 1) or ELISA (data not shown). However, only bFSHa protein
was efficiently secreted into the culture medium (Fig. 1A, lane 2),
whereas bFSHb under the same signal peptide was not (Fig. 1B, lane
2). This indicates that other components such as correct post-translational modifications might be required for secretion of bFSHb.
Improved secretion of bFSHa and bFSHb subunits by co-expression of
heterologous chaperone
To promote secretion of bFSHb and bFSHa, the representative
engineered H. polymorpha competent cells harboring a single copy
of bFSHb or bFSHa gene were transformed with the linearized
pHFMDZ-ScCne1. The resultant representative positive clones, designated H. polymorpha/bFSHb/ScCne1 and H. polymorpha/bFSHa/
ScCne1, respectively, along with the representative clones H. polymorpha/bFSHb and H. polymorpha/bFSHa, were grown in a shake
flask to compare their expression levels. Co-expression of ScCne1
increased secretion of bFSHb and bFSHa proteins. ELISA analysis
showed that the H. polymorpha/bFSHb/ScCne1 clone secreted
bFSHb protein into culture medium at levels (4.7 mg/L culture)
that were up to 6-fold higher than those (0.8 mg/L culture) secreted by H. polymorpha/bFSHb. Similarly, the yield of secreted
bFSHa reached approximately to 6.2 mg/L culture in the H. polymorpha/bFSHa/ScCne1 clone, which was an increase of 1.5-fold
compared with H. polymorpha/bFSHa (4.2 mg/L culture). To obtain
further insight into the secretion levels of bFSH subunit by ScCne1,
the extracellular and intracellular levels were investigated by Western blotting using anti-bFSH antibodies. As shown in Fig. 1B,
bFSHb proteins secreted by H. polymorpha/bFSHb stained at only
one-sixth the intensity of those secreted by H. polymorpha/
bFSHb/ScCne1 (Fig. 1B; lanes 2 and 3). Similarly, as shown in
Fig. 1A, the secretion of bFSHa proteins was slightly increased in
strain H. polymorpha/bFSHa as compared to strain H. polymorpha/
bFSHa/ScCne1 (Fig. 1A; lanes 2 and 3). In contrast, the amounts
of intracellular accumulated bFSHa and bFSHb proteins were reduced in engineered strains with co-expression of ScCne1 as compared to the parental strains, respectively (Fig. 1A and B; lanes 5
and 6). In addition, as shown in Fig. 1D and E, Cne1 protein was
overexpressed intracellularly in engineered strains. These results
showed that heterologous chaperone ScCne1 might help to reduce
intracellular protein aggregation possibly due to incorrect folding
of protein, thereby increasing the secretion of a protein of interest.
Purification of bFSHa and bFSHb subunits
To facilitate the purification of recombinant bFSHa and bFSHb, a
6 His-tag coding sequence was added to the C-terminus of each
of the subunits. This allowed us to use only one step to purify
the subunits using Ni–NTA affinity chromatography. As is shown
in Fig. 2A, the purity of the preparations was estimated to be over
95% for both bFSHa and bFSHb proteins. Table 1 summarizes the
purification scheme of the two proteins.
Characterization of the recombinant bFSHa and bFSHb subunits
To determine of the molecular weights of intracellular and
extracellular bFSH subunits, samples from the soluble cytosolic
fraction and culture supernatant of H. polymorpha/bFSHb and H.
