Optimization of the wheat puroindoline-a production inPichia pastoris
N. Issaly, O. Solsona, P. Joudrier1, M.F. Gautier1, G. Moulin and H. BozeUFR Microbiologie Industrielle et de GeÂneÂtique des Microorganismes, ENSA.M-INRA, Montpellier and 1UniteÂ
de Biochimie et Biologie MoleÂculaire des CeÂreÂales, INRA Montpellier, France
494/08/00: received 2 August 2000, revised 9 November 2000 and accepted 14 November 2000
N. ISSALY, O. SOLSONA, P. JOUDRIER, M.F . GAUTIER, G. MOULIN AND H. BOZE. 2001.
Aims: A recombinant puroindoline-a (rPIN-a) was produced using the methylotrophic yeast
Pichia pastoris.
Methods and Results: In fed-batch culture, the production of rPIN-a decreased after 24 h of
methanol induction. Most of the rPIN-a was not soluble in the culture medium remaining
bound to the cell walls. Soluble and membrane-bound rPIN-a were quanti®ed by ELISA after
Triton X-114 phase partitioning. In order to improve the production of rPIN-a, the in¯uence
of pH, speci®c growth rate and the addition of TX-114 was tested on two independent
continuous cultures. The production of rPIN-a was improved when continuous culture was
carried out at 29°C under acid conditions (pH 5) with a low dilution rate (D � 0á025 h)1). The
addition of 0á01% TX-114 to the medium inverted the ratio between the secreted and the
membrane-bound rPIN-a.
Conclusions: When a continuous culture was carried out under optimized conditions, the
rPIN-a production yield was increased 10-fold to 14 mg l)1 and 80% of the rPIN-a was
soluble.
Signi®cance and Impact of the Study: This study would be helpful to optimize the
expression of other membrane-bound proteins in P. pastoris.
INTRODUCTION
Puroindolines are basic cystine-rich proteins of about
13 kDa that have been isolated from wheat endosperm by
the Triton X-114 phase partitioning method (Blochet et al.1993). They belong to cystine-rich proteins of wheat seed as
purothionins, lipid transfer proteins and CM proteins. Two
cDNA clones encoding puroindoline-a (PIN-a) and puro-
indoline-b (PIN-b) have been characterized. They are 55%
similar and contain 10 cystine residues involved in ®ve
disulphide bonds which maintain their folded conforma-
tions. Furthermore, these proteins exhibit a unique trypto-
phan-rich domain which is partially truncated in PIN-b
(Gautier et al. 1994). Puroindolines are related to basic
friabilins (Greenblatt et al. 1994) and starch granule proteins
(Rahman et al. 1994). The presence of these proteins has
been correlated to wheat kernel softness (Jolly et al. 1993;
Sourdille et al. 1996). In wheat endosperm, puroindolines
are localized in the starchy endosperm and in the aleurone
layer (Dubreil et al. 1998). Phase partitioning in a non-ionic
detergent, TX-114, a method that usually speci®cally
extracts membrane proteins (Bordier 1981), suggests that
puroindolines might interact with polar lipids. Whereas
PIN-b interacts only with negative phospholipids, PIN-a is
capable of binding tightly to both wheat phospholipids and
glycolipids (Dubreil et al. 1997). This tight binding of
puroindolines to polar lipids, which has also been observed
with purothionins, could be responsible for their membra-
notoxic effects which play a role in the defence mechanism
against pathogens (Stuart and Harris 1942; Le GuerneveÂ
et al. 1998; Mourgues et al. 1998). Puroindoline-a has
intrinsically good foaming properties and prevents the
lipid-induced destabilization of protein beer foam (Clark
et al. 1994), suggesting that lipid binding may also occur at
the air±water interface. Furthermore, the unique surface
properties of puroindolines make them attractive in bread-
making technology (Dubreil et al. 1998). Le Bihan et al.(1996) observed, by infrared and Raman spectroscopy, that
Correspondence to: H. Boze, UFR Microbiologie Industrielle et de GeÂneÂtique
des Microorganismes, Institut National de la Recherche Agronomique, 2 place
Viala, 34060 Montpellier cedex 01, France (e-mail: [email protected]).
ã 2001 The Society for Applied Microbiology
Journal of Applied Microbiology 2001, 90, 397±406
under acidic and high ionic strength conditions PIN-a tends
to form aggregates.
Considering puroindolines to be important proteins in
future breeding programs and regarding their numerous
technological applications we produced recombinant PIN-a
(rPIN-a) in order to understand its structure±function
relationships. The methylotrophic yeast Pichia pastoris has
the ability to utilize methanol as a sole source of carbon and
energy and has been used extensively as a host system for
the expression of foreign genes (Sreekrishna et al. 1988;
Cregg et al. 1993). The highly ef®cient P. pastoris alcohol
oxidase 1 gene (AOX1) promoter was used for the high-level
expression of foreign genes. This expression system is
particularly valuable for its ability to secrete heterologous
proteins with high ef®ciency. In this paper we report the
production of rPIN-a in a 1á5-l working volume fermenter.
Because of its lipid-binding properties, a large proportion
of rPIN-a is not soluble in the culture medium. To test
whether rPIN-a remains bound to the cell membrane, we
investigated a derivated TX-114 phase partitioning protocol
to make rPIN-a soluble. Firstly, we tried to optimize rPIN-a
production in both fed-batch and continuous culture. The
in¯uence of parameters such as pH, temperature and growth
rate during continuous fermentation was studied, as well as
the addition of Triton X-114 and casamino acids to the
culture medium. The production of rPIN-a was then carried
out using these optimized parameters.
MATERIALS AND METHODS
Yeast and Escherichia coli strains
Escherichia coli strain JM109 was used for all plasmid
constructions. Pichia pastoris strain GS115 (his4) was
obtained from Invitrogen BV (Leek, The Netherlands)
and vector pYAM7SP8 from Dr Laroche (Leuven Catholic
University, Belgium).
