Appl Microbiol Biotechnol (2013) 97:1201–1212

DOI 10.1007/s00253-012-4270-2

APPLIED GENETICS AND MOLECULAR BIOTECHNOLOGY

Constitutive high-level expression of a codon-optimized

β-fructosidase gene from the hyperthermophile
Thermotoga maritima in Pichia pastoris
Carmen Menéndez & Duniesky Martínez &
Luis E. Trujillo & Yuliet Mazola & Ernesto González &
Enrique R. Pérez & Lázaro Hernández

Received: 9 May 2012 / Revised: 24 June 2012 / Accepted: 25 June 2012 / Published online: 21 July 2012
# Springer-Verlag 2012

Abstract Enzymes for use in the sugar industry are preferred
to be thermotolerant. In this study, a synthetic codon-
optimized gene encoding a highly thermostable β-
fructosidase (BfrA, EC 3.2.1.26) from the bacterium Thermo-
toga maritima was expressed in the yeast Pichia pastoris. The
gradual increase of the transgene dosage from one to four
copies under the control of the constitutive glyceraldehyde 3-
phosphate dehydrogenase promoter had an additive effect on
BfrA yield without causing cell toxicity. Maximal values of
cell biomass (115 g/l, dry weight) and overall invertase activ-
ity (241 U/ml) were reached at 72 h in fed-batch fermentations
using cane sugar as the main carbon source for growth.
Secretion driven by the Saccharomyces cerevisiae α-factor
signal peptide resulted in periplasmic retention (44 %) and
extracellular release (56 %) of BfrA. The presence of N-linked
oligosaccharides did not influence the optimal activity, ther-
mal stability, kinetic properties, substrate specificity, and exo-
type action mode of the yeast-secreted BfrA in comparison to
the native unglycosylated enzyme. Complete inversion of
cane sugar at initial concentration of 60 % (w/v) was achieved
by periplasmic BfrA in undisrupted cells reacting at pH 5.5
and 70 °C, with average productivity of 4.4 g of substrate
hydrolyzed per grams of biomass (wet weight) per hour. The
high yield of fully active glycosylated BfrA here attained by
recombinant P. pastoris in a low-cost fermentation process
appears to be attractive for the large-scale production of this
thermostable enzyme useful for the manufacture of inverted
sugar syrup.

Keywords Invertase . Fructosidase . Invert sugar . Pichia

C. Menéndez (*) : L. E. Trujillo : L. Hernández
pastoris . GAP promoter . Thermotoga maritima
Plant–Microbe Interactions Laboratory,
Center for Genetic Engineering and Biotechnology (CIGB),
Ave 31 entre 158 y 190, Apartado Postal 6162,
Havana 10600, Cuba Introduction
e-mail: carmen.menendez@cigb.edu.cu