polymorpha/bFSHa, respectively, were subjected to Western blotting. As is shown in Fig. 1, the molecular weights of the intracellular bFSHa and bFSHb subunits were larger than those of the
extracellular bFSHa and bFSHb subunits, respectively, which might
be attributed to insufficient cleavage of the a-mating factor signal
peptide attached to each subunit. To further determine of the
molecular weights of extracellular bFSH subunits, purified proteins
were analyzed by SDS–PAGE and MALDI-TOF. Mass spectral analysis showed that the molecular weights of secreted recombinant
bFSHa and bFSHb subunits were approximately 13.2 (Fig. 2B)
and 15.3 kDa (Fig. 2C), respectively, which were close to the predicted molecular mass estimated by SDS–PAGE (Fig. 2A) and
slightly larger than the theoretical size of the unglycosylated proteins (12.1 and 13.7 kDa, respectively). To investigate glycosylation
Fig. 1. Western blotting of the recombinant proteins. The samples from the culture supernatant and the soluble cytosolic fraction of engineered stranis were analyzed using
Western blotting in combination with anti-bovine FSH polyclonal antibodies against bFSHa and bFSHb subunits. The results were shown in (A) and (B), respectively. In
parallel, the samples from the soluble cytosolic fraction of engineered strains were analyzed using Western blotting in combination with anti-Cne1p polyclonal antibodies.
The results were shown in (D) and (E), respectively. (A and D) Lanes 2 and 3, the supernatants of H. polymorpha/bFSHa and H. polymorpha/bFSHa/ScCne1, respectively; lanes 5
and 6, the soluble cytosolic fractions of H. polymorpha/bFSHa and H. polymorpha/bFSHa/ScCne1, respectively. (B and E) Lanes 2 and 3, the supernatants of H. polymorpha/
bFSHb and H. polymorpha/bFSHb/ScCne1, respectively; lanes 5 and 6, the soluble cytosolic fractions of H. polymorpha/bFSHb and H. polymorpha/bFSHb/ScCne1, respectively.
Lanes 1 and 4, the culture supernatant and intracellular accumulated fraction of yeast transformed with an expression vector without any inserts (A and B), respectively. (C)
Two hundred nanograms of commercial pituitary porcine FSH standard for reference.
W. Qian et al. / Protein Expression and Purification 68 (2009) 183–189
187
Fig. 2. Identification of the purified bFSHa and bFSHb subunits by SDS–PAGE and MALDI-TOF MS. (A) The purified proteins by gel-filtration chromatography were separated
by a 15% (v/v) SDS–PAGE gel and stained with Coomassie blue R-250. Lane M, protein molecular weight markers; lane 1, bFSHa subunits; lane 2, bFSHb subunits. (B and C)
The molecular weights of bFSHa and bFSHb were 13,275 Da (B) and 15,372 Da (C), respectively.
status of the recombinant bFSH subunits, we employed PNGaseF
assay as described above. Slight size difference between the
digested and undigested bFSHa, bFSHb subunits was observed
(Fig. 3). By contrast, the size difference of bFSH heterodimers from
Table 1
Purification yield of recombinant bFSHa and bFSHb in shake flask culture.
Protein
Wet weight
cells (g/L)
Total protein before
purification (mg/L)a
Ni–NTA
resin
Sephadex
G-25
bFSHa
bFSHb
47.5
43.2
6.2
4.7
5.1
3.2
3.8
2.3
a
Total bFSHa or bFSHb was estimated from supernatants of 1 L culture by
quantitative ELISA with a standard curve of known amounts of pituitary porcine
FSH. The bFSHa or bFSHb purified from Ni–NTA resin column and Sephadex G-25
desalting column was estimated by Bradford method. The data represent the
average value of triplicate results.
P. pastoris was clearly revealed after being treated with PNGaseF
(Fig. 3). This was due to the hyperglycosylation and larger molecular weight of the heterodimer. These results demonstrated that
the extent of glycosylation of recombinant bFSH subunits in H.
polymorpha was significantly reduced.