Construction of pYAMPIN-a expression vectorand Pichia pastoris transformation
The PIN-a coding sequence from the pTa-31 clone (Gautier
et al. 1994) was ampli®ed by polymerase chain reaction
(PCR) creating a HindII restriction site at the 5¢ end
and an XhoI restriction site at the 3¢ end using the
AP1 (5¢-AGTGTCGACGTTGCTGGCGG-3¢) and AM1
(5¢-TATAGACACCTCGAGCAGGC-3¢) primers. The
HindII-XhoI sequence was then inserted in the pYAM7SP8
vector (Laroche et al. 1994) after cleavage by StuI-XhoI
restriction enzymes. Pichia pastoris strain GS115 (his4) was
transformed with either 20 lg NotI-linearized pYAMPIN-a
or SalI-linearized pYAMPIN-a using the spheroplast
method (Invitrogen) and plated on RDB plates [1 mol l)1
sorbitol, 1% dextrose, 1á34% yeast nitrogen base with
ammonium sulphate and without amino acids, 4 ´ 10)5 %
biotin, 5 ´ 10)3 % amino acids (LL-glutamic acid, LL-lysine,
LL-methionine, LL-leucine and LL-isoleucine) and 2% agar].
His+ transformants were screened for the Mut+/Muts
phenotype by spotting the His+ transformants on minimal
dextrose (MD) medium agar plates (1á34% yeast nitrogen
base with ammonium sulphate and without amino acids,
4 ´ 10)5% biotin, 1% dextrose and 1á5% agar) and minimal
methanol (MM) medium agar plates (as for MD but with
0á5% methanol).
Selection and screening of transformants
The rPIN-a production of different transformed yeasts was
tested using the following method. Each clone was inocu-
lated in 10 ml YMPG medium (0á3% yeast extract, 0á3%
malt extract, 0á5% bactopeptone and 1% glycerol) and
grown overnight in shake ¯asks at 28 °C (80 oscillations
min)1, amplitude 7 cm). After a 10-min centrifugation at
1000 g, the pellet of each clone was diluted with ultra-pure
H2O so that cultures had the same A600 value, approximately
equal to 1. Serial dilutions to 10)5 with ultra-pure water
were carried out and 5 ll of each dilution were spotted onto
MM agar plates. After 24, 48 and 72 h incubation at 28 °C,
the colony-forming units were blotted onto Hybond C extra
(Amersham France S.A., Courtaboeuf, France). Membranes
were soaked for 30 min at room temperature in TBS buffer
(20 mmol l)1 Tris-HCl, pH 7á5 and 0á15 mol l)1 NaCl)
containing 0á1% (v/v) Tween 20 and then incubated for 1 h
in the same buffer containing 1% (w/v) bovine serum
albumin (BSA) and a 1:1000 dilution of an antiserum raised
against the wheat puri®ed PIN-a. Membranes were washed
three times in TBS buffer (10 min each) and primary
antibodies revealed by incubation for 1 h with a 1:10 000
dilution of phosphatase alkaline-labelled goat antirabbit IgG
(Promega, Sigma Chemical Co., St. Louis, MO, USA) in
BSA-Tween TBS buffer. Detection was carried out using an
enzymatic colourimetric assay with a mix of 5-bromo-4-
chloro-3-indolyl phosphate and nitroblue tetrazolium salt in
a 0á1 mol l)1 Tris-HCl, pH 9á5, 0á1 mol l)1 NaCl buffer.
Culture conditions
Cultures were carried out largely as described by Klein et al.(1998). All cultures in shake ¯asks were performed at 29 °C
and under aerobic conditions (80 oscillations min)1, ampli-
tude 7 cm). A 10-ml seed culture of GS115 pYAMPIN-a in
YMPG medium was grown for 24 h and then used to
inoculate 20 ml of a synthetic medium [4% glycerol,
1 ´ PTM1 trace elements, 10 ´ FM21 basal salt and
120 lg l)1 d-biotin (Sreekrishna et al. 1989), buffered at
pH 5á4 with a nitrogen source buffer (0á75% tartric acid,
398 N. ISSALY ET AL .
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 397±406
3á58% Na2HPO4, 12 H2O (dodeca hydrated), 0á2%
SO4(NH4)2 and 0á6% H2PO4NH4)]. This culture was
incubated for 24 h and then transferred into 200 ml of the
same synthetic medium. After 24 h of culture, 200 ml
inoculum were transferred into a 3-l fermenter vessel
(Applikon, Schiedam, The Netherlands) containing 1á5 l
synthetic medium (4% glycerol, 1 ´ PTM1 trace elements,
10 ´ FM21 basal salt and 80 lg l)1 d-biotin). The pH was
regulated with 15% (v/v) ammonium hydroxide, which was
also used as a nitrogen source, and 2 NN H2SO4. The fed-
batch feeding rate was increased from 1á2 to 7á2 ml h)1
during 144 h of induction with a synthetic medium
(682 g l)1 methanol, 1á2 ´ PTM1 and 2á3 mg l)1 d-biotin).
The synthetic medium used for continuous cultures was
composed of 226 g l)1 methanol, 17á5 ´ FM21, 7á5 ´ TM1
and 1á5 mg l)1 d-biotin. During the batch and methanol
induction phases, the pH and temperature were controlled
and dissolved oxygen maintained above 30%.
Determination of cell density
Cell growth was monitored by measuring the absorbance at
600 nm using a spectrometer (DU7; Beckman Instruments
Inc., Fullerton, CA, USA). One A600 unit corresponds to
0á236 � 0á001 g dry weight l)1 on glycerol growth in batch
culture and 0á221 � 0á085 g dry weight l)1 on methanol
growth in fed-batch or continuous cultures.
Analyses of substrates and metabolites
The separation of the cells and supernatant ¯uid was
performed by automatic sterile ®ltration with a Water
Filter/Acquisition module. Glycerol, methanol and formal-
dehyde were separated by a Watersä Fast Fruit Juice
column (Waters Corporation, Milford, MA, USA) in ion
exclusion mode. The mobile phase was 3 mmol l)1 ortho-
phosphoric acid and the ¯ow rate was 1 ml min)1. The
detection was performed by refractometry (410; Waters).
The lower detection limit of the high performance liquid
chromatography was 0á1 g glycerol l)1, 0á1 g methanol l)1
and 0á2 g formaldehyde l)1.