D. Martínez : E. R. Pérez Invertase (β-D-fructofuranosidase, EC 3.2.1.26) cleaves su-

Fermentation Laboratory, Center for Genetic Engineering
and Biotechnology Sancti Spíritus (CIGBSS),
Circunvalante Norte S/N, Olivos 3, Apartado Postal 83,
Sancti Spíritus 60200, Cuba
Y. Mazola
Bioinformatics Department,
Center for Genetic Engineering and Biotechnology (CIGB),
Ave 31 entre 158 y 190, Apartado Postal 6162,
Habana 10600, Cuba
E. González
Protein Purification Laboratory,
Center for Genetic Engineering and Biotechnology (CIGB),
Ave 31 entre 158 y 190, Apartado Postal 6162,
Habana 10600, Cuba
crose into glucose and fructose. The equimolar mixture of
the two products is commercially known as invert sugar,
which is non-crystallizable and sweeter than sucrose be-
cause of the fructose content. Due to the hygroscopic nature
of glucose, invert syrup is used as a humectant in the
manufacture of confectionery and bakery products. Inver-
tase finds also application whenever sucrose-containing
substrates are subjected to fermentation (production of alco-
holic beverages, lactic acid, glycerol, etc) and for the hy-
drolysis of β(2,1)-linked fructans by its associated inulinase
activity. Other uses of the enzyme include the manufacture
of artificial honey and plasticizing agents used in cosmetics,
drug, and paper industries (Kotwal and Shankar 2009).
1202
Invertases are ubiquitous in plants and also exist in a
wide range of saccharolytic microorganisms (Sturm 1999;
Kotwal and Shankar 2009). Microbial invertases are gener-
ally more stable and easier to obtain for use in biotechnol-
ogy. Saccharomyces cerevisiae is by far the main source of
the enzyme for the commercial production of invert sugar
(Kotwal and Shankar 2009). This yeast synthesizes a glyco-
sylated periplasmic invertase and an unglycosylated cyto-
solic form from two different mRNAs that are transcribed
from the same gene (Carlson and Botstein 1982). S. cerevi-
siae invertase exhibits maximal activity at pH 4.5–5.0 and
55–60 °C with high catalytic efficiency (kcat/KM 9–40×104
M−1 s−1) (Reddy and Maley 1996; Lafraya et al. 2011).
However, the enzyme suffers from substrate inhibition at
sucrose levels above 20 % (w/v) and it is not resistant to heat
denaturation even after immobilization, hampering its con-
tinuous reuse when the operating temperature exceeds 40–
45 °C (Hasal et al. 1992; Kotwal and Shankar 2009).
Enzymes in the sugar industry are preferred to stably
operate at relatively high temperatures (55–70 °C) in con-
centrated substrate solutions (about 60 %, w/v). The hyper-
thermophilic bacterium Thermotoga maritima produces an
exo-type β-fructosidase (BfrA) involved in the utilization of
sucrose and fructans. BfrA produced in Escherichia coli
displayed similar catalytic efficiencies (kcat/KM 3–4×104
M−1s−1 at pH 5.5 and 75 °C) for the hydrolysis of sucrose
and inulin (Liebl et al. 1998). The recombinant BfrA was
markedly insensitive to inactivation at temperatures up to
70 °C, and sucrose cleavage was not severely inhibited at
the highest assayed substrate concentration of 27.4 % (w/v)
(Liebl et al. 1998). Due to these properties, BfrA constitutes
an attractive candidate for use in the industrial production of
invert sugar.
The expression of genes from thermophilic microorgan-
isms into more suitable mesophilic hosts is at hand to
produce stable biocatalysts. E. coli has been manipulated
to overproduce active thermostable enzymes from T. maritima
(Liebl et al. 1998; Wang and Zhang 2009), but it is not an
appropriate host for protein secretion. The yeast Pichia pas-
toris was successfully used for the secretion of bacterial
enzymes of biotechnological interest, such as levansucrase
and levanase from Gluconacetobacter diazotrophicus
(Trujillo et al. 2001 and 2004; Menéndez et al. 2004)
and α-amylase from Bacillus licheniformis (Paifer et al.
1994). For the cost-effective production and secretion of
foreign proteins in P. pastoris, the availability of com-
mercial vectors with strong constitutive or methanol-
regulated promoters and efficient signal peptide sequen-
ces it is significant, as well as the yeast ability to reach
high cell densities in relatively cheap media at the
optimal growth temperature of 30 °C. When N-glyco-
sylation occurs, P. pastoris generates protein-bound oli-
gosaccharides that are of shorter chain length than those
Appl Microbiol Biotechnol (2013) 97:1201–1212
found in S. cerevisiae (Bretthauer and Castellino 1999;
Cereghino and Cregg 2000). P. pastoris has the gener-
ally regarded as safe status and lacks endogenous
sucrases; thus, it is particularly useful for the recombi-
nant production of sucrose-converting enzymes with
applications in the food industry.
In this study, a synthetic codon-optimized gene encoding
the intact amino acid sequence of BfrA was fused to the
S. cerevisiae α-factor signal peptide and expressed un-
der the control of the strong constitutive glyceraldehyde
3-phosphate dehydrogenase (GAP) promoter in P. pas-
toris. BrfA was secreted as a glycoprotein with thermal
stability and catalytic properties similar to those
reported for the recombinant enzyme produced in E.
coli. To our knowledge, this is the first report on the
expression of a bacterial thermostable invertase in yeast.
Materials and methods
Microorganisms, media, and substrates
E. coli strain DH5α (Novagen) was used as a cloning host
and for plasmid propagation. P. pastoris strains X-33 (wild-
type) and GS115 (his4-deficient) from Invitrogen were used
as the expression hosts. E. coli was grown at 37 °C in LB
medium or low-salt LB medium with zeocin (25 μg/ml).
Ampicillin (50 μg/ml) was added to LB medium when
needed. P. pastoris was grown at 30 °C in YPG medium
[1 % (w/v) yeast extract, 2 % (w/v) peptone, and 2 % (v/v)
glycerol] or YNB-G plates [1.34 % (w/v) yeast nitrogen base
without amino acids and 2 % (w/v) glucose] supplemented
or not with L-histidine (100 g/ml). Solid SPY medium [5 %
(w/v) sucrose, 2 % (w/v) peptone, 1 % (w/v) yeast extract,
and 0.025 % (w/v) bromothymol blue, 1.5 % (w/v) agar, pH
6.5] was used for phenotypic screening of clones expressing
active BfrA. The medium for fermentation of P. pastoris strain
PpBfrA(4×) during the batch phase consisted of 1 % (w/v)
cane sugar, 4 % (v/v) glycerol, 0.5 % (w/v) yeast extract, 2.2 %
(w/v) (NH4)2SO4, 1.82 % (w/v) KH2PO4, 0.75 % (w/v)
MgSO4·7H2O, and 0.05 % (w/v) CaCl2·2H2O, with vitamins
and traces prepared as recommended by Cregg et al. (1987).
The substrates levan from Erwinia herbicola, inulin from
Dahlia tubers, sucrose, raffinose, and melezitose assayed for
BfrA activity were all from Sigma. The FOS 1-kestose and
nystose were kindly donated by Dr. Hidetoshi Kubota from
Meiji Seika Kaisha Ltd.
Genetic constructs, yeast transformation, and phenotypic
screening of clones expressing active BfrA
The β-fructosidase gene (bfrA) from T. maritima MSB8
(GenBank Accession No. AJ001073) with triplet sequences
Appl Microbiol Biotechnol (2013) 97:1201–1212 1203

Numbers indicate the quantity of each codon in the mRNA. Codons of preferred usage in the host yeast are underlined. Gene modification kept unchanged the protein sequence, as it is reported for
UAA

UAG

UGA
Stop
optimized for expression in P. pastoris (Table 1) was chem-

1
ically synthesized (GeneArt) with flanking EcoRI and SalI
restriction sites and supplied in the pMA vector. Enhance-

UGG
11
Trp
ment of mRNA stability was considered in addition to
codon optimization. The bfrA coding sequence in the syn-

AUG
Met
thetic EcoRI-SalI 1.3-kb fragment was fused to the N-

8
terminal α-factor signal sequence of S. cerevisiae and to
the C-terminal polyhistidine tag in the zeocin-resistant vec-

CAU
CAC
His

5
tor pGAPZαC (Invitrogen). The resulting plasmid pGAPb-
frA(1×)zeo carries a single copy of the chimeric bfrA gene

UUU

UUC
21
under the control of the constitutive GAP promoter and the

Phe

4
AOX1 terminator. To duplicate the gene dosage, the 2.4-kb
BglII-BamHI fragment consisting of the bfrA expression

UGU
UGC
Cys

5
cassette was incorporated at the BamHI site of the own
plasmid pGAPbfrA(1×)zeo creating the plasmid pGAPb-

AAU
AAC
21
frA(2×)zeo. The resulting 4.8-kb BglII-BamHI fragment

Asn

2
from pGAPbfrA(2×)zeo was placed at the BamHI site of
the ampicillin-resistant vector pAO815 (Invitrogen) to cre-

AAA

AAG
24
Lys

5
ate the plasmid pGAPbfrA(2×)his carrying the two in-
tandem copies of the bfrA expression cassette and the his4

GAU
GAC
gene. The three constructed plasmids were digested with the

21
Asp

5
appropriate restriction enzyme and electroporated into X-33
or GS115 using the procedure described in the manual for
GAA

GAG
36
Glu

protein expression in Pichia from Invitrogen. GS115 trans-

6
formants were selected on histidine-deficient YNB-G plates.
X-33 transformants or GS115 double transformants were AUU

AUA
AUC

selected on YPG plates with zeocin (100 μg/ml). Control 19

2

0
Ile

recombinant P. pastoris strains were created by transforming

UUA

UUG
CUU

CUA

CUG
CUC
Table 1 Codon usage of the synthetic β-fructosidase gene (bfrA) expressed in P. pastoris

X-33 and GS115 with the empty vectors pGAPZαC and

34
Leu

0
pAO815, respectively.
Single zeocin-resistant and histidine-complemented col-

AGU
UCU

UCA

UCG

AGC
UCC

22
Ser

onies transformed with the expression plasmids pGAPbfrA

4

0
(1×)zeo and pGAPbfrA(2×)his, respectively, were screened

the native β-fructosidase from T. maritima strain MSB8 (Liebl et al. 1998)
for sucrolytic activity by streaking on SPY plates and incu-
AGA

AGG
CGU

CGA

CGG
CGC

22
Arg

1
bating for 3–5 days at 30 °C. Suc+ colonies turned the
medium color from initial green (pH 6.5) to yellow (pH
CAA

CAG

≤6.0) due to the formation of organic acids during the

Gln

catabolism of glucose and fructose released from sucrose.