In vivo bioactivity of recombinant bFSHa and bFSHb subunits
To evaluate in vivo bioactivity of the recombinant bFSHa and
bFSHb subunits expressed in H. polymorpha, the rat ovarian weight
bioassay was performed. The group treated with the mixture of
Hansenula-derived bFSH subunits had a mean ± SD ovarian weight
3 days post-injection (452.6 ± 6.7) significantly greater than groups
treated with Pichia-derived bFSH and pituitary porcine FSH
(253.2 ± 5.6 and 251.7 ± 7.3 mg, respectively, P < 0.05) (Fig. 4). This
confirms that the bioactivity of the recombinant bFSH subunits
Fig. 3. Deglycosylation analysis of recombinant proteins. The purified Hansenula-derived bFSHa-, b-subunits and Pichia-derived bFSHa/b heterodimer were incubated,
respectively, with the manufacture supplied denaturing buffer at 100 °C for 15 min. After the samples were cooled, the manufacturer supplied G7 reaction buffer and NP-40
(final concentration 1% (v/v)) were added. Then the samples were harvested and subjected to SDS–PAGE analysis after at 37 °C for 24 h digestion with or without PNGaseF.
188
W. Qian et al. / Protein Expression and Purification 68 (2009) 183–189
Fig. 4. In vivo bioactivity of recombinant less-glycosylated bFSHa and bFSHb. The
in vivo bioactivity was determined by the ovarian augmentation bioassay as
described in the Materials and methods section. The results were expressed as the
means ± standard deviation from representative experiments in the size of ovarian
induced by the mixture of Hansenula-derived less-glycosylated bFSHa and bFSHb,
Pichia-derived bFSH and pituitary porcine FSH with hCG, hCG only, and PBS only,
respectively. Similar results were obtained in three independent experiments.
was possibly affected by some factors such as the decreased glycosylation and more efficient folding enhanced by the recombinant
calnexin.
Discussion
In this study, we investigated the possibilities of improving
expression and in vivo bioactivity of recombinant bFSH subunits
using the methylotrophic yeast H. polymorpha expression system.
Firstly, by using codon usage optimization scheme, we have considerably improved the amounts of expression of bFSHa and
bFSHb. However, while altering the codons can lead to an increase
in levels of intercellular expression, it fails to increase the secretion
of protein probably due to the incorrect protein folding. Hence, in
order to help these post-translational manipulations, we next explored the feasibility of using a chaperone co-expression system.
The combination of Western blot analysis and quantitative ELISA
revealed that the secretion of bFSHa and bFSHb subunits was enhanced using ScCne1 as a chaperone. The mechanism by which
ScCne1 enhancing the secretion of recombinant bFSH subunits
might be consistent with the multifunctions of this protein, which
could act as a N-glycosylation-binding lectin or an ER chaperone in
yeast [10]. When acting as a lectin, the ScCne1 interacts with Nlinked Glc1Man9GlcNac2 oligosaccharides of glycoproteins until
correct conformation is formed [29]. Both the bFSHa and bFSHb
subunits are glycoproteins carrying two N-linked carbohydrate
moieties. Interestingly, such complex formation between calnexin
and unglycosylated substrates has been reported in Schizosaccharomyces pombe [30]. This indicates that calnexin can perform the
general chaperone functions by interacting directly with the polypeptide backbone, in addition to binding to carbohydrate moieties.
Thus, acting as lectin and the chaperone activity of ScCne1 might
account for the increased secretion of the recombinant bFSHa
and bFSHb subunits.
It is also well-known that the glycosylation of bFSH subunit and
its heterodimerization are essential for its specific biological activity [15]. As previously demonstrated for hCG in animal cells, ungly-
cosylation of the subunit leads to the inhibition of subunit
assembly [17]. The Hansenula-derived bFSH subunits showed a
higher in vivo bioactivity compared with Pichia-derived bFSH, indicating that these subunits might be properly less-glycosylated,
folded and assembled as functional heterodimers. The glycosylation status could cause subsequent bioactivity changes by regulation the conformation of proteins [31]. The difference in
glycosylation observed in this study might be attributed to specific
regulation of protein glycosylation in H. polymorpha [20]. Actually,
further studies of the decreased glycosylation in H. polymorpha
were also currently under evaluation.