Extraction and solubilization of recombinantpuroindoline-a
The TX-114 phase partitioning was performed on 50 ml
culture during the fed batch phase. Soluble rPIN-a was
extracted for 2 h at 4 °C by gently stirring the sample. After
a ®rst 15-min centrifugation at 8000 g at 4 °C the pellet was
stirred again with 50 ml TX-114 buffer [0á1 mol l)1 Tris-
HCl, pH 7á8, 5 mmol l)1 EDTA, 0á1 mol l)1 KCl and 4%
(v/v) Triton X 114]. This operation was repeated twice on
the pellet. The supernatant ¯uid containing rPIN-a was
heated to 38 °C for 1 h and centrifuged for 15 min at
8000 g at 4 °C. The upper detergent-poor phase was
discarded and rPIN-a present in the rich detergent phase
was precipitated overnight with two volumes of 80% (v/v)
acetone at )20°C. After a 15-min centrifugation at 8000 gand 4 °C, the protein pellet was washed three times with
80% acetone, dried and concentrated 16-fold in ultra-pure
water.
Sodium dodecyl sulphate-polyacrylamide gelelectrophoresis and Western blot analysis
Both the culture supernatant ¯uids and TX-114-rich phase
were analysed on denaturing sodium dodecyl sulphate-
polyacrylamide (18%, w/v) gel electrophoresis under redu-
cing conditions (Laemmli 1970) followed by staining with
Coomassie brillant blue R or immunoblotting. Molecular
weight markers were obtained from GibcoBRL (Life
Technologies, USA) (2850±43 000 molecular range) and
PIN-a isolated from wheat was used as control. For Western
blot analysis, proteins were blotted onto Hybond C extra
(Amersham France S. A., Courtaboeuf, France) with a semi-
dry electro transfer Biometra (Biometra Rudolf-Wissell,
GoÈttingen, Germany) apparatus. Membranes were soaked
for 30 min in TBS buffer (20 mmol l)1 Tris-HCl, pH 7á5,
0á15 mol l)1 NaCl) containing 1% (w/v) BSA and 0á1% (v/
v) Tween 20. The same procedure as described for screening
of the transformant was used.
Immunochemical determination of recombinantpuroindoline-a content
The enzyme linked immunosorbent assay (ELISA) was
performed by immobilizing direct supernatant ¯uid or
TX-114 extracts on 96-well microtiter plates. Each well
was coated overnight at 4 °C with 100 ll of the wheat
puri®ed PIN-a in the 2á5±80 ng ml)1 range or with
suitable sample dilutions in a phosphate-buffered saline
(PBS) buffer, pH 6á9 (8 mmol l)1 Na2HPO4, 1á5 mmol l)1
KH2PO4, 0á14 mol l)1 NaCl and 2á7 mmol l)1 KCl)
containing 1% (w/v) BSA. The plate was washed three
times with the PBS buffer containing 0á05% (v/v) Tween
20. Phosphate-buffered saline-BSA (200 ll) was added
and incubated for 1 h at 37 °C. The plate was washed a
further three times and 100 ll of the polyclonal PIN-a
antibody (dilution 1:2000) added and incubated for 1 h at
37 °C. After three washes, primary antibodies were
revealed by the antirabbit IgG biotin (dilution 1:1000)
conjugated with streptavidin-biotinylated horseradish per-
oxidase complex (dilution 1:1000) and incubated for 1 h at
37 °C. The wells were washed a ®nal three times and
100 ll of TMB peroxidase substrate (Kirkegaard & Perry
Laboratory, Goithersburg, Maryland, USA) and hydrogen
PUROINDOLINE-A PRODUCTION BY PICHIA PASTORIS 399
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 397±406
peroxide solution (H2O2) added. The plate was incubated
at room temperature and the enzyme activity stopped by
the addition of 100 ll 1 mol l)1 H3PO4. The absorbance
measurements were performed at 450 nm.
Subcellular fractionation
Yeast cells were harvested by centrifugation at 8000 g for
10 min at 4 °C and washed twice with a equal volume of the
TBS buffer used in the ELISA procedure. Glass beads (8 g
Glasperlen beads, 0á45±0á5 mm, Braun Melsengen, Postfach,
Germany) were added to 10 ml culture and yeasts were
broken for 2 min using a CO2-cooled Braun MSK bead
grinder apparatus (Braun Melsengen, Postfach, Germany).
The homogenate was centrifuged for 20 min at 3000 g at
4 °C and the protein content measured with the biuret
procedure (Stickland 1951) and the ELISA procedure in
both soluble protein fractions within the supernatant ¯uid
and insoluble protein fractions within the pellet.
RESULTS
Construction of the expression vectorand selection of a productive clone
The PIN-a coding sequence was ampli®ed by PCR from the
cDNA clone pTa31 (Gautier et al. 1994) creating a HindII
restriction site at the 5¢ end and a XhoI restriction site at the
3¢ end (Fig. 1). The PCR products were ®rst inserted into
pGEM-T to give rise to the pGEM-T PIN-a and then the
sorted HindII/XhoI fragment was cloned into the StuI-
XhoI-restricted shuttle vector pYAM7SP8 (Laroche et al.1994). The PIN-a encoding sequence was expressed under
the control of the methanol-inducible alcohol oxidase
promoter using the P. pastoris acid phosphatase secretion
signal fused to a synthetic 19 amino acid pro-sequence
ending with a Lys-Arg processing site for the KEX2
protease. After transformation of P. pastoris, 45 His+
transformants were checked for methanol utilization by
replica plating on MM and MD plates in order to
demonstrate phenotypically if the AOX1 gene remained
intact. Each clone at the same cell density was then spotted
onto MM plates for immunodetection. Among the 28 clones
recognized by the PIN-a antibody, two were chosen
according to their higher level of protein synthesis. These
two clones, 44.8.26 and 44.8.36, seemed to be Mut+
according to the MM/MD screening.