Suc− colonies turned the medium color from initial green
UAU
UAC
17
Tyr

(pH 6.5) to blue (pH ≥7.5), as a consequence of the ammo-

nia released by the catabolic oxidation of the nitrogen-
CCU

CCA

CCG
CCC

17

containing carbon sources yeast extract and peptone.

Pro

0
ACU

ACA

ACG
ACC

22
Thr

BfrA production in shaking batch cultures and fed-batch

fermentation
GCU

GCA

GCG
GCC

12
Ala

For shaking batch experiments, 2-l flasks containing 200 ml

of YPG medium were inoculated from a fresh colony of P.
GUU

GUA

GUG
GUC

36

pastoris transformed with either the expression plasmids

Val

pGAPbfrA(1×)zeo, pGAPbfrA(2×)zeo, pGAPbfrA(2×)his,

or the empty vector and incubated for 3 days at 30 °C with
GGU

GGA

GGG
GGC

30
Gly

shaking at 200 rpm. Cells were spun down and invertase

1204
activity was assayed in the undisrupted biomass and the
culture supernatants.
Fed-batch fermentation of the four copy clone PpBfrA
(4×) was performed in a 7.5-l fermenter (Marubishi) con-
taining 3.5 l of the fermentation medium and inoculated
with 0.5 l of a shaking batch culture to an initial OD600
around 2. The operation conditions during the batch phase
were temperature of 30 °C, pH 5.5, agitation 500 rpm, and
aeration 1 vvm. The end of the batch phase upon depletion
of the carbon source occurred at the approximate time of
20 h as judged by a sharp increase of pH and pO2. The
feeding medium [50 % (w/v) cane sugar and 0.5 % (w/v)
yeast extract] was added at a constant flow of 8 ml/h/l of
initial volume. The operation conditions during the feeding
phase were temperature of 30 °C, pH 5.5, agitation 900 rpm,
and aeration 2 vvm. The total fermentation time was 4 days.
Samples (5 ml) were collected at 12-h intervals and assayed
for sucrose-hydrolyzing activity in undisrupted cells and the
culture supernatant.
Cell fractionation and disruption
P. pastoris cells from 5 ml of the shaking batch cultures
were collected by centrifugation, washed in distilled water,
and incubated with lyticase (Sigma) in 1 M sorbitol, 1 mM
EDTA, 0.5 % (w/v) sodium dodecyl sulfate (SDS), and
10 mM sodium citrate (pH 7.5) for 45 min to remove the
cell wall. Spheroplasts were harvested by centrifugation at
10,000×g for 5 min. Periplasmic proteins in the supernatant
were assayed for BfrA activity. Spheroplasts in the pellet
fraction were washed twice in 1.5 M sorbitol, 1 mM EDTA
and resuspended in 2 ml of lysis buffer containing 1 M
sorbitol, 10 mM MgCl2, 2 mM DTT, 1 mM EDTA, 1 mM
phenylmethanesulfonylfluoride (PMSF), 50 mM NaCl, and
50 mM sodium phosphate (pH 6.0) followed by mechanical
lysis by vortexing four times for 1 min in the presence of
acid-washed 500 μm glass beads (Sigma). The cell extract
corresponding to the cytoplasmic fraction was assayed for
invertase activity.
For direct yeast disruption, the collected cells were
washed in distilled water and resuspended in 0.4 ml of
breaking buffer [5 % (v/v) glycerol, 1 mM PMSF, 1 mM
EDTA, and 50 mM sodium phosphate pH 6.0]. After addi-
tion of equal volume of acid-washed 500-μm glass beads
(Sigma), the cells were mechanically lysed by ten cycles of
vortex for 30 s with intermediate incubations on ice for 30 s.
The cell debris and the soluble extract fraction were sepa-
rated by centrifugation and assayed for invertase activity.
Southern blot analysis
Genomic DNA (10 μg) from P. pastoris cells was digested
with EcoRI, electrophoresed in a 0.8 % (w/v) agarose gel in
Appl Microbiol Biotechnol (2013) 97:1201–1212
TBE buffer, transferred to a nylon membrane (Hybond N,
Amersham) by capillarity, and cross-linked to the membrane
by UV. High-specific activity DNA probe was generated by
using a Prime-a-Gene labeling kit (Promega) with [α32P]
dATP (>3,000 Ci mmol−1; Amersham Pharmacia Biotech)
using as template the 1.3-kb EcoRI-SalI fragment compris-
ing the coding region of the bfrA gene in plasmid pGAPbfrA
(1×)zeo. Prehybridization for 4 h and hybridization for 16 h
were performed at 65 °C in a solution of 7 % (w/v) SDS,
1 mM EDTA, 1 % (w/v) bovine serum albumin (BSA) in
100 mM sodium phosphate pH 7.2. High stringency wash-
ing was performed at 65 °C as follows: first wash in 2×
SSC, 0.1 % (w/v) SDS for 10 min; second and third washes
in 1× SSC, 0.1 % (w/v) SDS for 15 min; and fourth and fifth
washes in 0.1× SSC, 0.1 % (w/v) SDS for 15 min. Mem-
branes were then autoradiographed using Kodak X-OMAT
X-K1 film at −70 °C, with intensifying screens.
BfrA purification
Purification of recombinant BfrA containing a polyhistidine
tag was performed by immobilized metal ion affinity chro-
matography (IMAC). The culture supernatant (5 l) of fed-
batch fermented clone PpBfrA(4×) was tenfold concentrated
and dialyzed against the binding buffer (2 mM imidazole
and 100 mM NaCl in 10 mM Tris–HCl pH 7.5) by ultrafil-
tration using a YM-10 membrane (cutoff 10 kDa)
(Millipore). The dialyzed fraction was loaded on a column
containing 10 ml Ni2+ Chelating Sepharose Fast Flow
(Amersham Biosciences) previously equilibrated with the
binding buffer and operating at a flow rate of 4 ml/min.
After passing 60 ml of the washing buffer (20 mM imidaz-
ole and 100 mM NaCl in 10 mM Tris–HCl pH 7.5), the
bound enzyme was recovered with 50 ml of the elution
buffer (200 mM imidazole and 100 mM NaCl in 10 mM
Tris–HCl pH 5.5). The eluted fractions with invertase activ-
ity were pooled and dialyzed against 50 mM sodium acetate
(pH 5.5).
Enzyme assays
Purified BfrA (3 μg/ml) was incubated for 30 min at 60 °C
in 100 mM sodium acetate buffer (pH 5.5) containing
120 mM sucrose, raffinose, melezitose, 1-kestose, and nys-