In addition, as shown in Fig. 4, in the bioassay recombinant
hexahistidyl tagged bFSH subunits resulted in a slightly higher
augmentation of rat ovaries compared to the pituitary porcine
FSH, suggesting that 6 His-tag fused to the C terminus of bFSH
subunit had no significant effect on the normal folding and function of the recombinant protein. These results are consistent with
those reported in previous recombinant hexahistidyl tagged protein studies [32,33]. However, potential clinical use of the hexahistidyl tagged bFSH subunits still needs further assessment. In
our opinion, we should not rule out the possibility that the hexahistidyl tagged bFSH proteins might be harmful to the animals,
which needs further investigation.
The expression levels here are a little higher than those that
have been reported for transgenic mice [2], insect cells [3], plants
[4] and P. pastoris [6,7]. Meanwhile, recombinant Hansenula-derived bFSH showed a 1.8-fold increase in in vivo bioactivity compared with recombinant Pichia-derived hyperglycosylated bFSH.
In addition, the yeast expression system can be scaled up better
than higher eukaryotic cell systems [34]. Considering the production of recombinant proteins could increase by up to 5–20 times
in high-cell-density fermentation instead of in flasks [19], we believe that these amounts can be scaled up by further optimization
of fermentation to produce sufficient amount of protein for clinical
trials. Taken together, this makes H. polymorpha a better candidate
to produce large quantities of recombinant bFSH proteins for clinical and veterinary purposes.
In summary, a large quantity of bFSH subunits was produced
using a H. polymorpha expression system by codon usage optimization scheme and genetic tools to mediate the ER folding environment. In addition, the present study is the first report evaluating
bioactivity of less-glycosylated recombinant bFSH subunits, which
exhibited higher in vivo bioactivity compared with hyperglycosylated. Therefore, this expression system would be very useful for
production and investigation of the less-glycosylated glycoprotein
hormone or its subunits, permitting further examination of the effect of glycosylation status on hormone conformation and bioactivity in vivo or in vitro.
Acknowledgments
The authors are indebted to Dr. Xiangdong Huo (Institute of
Microbiology, Xinjiang Academy of Agricultural Science) for his
assistance in preparing Pichia-derived bFSH. The authors would
also like to thank Dr. Houhui Song (Qingdao Institute of BioEnergy
and Bioprocess Technology, Chinese Academy of Sciences) and Dr.
Jinyong Wang (University of Wisconsin-Madison) for their critical
reading of the paper. This work was supported by the innovation
funds of the Chinese Academy of Sciences.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.pep.2009.07.008.
W. Qian et al. / Protein Expression and Purification 68 (2009) 183–189
References
[1] J.G. Pierce, T.F. Parsons, Glycoprotein hormones: structure and function, Annu.
Rev. Biochem. 50 (1981) 465–495.
[2] N.M. Greenberg, J.W. Anderson, A.J. Hsueh, K. Nishimori, J.J. Reeves, D.M.
deAvila, D.N. Ward, J.M. Rosen, Expression of biologically active heterodimeric
bovine follicle-stimulating hormone in milk of transgenic mice, Proc. Natl.
Acad. Sci. USA 88 (1991) 8327–8331.
[3] D.F. van de Wiel, P.A. van Rijn, R.H. Meloen, R.J. Moormann, High-level
expression of biologically active recombinant bovine follicle stimulating
hormone in a baculovirus system, J. Mol. Endocrinol. 20 (1998) 83–98.
[4] D. Dirnberger, H. Steinkellner, L. Abdennebi, J.J. Remy, D. van de Wiel, Secretion
of biologically active glycoforms of bovine follicle stimulating hormone in
plants, Eur. J. Biochem. 268 (2001) 4570–4579.
[5] S. Coulibaly, U. Besenfelder, I. Miller, N. Zinovieva, C. Lassnig, T. Kotler, J.L.
Jameson, M. Gemeiner, M. Muller, G. Brem, Expression and characterization of
functional recombinant bovine follicle-stimulating hormone (boFSHalpha/
beta) produced in the milk of transgenic rabbits, Mol. Reprod. Dev. 63 (2002)
300–308.