Production of recombinant puroindoline-aby fed-batch fermentation
The production of rPIN-a was ®rst tested with clone
44.8.26. Fermentations consisted of both batch and fed-
batch culture procedures. The temperature and pH were
kept, respectively, at 29 °C and 6 during the fermentation. A
®rst fermentation was carried out using glycerol 4% (v/v) in
the initial batch phase. The methanol fed-batch culture was
then initiated when the glycerol was exhausted. In order to
Fig. 1 Construction of the pYAMPIN-a
expression vector. The puroindoline-a
(PIN-a) coding sequence ¯anked by the
HindII and XhoI restriction sites was cloned
into the pGEM-T vector to give rise to
pGEM-TPIN-a. The HindII/XhoI insert was
then inserted into StuI/XhoI-digested
pYAM7SP8 to give rise to pYAMPIN-a.
PCR, Polymerase chain reaction
400 N. ISSALY ET AL .
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 397±406
adapt cells to methanol, an initial period of slow methanol
feeding was ®rst carried out. Methanol feeding was then
regulated in order to keep the speci®c growth rate equal to
0á01 h)1. Under these conditions, no methanol accumulation
was detected during the induction phase. After 143 h of
induction, the ®nal biomass was 73 g l)1 and the biomass
yield on methanol was 0á13 g g)1 (Fig. 2a). Because pre-
liminary dot-blot assays have suggested that rPIN-a behaved
like a membrane-bound protein (data not shown), a
derivated TX-114 phase partitioning was used to follow
the rPIN-a production. The kinetics of rPIN-a production
was followed by analysis of both the culture supernatant
¯uid and the TX-114 lower phase. The production of rPIN-
a reached 1á11 mg l)1 after 22 h of induction, decreased to
0á12 mg l)1 and remained unchanged during the fermenta-
tion process. This decrease corresponds to the growth phase
after the initial adaptation phase. In the same time the rPIN-
a yield decreased from 46á5 to 1á5 lg g)1 and the speci®c
rPIN-a productivity decreased from 0á46 to 0á01 lg g)1 h)1
at the end of the fed-batch culture (Table 1). From the
beginning of induction to 123 h, the proportion of rPIN-a
secreted in the supernatant ¯uid was insigni®cant compared
with the rPIN-a extract from the TX-114 lower phase. An
increase in secreted rPIN-a was observed after 143 h and
36% of rPIN-a was detected in the culture supernatant ¯uid
at the end of the fed-batch culture. The rPIN-a was still
detected after two TX-114 pellet treatments which indicated
the high af®nity of rPIN-a to polar lipids. Three proteins
recognized by the PIN-a antibody were synthesized during
the production phase. According to their molecular mass,
these proteins might correspond to processed (15 kDa) and
unprocessed rPIN-a (about 17 kDa) and the higher one
might correspond to rPIN-a aggregates (Fig. 2b). Most
rPIN-a seemed to be accumulated as the unprocessed form.
In the ®rst few hours following methanol induction, only the
17-kDa rPIN-a was detected. The protein with the expected
Fig. 2 (a) Growth and production of
recombinant puroindoline-a (rPIN-a) during
the fed-batch methanol induction phase with
the 44á8á26 clone. The rPIN-a was estimated
by enzyme-linked immunosorbent assay in
both supernatant ¯uid culture (S1, j) and
after TX-114 phase partitioning (S2, m). d,
Biomass. (b) Western blot analysis of soluble
rPIN-a during fed-batch culture with the
44á8á26 clone. Culture supernatant ¯uid sam-
ples were taken during the methanol induc-
tion phase and analysed by sodium dodecyl
sulphate-polyacrylamide gel electrophoresis
and then by Western blotting with an anti-
serum raised against the wheat-puri®ed PIN-
a. Lanes: 1±6, 8 ll supernatant ¯uid samples
after 45á5, 69á5, 123, 143, 151 and 167 h of
cultivation; 7, 2 ll wheat-puri®ed PIN-a
PUROINDOLINE-A PRODUCTION BY PICHIA PASTORIS 401
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 397±406
molecular mass was detected at the beginning of the
induction phase and after 143 h of culture. In order to
determine whether variations in culture conditions have an
effect on the production of rPIN-a and on its solubilization,
the in¯uence of several parameters was tested during
continuous culture.
Optimization of the production of recombinantpuroindoline-a by continuous fermentation
Two continuous cultures were carried out independently with
the 44.8.26 clone. In the ®rst continuous culture, the in¯uence
of physico-chemical parameters and the addition of 5 g l)1
casamino acids on the production of rPIN-a were studied. In
the second continuous culture, the in¯uence of dilution rate
and the addition of 0á01% (v/v) TX-114 were analysed.
In¯uence of the pH, temperature and addition ofcasamino acids. In continuous culture no. 1, the dilution
rate was progressively increased to 0á05 h)1 and kept
constant to give rise to 60 g dry weight l)1. The biomass
yield on methanol was 0á3 g g)1 and the volume change time
was 20 h. Variations of pH, temperature and the addition of
5 g l)1 casamino acids were tested. We considered that a
steady state was reached roughly every 2 d when the
biomass and level of dissolved oxygen were stable. The cell
growth rate did not seem to be affected by the pH variation
from 5 to 7, unlike rPIN-a synthesis which seemed to be
optimal under acid conditions (Table 2). The concentration
of rPIN-a increased from 0á13 mg l)1 at pH 6 to
1á15 mg l)1 after 166 h of cultivation at pH 5. At pH 7,
the concentration of rPIN-a fell to 0á037 mg l)1. The
speci®c rPIN-a productivity was 30 times higher when the
pH was monitored at 5 (q rPIN-a � 0á96 lg g)1 h)1) than
when the pH was regulated at 7 (q rPIN-a � 0á03 lg
g)1 h)1).
In the second steady state, when the temperature was
raised to 37 °C, both the concentration and yield production
of rPIN-a decreased twofold and the solubility of rPIN-a
decreased from 33 to 6% (Table 3).
In order to test whether proteases were responsible for the
decrease in rPIN-a detected in the supernatant ¯uid during
the fed-batch culture, 5 g l)1 casamino acids were added in
the last steady state. The presence of casamino acids did not
have a signi®cant in¯uence on the production of rPIN-a but
seemed to affect the solubility of rPIN-a.