tose, or 1 % (w/v) inulin and levan. The reducing sugars
released from the hydrolysis of the mentioned substrates
were measured by using the dinitrosalicylic acid (DNSA)
colorimetric method (Miller 1959). The amount of substrate
hydrolyzed is calculated from a calibration curve which
relates the measured absorbance at 540 nm to an equimolar
mixture of glucose and fructose (for sucrose cleavage) or
defined amounts of fructose (for raffinose and fructan cleav-
age). One unit of BfrA is defined as the amount of enzyme
Appl Microbiol Biotechnol (2013) 97:1201–1212
releasing 1 μmol of reducing sugars per min from 120 mM
sucrose or 1 % (w/v) inulin in initial velocity reactions at pH
5.5 and 60 °C.
The effect of pH and temperature on activity was assayed
in 15-min reactions using purified enzyme (3 μg/ml) and the
substrate sucrose (120 mM). The rates of thermoinactivation
were determined by heating the enzyme in 100 mM sodium
acetate buffer (pH 5.5) to 50–90 °C for various time periods
and then reacting with 120 mM sucrose at 60 °C. For
assaying pH stability, the enzyme was preincubated with
buffers of varying pH (4.0–8.0) at 60 °C. Kinetic properties
of glycosylated BfrA were determined with sucrose (10–
200 mM) as substrate at pH 5.5 and 75 °C. Initial velocity
reactions were initiated by the addition of purified enzyme
at a final concentration of 1.8 nM. Aliquots were removed at
scheduled time, chilled quickly on ice, and reactions were
stopped by adding the DNSA solution. The rates were fitted
to the Michaelis–Menten equation.
Protein and carbohydrate analysis
Proteins were quantified by the procedure described by
Bradford (1976) using BSA as a standard. BfrA purity
in the culture supernatant or the purification fraction
was determined by densitometric analysis of 12.5 %
polyacrylamide gels stained with Coomassie brilliant
blue R-250 (Sigma). SDS-PAGE in 12.5 % gels was
performed according to Laemmli (1970). For Western
blot, proteins were electrotransferred on nitrocellulose
membranes (Amersham Pharmacia Biotech) using a
Mini Trans-Blot electrophoresis transfer (BIORAD) at
a constant current of 350 mA for 1.30 h. The presence
of mannose chains in the recombinant BfrA was assayed
using the Digoxigenin Detection Kit (Roche Diagnostics
GmbH) and Galanthus nivalis agglutinin as the
digoxigenin-labeled lectin. B ribonuclease containing
three N-glycosylation sites was used as the positive
control glycoprotein. A single-chain antibody expressed
in E. coli was used as unglycosylated protein (negative
control).
The carbohydrates resulted from the BfrA hydrolase and
transferase reactions at various sucrose concentrations were
quantified by high performance liquid chromatography
(HPLC) using an Aminex HPX-42 C column (0.78 ×
30 cm, BIORAD) equipped with a refractive index detector.
The column temperature was maintained at 85 °C. Water
was used as a mobile phase at a flow rate of 0.6 ml/min.
Samples were appropriately diluted before injection. Fruc-
tose, glucose, sucrose, 1-kestose, and nystose were used as
standards. The products of BfrA reactions on different sub-
strates were separated by thin layer chromatography (TLC)
on silica gel 60 plates (Merck) using acetone/water (9:1, v/v)
as the solvent. After two runs, the fructose-containing sugars
1205
were made visible by spraying the plates with a solution of
3 % (w/v) urea, 1 M phosphoric acid in water-saturated
butanol, and heating at 80 °C for about 20 min.
Endoglycosidase Hf treatment
Purified BfrA (10 μg) was denatured in 100 μl of
0.5 % (w/v) SDS, 1 % (v/v) β-mercaptoethanol at
100 °C for 10 min. After addition of 1/10 volume
1 M sodium citrate buffer pH 5.5 at 25 °C, the sample
was reacted with endoglycosidase Hf (New England Biolabs)
at 0.25 U/μg of total protein at 37 °C for 5 h.
Hydrolysis of cane sugar with biomass reuse in batch
experiments
Recombinant P. pastoris PpBfrA(4×) was assayed as a
whole cell biocatalyst for the inversion of cane sugar at
various temperatures in batch experiments. The yeast cul-
ture was heated to 70 °C for 30 min to kill the cells and
inactivate the thermolabile proteins. The biomass (15 g, wet
weight) was recovered by centrifugation, resuspended in 1 l
of 60 % (w/v) cane sugar in 100 mM sodium acetate buffer
(pH 5.5), and reacted at temperatures ranged from 50 to 70 °C
with constant agitation for 12 h. The reaction was repeated
during other four consecutive 12-h batches with an interme-
diate centrifugation step for recycling of yeast cells. The
residual activity of the harvested biomass was assayed at
initial reaction rates using 240 mM sucrose in 100 mM sodi-
um acetate buffer pH 5.5 at 60 °C.
Bioinformatics analysis
The prediction of potential N-linked glycosylation sites in
the amino acid sequence of BfrA was performed with the
EnsembleGly server (Caragea et al. 2007). The Chimera
software (Pettersen et al. 2004) was used to visualize the
protein 3D structure. The theoretical mass of BfrA secreted
by P. pastoris was calculated on the basis the primary
structure using the software Protein Calculator version 3.3
available at the website http://www.scripps.edu/~cdputnam/
protcalc.html.
Results
Constitutive expression of a synthetic codon-optimized
T. maritima β-fructosidase gene (bfrA) in P. pastoris
The β-fructosidase gene (bfrA) from T. maritima strain
MSB8 encodes a 432-residue polypeptide lacking a putative
signal peptide sequence (Liebl et al. 1998). For optimal
expression in P. pastoris, all codons of the native bfrA gene
1206
were adapted to match those of preferential usage in the host