[6] M. Samaddar, J.F. Catterall, R.R. Dighe, Expression of biologically active beta
subunit of bovine follicle-stimulating hormone in the methylotrophic yeast
Pichia pastoris, Protein Expr. Purif. 10 (1997) 345–355.
[7] X. Huo, Y. Liu, X. Wang, P. Ouyang, Z. Niu, Y. Shi, B. Qiu, Co-expression of
human protein disulfide isomerase (hPDI) enhances secretion of bovine
follicle-stimulating hormone (bFSH) in Pichia pastoris, Protein Expr. Purif. 54
(2007) 234–239.
[8] M. Samaddar, P.S. Babu, J.F. Catterall, R.R. Dighe, Identification of an
attenuating region in the bovine follicle-stimulating hormone beta subunit
mRNA that decreases its expression in E. coli, Gene 228 (1999) 253–260.
[9] L. Ellgaard, M. Molinari, A. Helenius, Setting the standards: quality control in
the secretory pathway, Science 286 (1999) 1882–1888.
[10] F. Parlati, M. Dominguez, J.J. Bergeron, D.Y. Thomas, Saccharomyces cerevisiae
CNE1 encodes an endoplasmic reticulum (ER) membrane protein with
sequence similarity to calnexin and calreticulin and functions as a
constituent of the ER quality control apparatus, J. Biol. Chem. 270 (1995)
244–253.
[11] I.G. Haas, M. Wabl, Immunoglobulin heavy chain binding protein, Nature 306
(1983) 387–389.
[12] J. Klabunde, S. Kleebank, M. Piontek, C.P. Hollenberg, S. Hellwig, A. Degelmann,
Increase of calnexin gene dosage boosts the secretion of heterologous proteins
by Hansenula polymorpha, FEMS Yeast Res. 7 (2007) 1168–1180.
[13] T.A. Hsu, M.J. Betenbaugh, Coexpression of molecular chaperone BiP improves
immunoglobulin solubility and IgG secretion from Trichoplusia ni insect cells,
Biotechnol. Prog. 13 (1997) 96–104.
[14] J.U. Baenziger, E.D. Green, Pituitary glycoprotein hormone oligosaccharides:
structure, synthesis and function of the asparagine-linked oligosaccharides on
lutropin, follitropin and thyrotropin, Biochim. Biophys. Acta 947 (1988) 287–
306.
[15] G.R. Bousfield, V.Y. Butnev, W.J. Walton, V.T. Nguyen, J. Huneidi, V. Singh, V.S.
Kolli, D.J. Harvey, N.E. Rance, All-or-none N-glycosylation in primate folliclestimulating hormone beta-subunits, Mol. Cell Endocrinol. 260–262 (2007) 40–
48.
[16] W.J. Walton, V.T. Nguyen, V.Y. Butnev, V. Singh, W.T. Moore, G.R. Bousfield,
Characterization of human FSH isoforms reveals a nonglycosylated betasubunit in addition to the conventional glycosylated beta-subunit, J. Clin.
Endocrinol. Metab. 86 (2001) 3675–3685.
189
[17] S.W. Trout, Y. Han, R.V. Myers, M.P. Bernard, W.R. Moyle, Deglycosylation of a
bifunctional lutropin–follitropin agonist reduced its follitropin activity more
than its lutropin activity, Fertil. Steril. 72 (1999) 1093–1099.
[18] G. Gellissen, Heterologous protein production in methylotrophic yeasts, Appl.
Microbiol. Biotechnol. 54 (2000) 741–750.
[19] G. Gellissen, G. Kunze, C. Gaillardin, J.M. Cregg, E. Berardi, M. Veenhuis, I. van
der Klei, New yeast expression platforms based on methylotrophic Hansenula
polymorpha and Pichia pastoris and on dimorphic Arxula adeninivorans and
Yarrowia lipolytica—a comparison, FEMS Yeast Res. 5 (2005) 1079–1096.