In¯uence of the dilution rate and addition of TX-114.The in¯uence of the dilution rate and addition of TX-114
were analysed in continuous culture no. 2. The continuous
culture was initiated after 120 h with temperature and pH,
respectively, set at 29 °C and 6. The biomass yield on
methanol was 0á3 g g)1. Two dilution rates were ®rstly
tested, 0á03 h)1 in the ®rst steady state and 0á05 h)1 in the
second. The concentration of rPIN-a decreased threefold
from 1á38 to 0á43 mg l)1 when the dilution rate increased,
while the speci®c productivity of rPIN-a decreased from 0á69
to 0á35 lg g)1 h)1 (Table 4). This was in good agreement
with the results obtained during the adaptation phase. The
Table 1 Recombinant puroindoline-a (rPIN-a) concentration
(mg l)1), speci®c productivity (q; lg g)1 h)1) and yield production (Y;
lg g)1) determined by ELISA during the fed-batch culture with the
44.8.26 clone
Time of
methanol
induction (h)
rPIN-a
(mg l)1)
q rPIN-a
(lg g)1 h)1)
Y rPIN-a
(lg g)1)
% rPIN-a
membrane-
bound (S2)
22á5 1á11 0á46 46á5 89
46á5 0á12 0á04 4á3 98
100 0á17 0á03 3á3 99
120 0á07 0á01 1á1 72
144 0á11 0á01 1á5 64
S2, Membrane-bound rPIN-a detected in the successive TX-114 phase
partitioning.
Table 2 In¯uence of pH on recombinant puroindoline-a (rPIN-a)
concentration (mg l)1), speci®c productivity (q; lg g)1 h)1) and
yield production (Y; lg g)1) during continuous culture no. 1 with the
44.8.26 clone
pH
rPIN-a
(mg l)1)
q rPIN-a
(lg g)1 h)1)
Y rPIN-a
(lg g)1)
% rPIN-a
membrane-
bound (S2)
5 1á16 0á96 19á3 87
6 0á13 0á11 2á20 55
7 0á04 0á03 0á62 98
A steady state was obtained at 29 °C when X = 60 g l)1 with a dilution
rate D of 0á05 h)1. S2, Membrane-bound rPIN-a.
Table 3 In¯uence of the temperature on recombinant puroindo-
line-a (rPIN-a) concentration (mg l)1), speci®c productivity
(q; lg g)1 h)1) and yield production (Y; lg g)1) during continuous
culture no. 1 with the 44.8.26 clone
Tempera-
ture (°C)
rPIN-a
(mg l)1)
q rPIN-a
(lg g)1 h)1)
Y rPIN-a
(lg g)1)
% rPIN-a mem-
brane-bound (S2)
29 0á06 0á05 0á98 67
29* 0á05 0á04 0á80 94
37 0á03 0á033 0á55 94
A steady state was obtained when X = 60 g l)1, at pH 6; with a
dilution rate D of 0á05 h)1. S2, Membrane-bound rPIN-a.
* 5 g l)1 casamino acids were added after 311 h when the temperature
was decreased to 29 °C.
402 N. ISSALY ET AL .
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 397±406
production of rPIN-a seemed to be improved when the
culture was carried out with a low speci®c growth rate.
A third steady state allowed us to test the effect of the
addition of 0á01% TX-114 with a dilution rate of 0á06 h)1.
The concentration of rPIN-a increased to 1á58 mg l)1,
which was more than when the culture was carried out with
a dilution rate of 0á03 h)1 but without TX-114. Further-
more, when TX-114 was present in the culture medium,
most rPIN-a was in the soluble form in the culture
supernatant ¯uid. The speci®c productivity of rPIN-a was,
therefore, 1á58 lg g)1 h)1 and 1á6 times higher than the
productivity of rPIN-a when the ®rst continuous culture
was carried out with a dilution rate of 0á05 h)1 and at pH 5.
Production of recombinant puroindoline-aby continuous fermentation with favourableconditions
Continuous culture no. 3 was then carried out in order to
monitor the production of rPIN-a using the parameters
de®ned above. The temperature and pH were, respectively,
set at 29 °C and 5. The dilution rate was about D � 0á03 h)1
and 0á01% (v/v) TX-114 was added to the culture medium.
The 44.8.36 clone was used for this experiment because it
produced eight times more rPIN-a than the 44.8.26 clone by
fed-batch culture (data not shown). The dilution rate was ®rst
monitored from 0á003 to 0á012 h)1 during 10 h in order to
adapt the cells to methanol. This adaptation phase was run too
fast and the accumulation of formaldehyde required methanol
feeding to be stopped. TX-114 was supplied gradually to the
fermenter vessel inducing foam production, volume changes
and steady state interruptions that prevented the continuous
culture from being carried out with the desired dilution rate.
Nevertheless, three steady states were obtained with different
dilution rates: (i) D � 0á025 h)1; (ii) D � 0á033 h)1 and (iii)
D � 0á038 h)1. During the ®rst steady state, the concentra-
tion of rPIN-a increased from 3á3 mg l)1 at the end of the
adaptation phase to 13á8 mg l)1 (10 times higher than the
concentration of rPIN-a previously detected). The speci®c
productivity of rPIN-a reached 5á75 lg g)1 h)1. The next
two steady states ran with a higher dilution rate which could
explain the decrease in the production of rPIN-a to 7á1 and
1á52 mg l)1, respectively (Table 5).
However, more than 80% of the rPIN-a was secreted and
detected in the culture supernatant ¯uid when the dilution
rate was set to 0á025 h)1. The addition of 0á01% (v/v) TX-
114 inverted the ratio between the soluble and the mem-
brane-bound rPIN-a. A subcellular fractionation was per-
formed on 10 ml of the culture when the dilution rate was
0á038 h)1. Secreted rPIN-a represented only 31% of the
total rPIN-a and 22% of the rPIN-a was contained within
the perisplasm and cytoplasm of the yeast. Only 6% of the
rPIN-a remained insoluble and present within the pellet
fraction (Table 6).