genome (Table 1) without changing the amino acid
sequence. Plasmid pGAPbfrA(1×)zeo was constructed
so as to carry the synthetic bfrA gene fused in frame
to the S. cerevisiae α-factor signal sequence and a C-terminal
polyhistidine tag under the control of the constitutive GAP
promoter and the AOX1 terminator. The empty vector pGAP-
ZαC (Invitrogen) and plasmid pGAPbfrA(1×)zeo were AvrII-
linearized to favor single crossover recombination at the GAP
locus of P. pastoris strain X-33.
All the zeocin-resistant colonies that resulted from the
integration of plasmid pGAPbfrA(1×)zeo secreted active
BfrA as judged by the shift of the SPY medium color from
initial green (pH 6.5) to yellow (pH ≤6.0). The medium
acidification is indicative of sucrose utilization (see
“Materials and methods”). As expected, the cells elec-
troporated with the empty vector showed the non-
saccharolytic phenotype revealed by alkalinization of
the growth medium. In addition, the bfrA-expressing
yeast could grow on minimal medium containing sucrose
(3 %, w/v) as the sole carbon source, despite the fact that this
enzyme is known to be suboptimally active at 30 °C (Liebl et
al. 1998). In this sense, the acquired Suc+ phenotype offers a
dominant marker for the selection of P. pastoris transformants
and it is also of interest for the yeast fermentation at lower
costs.
The P. pastoris clone with the highest BfrA activity in the
undisrupted cells (24 U/ml of culture, 84 U/g of wet bio-
mass) and the culture supernatant (59 U/ml) during growth
under shaking batch conditions was referred to as PpBfrA
(1×) (Table 2). After yeast disruption, over 95 % of the
invertase activity was detected in the soluble cell extract,
indicating that the recombinant enzyme is not attached to
membranes but secreted into the periplasm where the sub-
strate sucrose readily enters by diffusion. Cell fractionation
experiments confirmed the absence of invertase activity in
the cytoplasmic fraction.
The increase of transgene dosage proportionally enhanced
BfrA yield in shaking batch cultures
P. pastoris clones carrying two, three, and four copies of the
bfrA gene were generated in different transformation steps.
First, the histidine-deficient strain GS115 was electropo-
rated with BglII-digested plasmid pGAPbfrA(2×)his con-
taining two in-tandem copies of the cassette pGAP-bfrA-
tAOX in vector pAO815. Twenty transformants that com-
plemented the his4 mutation and acquired the Suc+ pheno-
type were assayed for periplasmic and extracellular
invertase activities in shaking batch cultures. The elite
clone, named PpBfrA(2×), reached average invertase activ-
ity of 84 U/g in undisrupted biomass (wet weight) and
59 U/ml in the culture supernatant (Table 2). Southern
Appl Microbiol Biotechnol (2013) 97:1201–1212
blot analysis confirmed the integration of the two copies
of the foreign gene in the genome of PpBfrA(2×) with
replacement of the AOX1 locus by a double crossover event
(Fig. 1, lane 3). The hybridizing fragment with the expected
size of 2.47 kb comprises the region between the unique EcoRI
sites of each expression cassette.
A third copy of the expression cassette was integrated at
the resident GAP locus of clone PpBfrA(2×) by introducing
AvrII-linearized plasmid pGAPbfrA(1×)zeo. Simultaneous-
ly, clones with four bfrA copies were generated by retrans-
forming clone PpBfrA(2×) with plasmid pGAPbfrA(2×)zeo
digested with NdeI, a unique site created in the 5′AXO1
region of the vector to incorporate the two extra in-tandem
bfrA copies by single crossover recombination at the chro-
mosomal AOX1 locus. Elite clones PpBfrA(3×) and
PpBfrA(4×), selected by their highest overall invertase
activities among 20 zeocin-resistant retransformants in
both cases, were subjected to Southern blot analysis.
The three hybridization bands observed in clone
PpBfrA(3×) correspond to the sum of the hybridization
patterns of clones PpBfrA(1×) and PpBfrA(2×) (Fig. 1b,
lane 4). As expected, the hybridization pattern of clone
PpBfrA(4×) shows an increased intensity of the 2.47-kb
fragment in comparison with the two other bands
(Fig. 1b, lane 5).
The selected clones containing one to four bfrA copies
and the untransformed strain GS115 showed similar growth
curves and reached comparable biomass yields in shaking
batch cultures. This result indicates that the constitutive
expression of the codon-optimized bfrA gene in P. pastoris
does not cause cell toxicity. The sum of invertase activity in
the collected intact cells and the supernatant fraction
increased from 83 U per ml of growth culture in clone
PpBfrA(1×) to 146 U/ml, 195, and 229 U/ml in the
multicopy clones PpBfrA(2×), PpBfrA(3×), and PpBfrA
(4×), respectively (Table 2). The rate of the periplasmic
versus extracellular activity was proportionally enhanced
with the raise of the transgene dosage (Table 2).
Time course of BfrA production by the four copy clone
PpBfrA(4×) using cane sugar as the main energy source
for growth in fed-batch fermentation
Glucose and glycerol are the carbon sources recommended
to grow P. pastoris for the GAP promoter-driven expression
of recombinant proteins in high cell density cultures
(Cereghino and Cregg 2000). The acquired ability of
clone PpBfrA(4×) to utilize sucrose prompted us to
evaluate the use of cane sugar as a cheaper substrate
for the yeast growth and the constitutive expression of
the bfrA gene at a 5-l fermenter scale under fed-batch
conditions at fixed pH 5.5 and 30 °C. As this temper-
ature is far below the optimal value of 85 °C for BfrA
Appl Microbiol Biotechnol (2013) 97:1201–1212 1207