[20] S.Y. Kim, J.H. Sohn, Y.R. Pyun, E.S. Choi, Variations in protein glycosylation in
Hansenula polymorpha depending on cell culture stage, J. Microbiol. Biotechnol.
17 (2007) 1949–1954.
[21] H. Song, Y. Li, W. Fang, Y. Geng, X. Wang, M. Wang, B. Qiu, Development of a set
of expression vectors in Hansenula polymorpha, Biotechnol. Lett. 25 (2003)
1999–2006.
[22] X. Xu, H. Azakami, A. Kato, P-domain and lectin site are involved in the
chaperone function of Saccharomyces cerevisiae calnexin homologue, FEBS Lett.
570 (2004) 155–160.
[23] C. Withers-Martinez, E.P. Carpenter, F. Hackett, B. Ely, M. Sajid, M. Grainger, M.J.
Blackman, PCR-based gene synthesis as an efficient approach for expression of
the A+T-rich malaria genome, Protein Eng. 12 (1999) 1113–1120.
[24] K.N. Faber, P. Haima, W. Harder, M. Veenhuis, G. Ab, Highly-efficient
electrotransformation of the yeast Hansenula polymorpha, Curr. Genet. 25
(1994) 305–310.
[25] N. Brito, M.D. Perez, G. Perdomo, C. Gonza lez, P. Garcia-Lugo, J.M. Siverio, A set
of Hansenula polymorpha integrative vectors to construct lacZ fusions, Appl.
Microbiol. Biotechnol. 53 (1999) 23–29.
[26] D.P. Barr, M.R. Gunther, L.J. Deterding, K.B. Tomer, R.P. Mason, ESR spintrapping of a protein-derived tyrosyl radical from the reaction of cytochrome c
with hydrogen peroxide, J. Biol. Chem. 271 (1996) 15498–15503.
[27] S.L. Steelman, F.M. Pohley, Assay of the follicle stimulating hormone based on
the augmentation with human chorionic gonadotropin, Endocrinology 53
(1953) 604–616.
[28] G. Sinclair, F.Y. Choy, Synonymous codon usage bias and the expression of
human glucocerebrosidase in the methylotrophic yeast, Pichia pastoris, Protein
Expr. Purif. 26 (2002) 96–105.
[29] F.E. Ware, A. Vassilakos, P.A. Peterson, M.R. Jackson, M.A. Lehrman, D.B.
Williams, The molecular chaperone calnexin binds Glc1Man9GlcNAc2
oligosaccharide as an initial step in recognizing unfolded glycoproteins, J.
Biol. Chem. 270 (1995) 4697–4704.
[30] H. Beaulieu, A. Elagoz, P. Crine, L.A. Rokeach, Interaction of mammalian
neprilysin with binding protein and calnexin in Schizosaccharomyces pombe,
Biochem. J. 340 (Pt. 3) (1999) 813–819.
[31] M.R. Flack, J. Froehlich, A.P. Bennet, J. Anasti, B.C. Nisula, Site-directed
mutagenesis defines the individual roles of the glycosylation sites on
follicle-stimulating hormone, J. Biol. Chem. 269 (1994) 14015–14020.
[32] B.S. Wojczyk, M. Czerwinski, M.M. Stwora-Wojczyk, D.L. Siegel, W.R. Abrams,
W.H. Wunner, S.L. Spitalnik, Purification of a secreted form of recombinant
rabies virus glycoprotein: comparison of two affinity tags, Protein Expr. Purif. 7
(1996) 183–193.
[33] M. Mack, M. Wannemacher, B. Hobl, P. Pietschmann, B. Hock, Comparison of
two expression platforms in respect to protein yield and quality: Pichia pastoris
versus Pichia angusta, Protein Expr. Purif. 66 (2009) 165–171.
[34] G.P. Cereghino, J.L. Cereghino, C. Ilgen, J.M. Cregg, Production of recombinant
proteins in fermenter cultures of the yeast Pichia pastoris, Curr. Opin.
Biotechnol. 13 (2002) 329–332.