DISCUSSION
For the ®rst time an rPIN-a was produced in fermentation
culture by the methylotrophic yeast P. pastoris. The aim of
the present study was to test different culture conditions
to improve the production of rPIN-a. The pYAM7SP8
expression vector was chosen because it allows cDNA fusion
to a secretory leader sequence which has often been used
successfully for the production and secretion of foreign
Table 4 In¯uence of the addition of TX-114 and the dilution rate
(D) on recombinant puroindoline-a (rPIN-a) concentration
(mg l)1), speci®c productivity (q; lg g)1 h)1) and yield production (Y;
lg g)1) during continuous culture no. 2 with the 44.8.26 clone
D (h)1) rPIN-a
(mg l)1)
q rPIN-a
(lg g)1 h)1)
Y rPIN-a
(lg g)1)
% rPIN-a mem-
brane-bound (S2)
0á03 1á38 0á69 23á0 86
0á05 0á43 0á35 7á16 91
0á06* 1á58 1á58 26á3 44
A steady state was obtained when X = 60 g l)1, T = 29 °C and at
pH 6. S2, Membrane-bound rPIN-a.
* 0á01% TX-114 was added after 270 h when the dilution rate was
0á06 h)1.
Table 5 Recombinant puroindoline-a (rPIN-a) concentration
(mg l)1), speci®c productivity (q; lg g)1 h)1) and yield production (Y;
lg g)1) during continuous culture no. 3 with the 44.8.36 clone
D (h)1) rPIN-a (mg l)1) q rPIN-a (lg g)1 h)1) Y rPIN-a (lg g)1)
0á025 13á8 5á75 230
0á033 7á10 3á89 118
0á038 1á52 0á96 25á3
Fermentation was carried out under favourable conditions. A steady
state was obtained when X = 60 g l)1, T = 29 °C, at pH 5, and with a
dilution rate D of 0á03 h)1. 0á01% TX-114 was added to the culture
medium.
Table 6 Protein measurement with the biuret reaction (mg) and
recombinant puroindoline-a (rPIN-a) measurement by enzyme-
linked immunosorbent assay (mg) in subcellular fractions after
yeast grinding
Protein content (mg)
Total rPIN-a
Secreted proteins 24 0á018
Membrane-bound proteins 46á1 0á023
Soluble proteins after cell breakage 228 0á013
Insoluble proteins after cell breakage 152 0á004
Subcellular fractionation was performed on 10 ml of continuous
culture no. 3 when the dilution rate was 0á038 h)1.
PUROINDOLINE-A PRODUCTION BY PICHIA PASTORIS 403
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 397±406
proteins (Laroche et al. 1994; De Klein et al. 1998; Inan
et al. 1999). The prohormome-processing KEX2 protease
allows the cleavage, on the C side, of a pair of dibasic
residues, Lys-Arg (Brenner and Fuller 1992). However, the
translocation folding, intracellular transport and secretion of
a 9-kDa rLTP have already shown the in¯uence of the
medium on the ef®ciency of KEX 2 (De Klein et al. 1998).
Because several studies have reported that Muts transfor-
mants sometimes result in an enhanced production of
foreign proteins compared with Mut� proteins (Cregg and
Madden 1987; Tshopp et al. 1987; Chirulova et al. 1997),
both AOX1 and HIS4 insertion were carried out. Among
the 28 clones giving a positive response by the immuno-
bloting detection, only two seemed to be Muts
. In fact, we
did not observe any signi®cant relationships between the
phenotype of transformants and the production of rPIN-a.
In several examples, improved recombinant protein expres-
sion yields were observed for multicopy clones (Romanos
et al. 1992; Cregg et al. 1993). These results are in good
agreement with our own results (data not shown). Two
Mut+ transformants were chosen to test the production of
rPIN-a and allowed the feeding rate to be more freely
controlled than a culture with a Muts strain.
During fed-batch fermentation, mature rPIN-a was
detected in the ®rst hours of methanol induction and then
disappeared suddenly. Several hypotheses can be proposed
to explain this observation. Proteases could be responsible
for the decrease in rPIN-a in culture supernatant ¯uid. The
use of a protease-de®cient strain has been shown to be a
successful approach to improve the production yield of
heterologous proteins in both Saccharomyces cerevisiae and
P. pastoris (Sander et al. 1994; Weiss et al. 1995). The
utilization of strain SMD1168 could lead to better results in
the production of rPIN-a. In the same way, the addition of
protease inhibitors to the culture media could increase the
production of rPIN-a. Nevertheless, rPIN-a did not remain
at the same concentration level despite the addition of
casamino acids during continuous fermentation in order to
prevent protease activity (Clare et al. 1991).
The changeover from the secreted to the membrane-
bound rPIN-a could be another hypothesis explaining this
decrease. In high density fed-batch fermentation most rPIN-
a remains membrane-bound during the culture. It had been
observed that PIN-a tends to form aggregates under acid
and high ionic strength conditions (Le Bihan et al. 1996).
Despite the presence of ®ve disulphide bridges, the
secondary structure of PIN-a is pH dependant and the
reduction in disulphide bonds leads to a signi®cant loss of
solubility. We have observed that the solubility of rPIN-a
increased at low pH; this is contradictory to previous
observations. The decrease observed after 43 h during fed-
batch culture was repeatable with all the clones tested in the
same culture conditions. Another hypothesis explaining the
decrease in rPIN-a in the culture supernatant ¯uid could be
based on non-detection by the antiserum against PIN-a
when rPIN-a forms aggregates with other cell proteins.
The speci®c growth rate had an important in¯uence on
the productivity yield of rPIN-a. Maximum productivity of
rPIN-a was observed during the initial period of slow
methanol feeding of the fed-batch culture. There is no
signi®cant relationship between the production yield of
foreign protein and the speci®c growth rate but, in a similar
way, a low speci®c growth rate seemed to improve the
production of a recombinant porcine follicle-stimulating
hormone (FSH) (Boze, personal communication).