Table 2 Effect of gene dosage on the invertase activity in the transgenic yeast clones grown under shaking batch conditions

Pichia pastoris Number of expression Biomass yield Periplasmic activity Extracellular Overall Ratio periplasm/
cassettes (g/l) activity (U/ml) activity (U/ml) extracellular
(U/g) (U/ml)

GS115 0 240±19 – – – – –
PpBfrA(1×) 1 257±22 84 24 59 83±10 0.41
PpBfrA(2×) 2 234±27 141 50 96 146±18 0.52
PpBfrA(3×) 3 251±11 208 81 114 195±12 0.71
PpBfrA(4×) 4 232±23 296 102 127 229±17 0.80

The biomass yield (wet weight) and BfrA activity were analyzed in 0.2 l cultures grown in YPG medium for 72 h at 30 °C. The cells from 1 ml
culture samples were spun down at 6,500×g for 10 min, washed in TE buffer and assayed for periplasmic BfrA activity. Supernatant fluids were
assayed for extracellular BfrA activity. Reactions were conducted at initial rates with 120 mM sucrose in 100 mM sodium acetate buffer pH 5.5 at
60 °C. One unit of BfrA is defined as the amount of enzyme hydrolyzing 1 μmol of sucrose per min under the reaction conditions mentioned above.
Periplasmic activity is referred both to BfrA units per gram of wet biomass and units per milliliter of equivalent growth culture. Extracellular
activity is referred to BfrA units per milliliter of culture supernatant. The sum of periplasmic and extracellular activity (overall activity) varied up to
25 % among the 20 assayed clones with the same copy number of the bfrA expression cassette. Values represent means ± standard deviations (n06)
of the elite clones

activity (see below), glycerol at 4 % (v/v) was added in
the culture medium to support growth during the batch
period. Cane sugar was the sole energy source used
during the feeding phase, which was started at 20 h.
The invertase activity in the intact cells and the culture
supernatant increased gradually during the fed-batch fer-
mentation (Fig. 2). The time taken to reach the maximal
cell density (115 g/l dry biomass; 402 g/l wet biomass)
and the peak of overall activity (241 U per ml of
culture) was 72 h. At this point, secretion of BfrA
resulted in periplasmic retention (44 %) and extracellu-
lar release (56 %), while the total enzyme productivity
reached 3,347 U/l/h. Extensive incubation of either the
culture supernatant or the undisrupted cells with sucrose
(60 %, w/v) at pH 5.5 and 60 °C resulted in the
complete hydrolysis of the substrate.
Purification and enzymatic properties of glycosylated BfrA
BfrA represented 84 % of the total proteins in the culture
supernatant of clone PpBfrA(4×) (Fig. 3a). The extracellularly
released enzyme containing a C-terminal His6 tag fusion was
purified by Ni affinity chromatography, with a process recov-
ery of 90 %. The purified protein migrated as two defined
bands on SDS-PAGE with estimated molecular masses of 58
and 53 kDa (Fig. 3a). Both bands reacted with the
digoxigenin-labeled lectin GNA on a gel blot (Fig. 3c), re-
vealing the presence of mannose chains. The removal of N-
linked oligosaccharides by the endoglycosidase H treatment
resulted in a single protein band with a mobility corresponding
to an apparent size of about 50 kDa (Fig. 3d). This value is
consistent with the theoretical mass of 51.2 kDa calculated for
the mature N-deglycosylated BfrA with the polyhistidine tag
on the basis of the primary structure.
The effect of pH and temperature on the activity of glyco-
sylated BfrA was examined in the ranges 4.0–8.0 and 30–
100 °C, respectively (Fig. 4). Sucrose hydrolysis was max-
imum at pH 5.5 and 80–90 °C at initial reactions rates.
The enzyme activity was reduced almost twofold at the pH
values 4.5 and 7.0. Only 6 % of the highest activity was
measured at 30 °C. The purified enzyme retained at least
85 % of its initial activity when preincubated for 24 h at
temperatures up to 70 °C with constant pH 5.5 but became
rather inactive after 12-h incubation at 90 °C. One week
exposition at 60, 70, and 80 °C resulted in residual activities
of 75, 51, and 9 %, respectively. No significant variations in
the enzyme stability were observed for pH values between 5.0
and 7.0 at 60 °C. The addition of 1 mM NaCl, CaCl2, MgCl2,
or MnCl2, as well as the metal ion chelator EDTA did not
significantly influence the enzyme activity, indicating that
glycosylated BfrA does not need a metal cofactor for activity.
The substrate specificity and action mode of glycosylated
BfrA was investigated by incubating the purified enzyme
with sucrose [αglu(1,2)βfru], raffinose [αgal(1,6)αglu
(1,2)βfru], melezitose [αglu(1,2)βfru(3,1)αglu], 1-kestose
[αglu(1,2)βfru(1,2)βfru], nystose [αglu(1,2)βfru(1,2)βfru
(1,2)βfru], inulin [β(2,1)-linked fructan], and levan
[β(2,6)-linked polyfructan] at pH 5.5 and 75 °C. The en-
zyme released fructose from all the assayed substrates, ex-
cept melezitose. The highest specific activity was 932 U/mg
for the reaction on sucrose. The hydrolysis of inulin from
Dahlia tubers (Sigma) and levan from E. herbicola (Sigma)
resulted in the successive release of the terminal fructose
units without the formation of intermediate oligofructans, as
confirmed by TLC analysis (data not shown). The average
rate of inulin versus levan hydrolysis was 16.9±2.1. The
kinetics of glycosylated BfrA was analyzed according to the
Michaelis–Menten model. The values of the kinetic
1208 Appl Microbiol Biotechnol (2013) 97:1201–1212

Fig. 1 Multicopy integration of

a synthetic codon-optimized
a probe
AvrII EcoRI SalI
bfrA gene in P. pastoris. a Par-
tial map of plasmids pGAPbfrA pGAP SP bfrA tAXO1 zeocinr pGAPbfrA(1x)zeo
(1×)zeo, pGAPbfrA(2×)his, and
pGAPbfrA(2×)zeo showing the 1x
bfrA expression cassette and the
5` AOX1 3` AOX1
selection marker. The restriction
pGAP SP bfrA tAXO1 His4+ pGAPbfrA(2x)his
sites in the regions used to
direct homologous recombina-
BglII 2x BglII
tion are indicated with arrows.
b Southern blot analysis of total
DNA from P. pastoris strains 5` AOX1-linker
digested with EcoRI and probed pGAP SP bfrA tAXO1 zeocinr pGAPbfrA(2x)zeo
with the 32P-labeled 1.3-kb
EcoRI-SalI fragment from NdeI 2x
plasmid pGAPbfrA(1×)zeo.
Lanes correspond to strains: 1 b kb 1 2 3 4 5
GS115, 2 PpBfrA(1×), 3
PpBfrA(2×), 4 PpBfrA(3×), 5
PpBfrA(4×). Sizes in kilobase
reflect the migration of unla-
beled 1 kb DNA ladder
(Promega)

10.0
8.0
6.0
5.0
4.0
3.0
2.5
2.0

1.5

parameters kcat, KM, and kcat/KM determined for sucrose
hydrolysis showed values similar to those reported for
Fig. 2 Constitutive expression of the synthetic bfrA gene in P. pastoris
during fed-batch fermentation. The four copy clone PpBfrA(4×) was
grown in a 7.5-l fermenter at fixed pH 5.5 and 30 °C. The fed-batch
phase started at 20 h using cane sugar as the carbon source for growth.
Culture samples (5 ml) were collected at 12-h intervals and centri-
fuged. BfrA activity was assayed in the culture supernatant (extracel-
lular activity) and the undisrupted biomass (periplasmic activity).
Reactions were conducted at initial rates with 120 mM sucrose in
100 mM sodium acetate buffer pH 5.5 at 60 °C. One unit is defined
as the amount of recombinant BfrA hydrolyzing 1 μmol of sucrose per
min under the reaction conditions mentioned above
the unglycosylated enzyme produced in E. coli
(Table 3).
To evaluate the influence of high substrate concen-
tration on the activity of glycosylated BfrA, the purified
enzyme was reacted with sucrose in the 1–80 % (w/v)
concentration range at pH 5.5 and 60 °C (Fig. 5). Initial
velocity measurements show that sucrose was hydro-
lyzed gradually faster as the substrate concentration
was driven from 1 to 10 % (w/v). The enzyme suffered
from slight inhibition at sucrose levels above 20 % (w/
v). Under these conditions, the velocity of sucrose hy-
drolysis decreased and a minor transfructosylation reac-
tion occurred resulting in the synthesis of the
trisaccharide 1-kestose. At the highest substrate concen-
tration (70 %, w/v), the overall BfrA activity (hydrolase
plus transferase) was 1.6-fold reduced and 1-kestose
represented 19 % of the total products at initial reaction
rates. After 8 h incubation, sucrose and the synthesized
FOS were completely hydrolyzed (data not shown),
indicating that high levels of the released products glu-
cose and fructose did not restrain substrate cleavage. A
similar behavior was reported for unglycosylated BfrA
produced in E. coli (Liebl et al. 1998).
Appl Microbiol Biotechnol (2013) 97:1201–1212
a b
kDa 1 2 3 1 2 3 4 1
170
130
95
72
55
43
34
26
Fig. 3 Glycosylation analysis of BfrA purified by IMAC from the
culture supernatant of P. pastoris clone PpBfrA(4×). Proteins were
denatured, separated on SDS–polyacrylamide gel (12.5 %) and
revealed by Coomassie Blue staining (a, b, d) or transferred to a
nitrocellulose membrane and probed with lectin GNA conjugated with
digoxigenin for the detection of mannose chains (c). Lanes in a: 1,
PageRuler prestained protein ladder (Fermentas); 2, culture supernatant
Complete hydrolysis of cane sugar by heat-treated
recombinant P. pastoris cells
BfrA in the periplasmic space rendered 44 % of the total
invertase activity of clone PpBfrA(4×) after fed-batch fer-
mentation. Heat treatment of the yeast culture at 70 °C for
30 min caused cell death. The invertase specific activity of
the non-viable biomass (316 U/g, wet weight) increased 1.2-
fold in comparison to the untreated control (264 U/g, wet
weight), likely due to most favored diffusion of the substrate
and products through the cell wall. The culture incubation at
the high temperature had no effect on the activity of BfrA in
the supernatant fraction. Time course analyses of cane sugar
hydrolysis by the heat-treated intact cells (15 g/l, wet
weight) were conducted with cane sugar concentration of
60 % (w/v) in the temperature range of 50–70 °C and pH 5.5
(Fig. 6). The complete substrate hydrolysis took 10, 12, and
16 h for the reactions at 70, 60, and 50 °C, respectively. The
average productivity in the reaction at 70 °C was 4.4 g of
cane sugar hydrolyzed per gram of wet biomass per hour.
The highest productivity value of 7.2 was reached at the
reaction time interval between 4 and 5 h when almost
half of the initial substrate was remaining. Similar reac-
tion curves were observed when the cells were recov-
ered by centrifugation and reused under the same
reaction conditions (data not shown). The biomass
retained about 55, 91, and 95 % of its original invertase
activity after recycling five times in 12 h batches at 70,
60, and 50 °C, respectively. The repetitive incubation at
70 °C promoted the occurrence of Maillard browning
reactions, while the sugar solution remained colorless at
50 and 60 °C.
1209
c d
2 3 4 kDa 1 2 3 4 5
212
116
66
45
29
20
14
(10 μg of total proteins); 3, purified BfrA (10 μg). Lanes in b and c: 1,
PageRuler prestained protein ladder (Fermentas); 2, purified BfrA
(10 μg); 3, unglycosylated control protein (5 μg); 4, glycosylated B
ribonuclease (5 μg). Lanes in d: 1, broad-range protein marker (BIO-
RAD); 2, purified BfrA (5 μg); 3, endoH-treated BfrA (5 μg); 4,
glycosylated B ribonuclease (5 μg); 5, endoH-treated glycosylated B
ribonuclease (5 μg)
In all experiments, cane sugar diffused into the periplasm
without the need of cell permeabilization treatments. The
intact cells did not react on levan and were tenfold less
efficient to hydrolyze inulin (22 U/g wet biomass,
9.2 U/ml of culture) compared to the free enzyme in
the culture supernatant.
Discussion
The highly thermostable β-fructosidase (BfrA) from T. mar-
itima is attractive for use in biotechnology as a hydrolyser of
sucrose and fructans. BfrA has been produced intracellularly
in E. coli requiring cell disruption to detect activity (Liebl et
al. 1998; Muñoz-Gutiérrez et al. 2009). Neither the native
bacterium nor the recombinant E. coli are well equipped for
the high-scale production and secretion of the enzyme. This
study reports, for the first time, the employ of a yeast host as
a source of secreted BfrA and investigates the effect of N-
glycosylation on the enzyme performance. As other pro-
karyotic proteins, BfrA is not glycosylated in its native host.
The GAP promoter-driven expression of the bfrA gene in
P. pastoris did not cause cell toxicity. We thus decided not to
utilize the inducible AOX1 promoter system, which requires
the use of methanol and it is not recommended for the
production of biocatalysts destined to the food industry.