The most striking result was obtained by the addition of
0á01% (v/v) TX-114 directly to the feeding medium. The
high af®nity of rPIN-a to polar lipids was a hindrance to
protein secretion. Phase separation under salt in¯uence for
non-ionic detergents is a simple and effective method for the
concentration of solubilized membrane proteins (Fricke
1993). The TX-114 phase partitioning allowed the remain-
ing cell-bound rPIN-a to be solubilized. Moreover, the
addition of TX-114 during fermentation was tested accord-
ing to the results, showing that the addition of 0á01% (v/v)
TX-100 to a feeding medium partially reduced the prote-
olysis of a urokinase-type plasminogen activator and
increased the secretion level (Tsujikawa et al. 1996). Finally,
TX-114 increased by 10-fold the production yield of rPIN-a
to 13 mg l)1 and inverted the ratio between secreted and
membrane-bound rPIN-a. This suggests that TX-114 may
perturb the adhesion of rPIN-a to the cell wall during yeast
growth and further secretion. Results of subcellular frac-
tionation have shown that an important part of rPIN-a was
still contained within the periplasm of P. pastoris. Whether a
secreted protein lodges in the periplasmic space, cell wall or
diffuses into the culture medium depends partly on the
size of that protein. The yeast cell wall strongly in¯uences
permeability in conventional yeasts and disulphide bridges
have been recognized to play a prominent role in determin-
ing the shape of the cell wall (Kidby and Davies 1970).
Thus, the cell wall represents an ion exchange ®lter with the
porosity being dependent on the cell wall charge (De Nobel
et al. 1990). Acidic conditions could increase the elasticity of
the cell wall and allow better secretion.
The aim of the production of rPIN-a was to provide large
quantities for the rPIN-a structural study. Contrary to all
expectations, the production yield was not the same as the
production yield obtained with other wheat seed cystine-rich
proteins (Klein et al. 1998). The production of rPIN-a could
be compared with the production yield (about 1 mg l)1) of
the heterologous membrane receptor secreted by P. pastoris(Cereghino and Cregg 2000). Our results allow a 10-fold
enhancement in the production yield of rPIN-a when
fermentation culture is carried out under precise conditions.
However, expected production yields were not reached. It
404 N. ISSALY ET AL .
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 397±406
will be interesting to consider other expression systems in
order to increase the production of rPIN-a.
ACKNOWLEDGEMENTS
The authors would like to thank Dr L. Quillien (INRA
Nantes) for providing the polyclonal anti-PIN-a antibody,
P. Brignon (TEPRAL Strasbourg) for puri®ed PIN-a and
Dr Y. Laroche (Leuven Catholic University, Belgium) for
the pYAM7SP8 vector. This work was supported by the
MinisteÁre de l'Enseignement SupeÂrieur et de la Recherche
(grant MESR 97C0101) and N.I. was the recipient of a grant
from MENRT.
REFERENCES
Blochet, E., Chevalier, C., Forest, E., Pebay-Peyroula, E., Gautier,
M.F., Joudrier, P., Pezolet, M. and Marion, D. (1993) Complete
amino sequence of puroindoline, a new basic and cystine-rich
protein with a unique tryptophan-rich domain, isolated from wheat
endosperm by Triton X-114 phase partitioning. FEBS Letters 3,
336±340.
Bordier, C. (1981) Phase separation of integral membrane proteins in
Triton X-114 solution. Journal of Biological Chemistry 256, 1604±
1607.
Brenner, C. and Fuller, R.S. (1992) Structural and enzymatic
characterisation of a puri®ed prohormone-processing enzyme:
secreted, soluble Kex2 protease. Proceedings of the National Academy
of Sciences of the United States of America 89, 922±926.
Cereghino, J.L. and Cregg, J.M. (2000) Heterologous protein expres-
sion in the methylotrophic yeast Pichia pastoris. FEMS Microbiology
Reviews 24, 45±66.
Chirulova, V., Cregg, J.M. and Meagher, M.M. (1997) Recombinant
protein production in an alcohol oxidase-defective strain of Pichia
pastoris in fed-batch fermentation. Enzyme and Microbial Technology
21, 277±283.
Clare, J.J., Romanos, M.A., Rayment, F.B., Rowedder, J.E., Smith,
M.A., Payne, M.M., Sreekrishna, K. and Henwood, C.A. (1991)
Production of mouse epidermal growth factor in yeast: high level
secretion using Pichia pastoris strains containing multiple gene
copies. Gene 105, 205±212.
Clark, C., Wilde, P.J. and Marion, D. (1994) The protection of beer
foam against lipid-induced destabilization. Journal of the Institute of
Brewing 100, 23±25.
Cregg, J.M. and Madden, K.N. (1987) Development of transformation
systems and construction of methanol-utilization-defective mutants
of Pichia pastoris by gene disruption. Biological Research on Industrial
Yeast II, 1±18.
Cregg, J.M., Vedvick, T.S. and Raschke, W.E. (1993) Recent advances
in the expression of foreign gene in Pichia pastoris. Bio/Technology 5,
479±485.
De Nobel, J.G., Klis, F.M., Priem, J., Munnik, T. and Van Der Ende,
H. (1990) An assay of relative cell wall porosity in Saccharomyces
cerevisiae, Kluyveromyces lactis and Schizosaccharomyces pombe. Yeast
6, 483±490.
Dubreil, L., Compoint, J.P. and Marion, D. (1997) Interaction of
puroindolines with wheat ¯our polar lipids determines their foaming
properties. Journal of Agricultural and Food Chemistry 45, 108±116.
Dubreil, L., Gaborit, T., Bouchet, B., Gallant, D.J., Broekaert, W.F.,
Quillien, L. and Marion, D. (1998) Spatial and temporal distribution
of the major isoforms of puroindolines (puroindoline-a and puro-
indoline-b) and non speci®c lipid transfer protein (ns-LTP) of
Triticum aestivum seeds. Relationships with their in vitro antifungal
properties. Plant Sciences 138, 121±135.
Dubreil, L., MeÂliande, S., Chiron, H., Compoint, J.P., Quillien, L.,
Branlard, G. and Marion, D. (1998) Effect of puroindolines on the
breadmaking properties of wheat ¯our. Cereal Chemistry 75,
222±229.
Fricke, B. (1993) Phase separation of nonionic detergents by salt
addition and its application to membrane proteins. Analytical
Biochemistry 212, 154±159.
Gautier, M.F., Aleman, M.E., Guirao, A., Marion, D. and Joudrier, P.
(1994) Triticum aestivum puroindolines, two basic cystine-rich seed
proteins: cDNA sequence analysis and developmental gene expres-
sion. Plant Molecular Biology 25, 43±57.
Greenblatt, G.A., Bettge, A.D. and Morris, C.F. (1994) Relationship
between endosperm texture and the occurrence of friabilin and
bound polar lipids on wheat starch. Cereal Chemistry 72, 172±176.
Inan, M., Chirovolu, V., Eskridge, K.M., Vlasuk, G.P., Dickerson, K.,
Brown, S. and Meagher, M.M. (1999) Optimization of temperature-
glycerol-pH conditions for a fed-batch fermentation process for
recombinant hookworm (Ancylostoma caninum) anticoagulant peptide
(AcAP-5) production by Pichia pastoris. Enzyme and Microbial
Technology 24, 438±445.
Jolly, C.J., Rahman, S., Kortt, A.A. and Higgins, T.G.V. (1993)
Characterization of the wheat Mr 15000 `grain softness protein' and
analysis of the relationship between its accumulation in the whole
seed and grain softness. Theoretical Apply Genetic 86, 589±597.
Kidby, D.K. and Davies, R. (1970) Invertase and disul®de bridges in
the yeast wall. Journal of General Microbiology 61, 327±333.
Klein, C., De Lamotte-GueÂry, F., Gautier, F., Moulin, G., Boze, H.,
Joudrier, P. and Gautier, M.F. (1998) High level secretion of a wheat
lipid transfer protein in Pichia pastoris. Protein Expression and
Puri®cation 13, 73±82.
Laemmli, U.K. (1970) Cleavage of structural proteins during the
assembly of the head of bacteriophage T4. Nature 227, 680±685.
Laroche, Y., Storme, V., De Meutter, J., Messens, J. and Lauwereys,
M. (1994) High level secretion and very ef®cient isotopic labeling of
tick anticoagulant peptide (TAP) expressed in the methylotrophic
yeast Pichia pastoris. Bio/Technology 2, 1119±1124.
Le Bihan, T., Blochet, J.E., Desormeaux, A., Marion, D. and PeÂzolet,
M. (1996) Determination of the secondary structure and conforma-
tion of puroindolines by infrared and raman spectroscopy. Biochem-
istry 35, 12712±12722.
Le GuerneveÂ, C., Seigneuret, M. and Marion, D. (1998) Interaction of
the wheat endosperm lipid-binding protein puroindoline-a with
phospholipids. Archives of Biochemistry and Biophysics 360, 179±186.
Mourgues, F., Brisset, M.N. and Chevreau, E. (1998) Activity of
different antibacterial peptides on Erwinia amylovora growth, and
evaluation of the phytotoxicity and stability of cecropins. Plant
Sciences 139, 83±91.
PUROINDOLINE-A PRODUCTION BY PICHIA PASTORIS 405
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 397±406
Rahman, S., Jolly, C.J., Skerrit, J.H. and Wallosheck, A. (1994)
Cloning of a wheat 15-kDa grain softness protein (GSP). GSP is a
mixture of puroindoline-like polypeptides. European Journal of
Biochemistry 223, 917±925.
Romanos, M.A., Scorer, C.A. and Clare, J.J. (1992) Foreign gene
expression in yeast: a review. Yeast 8, 423±488.
Sander, P., Grunewald, S., Bach, M., Haase, W., Reilander, H. and
Michel, H. (1994) Heterologous expression of the human D2S
dopamine receptor in protease-de®cient Saccharomyces cerevisae
strains. European Journal of Biochemistry 226, 697±705.
Sourdille, P., Perretant, M.R., Charmet, G., Leroy, P., Gautier, M.F.,
Joudrier, P., Nelson, J.C., Sorrells, M.E. and Bernard, M. (1996)
Linkage between RFLP markers and genes affecting kernel hardness
in wheat. Theoretical Apply Genetic. 93, 580±586.
Sreekrishna, K., Nelles, L., Potenz, R., Cruze, J., Mazzaferro, P., Fish,
W., Fuke, M., Holden, K., Phelps, D., Wood, P. and Parker, K.
(1989) High level expression, puri®cation and characterization of
recombinant human tumour necrosis factor synthesized in the
methylotrophic yeast Pichia pastoris. Biochemistry 28, 4117±4125.
Sreekrishna, K., Potenz, R.B., Cruze, J.A., McCombie, W.R., Parker,
K.A., Nelles, L., Mazzaferro, P.K., Holden, K.A., Harrison, R.G.,
Wood, P.J., Phelps, D.A., Hubbard, C.E. and Fuke, M. (1988) High
level expression of heterologous proteins in methylotrophic yeast
Pichia pastoris. Journal of Basic Microbiology 28, 265±278.
Stickland, L.H. (1951) The determination of small quantities of
bacteria by means of biuret reaction. Journal of General Microbiology
5, 698±701.
Stuart, L.S. and Harris, T.H. (1942) Bactericidal and fungicidal
properties of a crystalline protein isolated from unbleached wheat
¯our. Cereal Chemistry 19, 288±300.
Tshopp, J.F., Sverlow, G., Kosson, R., Craig, W. and Grinna, L.
(1987) High level secretion of glycosylated invertase in the
methylotrophic yeast Pichia pastoris. Bio/Technology 5, 1305±1308.
Tsujikawa, M., Okabayashi, K., Morita, M. and Tanabe, T. (1996)
Secretion of a variant of human single-chain urokinase-type
plasminogen activator without an N-glycosylation site in the
methylotrophic yeast Pichia pastoris and characterisation of the
secreted product. Yeast 12, 541±553.
Weiss, H.M., Haase, W., Michel, H. and Reilander, H. (1995)
Expression of functional mouse 5-HT5A serotonin receptor in the
methylotrophic yeast Pichia pastoris: pharmacological characteriza-
tion and localisation. FEBS Letters 337, 451±456.
406 N. ISSALY ET AL .
ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 397±406