Secretion of the constitutively expressed BfrA was achieved
using the S. cerevisiae α-factor signal peptide. The gradual
raise of the transgene dosage from one to four copies with
optimized codon composition successfully improved the
enzyme yield in the cell periplasm and the growth medium.
The periplasmic/extracellular rate of invertase activity
1210
Fig. 4 Effect of pH (a) and temperature (b) on the activity of glyco-
sylated BfrA. The influence of pH was examined at 75 °C using
100 mM sodium acetate buffer in the pH range 4.0–6.0 and 100 mM
sodium phosphate buffer in the pH range 6.0–8.0. The influence of
increased proportionally in the multicopy clones (Table 2),
indicating that the transfer across the cell wall is a limiting step
for the complete secretion of glycosylated BfrA in P. pastoris.
The Bacillus subtilis β-exofructanase (SacC) was hypergly-
cosylated during secretion by S. cerevisiae and only 30 % of
its total activity was found in the culture supernatant (Wanker
et al. 1995). Methanol-induced expression in P. pastoris of the
SUC2 gene gave a secreted glycosylated invertase containing
less carbohydrate, due to shorter N-linked oligosaccharides,
than the endoplasmic reticulum form of the enzyme synthe-
sized by the native host S. cerevisiae (Tschopp et al. 1987;
Trimble et al. 1991). On the other hand, the G. diazotrophicus
exolevanase (LsdB), which is functionally and structurally
related to BfrA, SacC, and Suc2, was smoothly released out
of the P. pastoris cells without transit retention in the peri-
plasm. Contrary to its orthologues, LsdB secreted by the yeast
was not a glycoprotein despite containing a potential N-
glycosylation site (Menéndez et al. 2004).
The extracellularly released BfrA migrated in SDS-
PAGE as two defined bands with estimated masses of 58
and 53 kDa, which generated the expected 50 kDa form of
Table 3 Catalytic behavior of glycosylated BfrA
Host KM Vmax kcat kcat/KM
(mM) (μmol min−1 mg−1) (s-1) (M−1 s−1)
P. pastoris 51 3,333 2.8×103 5.6×104
E. colia 64 3,117 2.6×103 4.1×104
The kinetic parameters KM and Vmax for sucrose hydrolysis were
estimated from Lineweaver–Burk plots of initial velocity data mea-
sured at 75 °C. The purified enzyme (1.8 nM) was reacted with the
substrate in the concentration range of 10–200 mM in sodium acetate
buffer (pH 5.5). The products glucose and fructose were quantified
using the DNSA test. The kcat value was calculated assuming an
average molecular mass of 55 kDa for glycosylated BfrA. Shown are
the mean values from ten independent experiments a Results from Liebl
et al. (1998)
Appl Microbiol Biotechnol (2013) 97:1201–1212
temperature was at constant pH 5.5. Purified BfrA (final concentration
3 μg/ml) was incubated with 120 mM sucrose for 15 min. The reduc-
ing sugars released from the substrate were quantified by the DNSA
method. Values represent means±standard deviations (n03)
the protein after deglycosylation with EndoH. This finding
suggests the presence of oligosaccharide chains linked to at
least two Asn residues in the band of higher size. The
encoded BfrA contains in its mature sequence three potential
N-glycosylation sites, following the general rule N-X-T/S,
where X is not proline. The occurrence of N-glycosylation at
positions 142, 340, and 395 was predicted with respective
scoring values of 0.76, 0.81, and 0.85 using the EnsembleGly
web-online server. In general, N-glycosylation occurs at or
just after points of change in secondary structure (Petrescu et
al. 2004), which is not the case of Asn-142 residing in the
middle of a β-strand from Blade III. By contrast, the residues
Asn-340 and Asn-395 occur on β-turns and appear to be more
convenient sites for the occurrence of N-glycosylation. The
crystal structure of BfrA reveals a bimodular arrangement
with an N-terminal catalytic domain folding in a five-bladed
Fig 5 Effect of the substrate concentration on the hydrolase and
transferase activities of glycosylated BfrA at initial reaction rates.
Purified BfrA (final concentration 3 μg/ml) was reacted with sucrose
in the 1–80 % (w/v) concentration range at pH 5.5 and 60 °C. Plotted
values correspond to initial velocity measurements of the fructose
released (square) and the 1-kestose synthesized (circle) in the reaction
mixtures. The reaction products were quantified by HPLC. The exper-
iment was replicated three times. Standard deviation of the means was
below 10 %
Appl Microbiol Biotechnol (2013) 97:1201–1212
Fig. 6 Time course of cane sugar hydrolysis by non-viable cells of
clone PpBfrA(4×) at various temperatures. Intact cells (15 g, wet
weight) from fed-batch cultures were killed by heat treatment and
reacted with 1 l of 60 % (w/v) cane sugar in 100 mM sodium acetate
buffer (pH 5.5) under constant agitation for 16 h in the temperature
range 50–70 °C. Samples (1 ml) were withdrawn every hour and the
release of reducing sugars was monitored using the DNSA method.
The experiment was replicated three times. Standard deviation of the
means was below 10 %. Symbols represent the reaction temperature:
50 °C (triangle), 60 °C (square), and 70 °C (circle)
β-propeller linked to a C-terminal β-sandwich domain of
unknown function (Alberto et al. 2004). Asn-142 is located
in the catalytic β-propeller domain while Asn-340 and Asn-
395 are placed in the β-sandwich module (Fig. 7). Since none
of these residues points toward (or are close) to the active site,
N-glycosylation was predicted not to compromise the enzyme
folding and activity. Indeed, the values of the kinetic constants
kcat and KM here obtained for the reaction on sucrose are in the
same order of magnitude as those reported for BfrA produced
Fig. 7 Mapping of the three
potential N-glycosylation sites in
the monomeric tertiary structure
of BfrA. The β-propeller and
β-sandwich domains are indi-
cated. The color is ramped from
N- (blue) to C- (purple) termi-
nus. The catalytic triad (D17,
D138, and E190) in the active
site and the predicted
N-glycosylation sites (N142,
N340, and N395) are shown in
ball-and-stick representation.
The ribbon diagram was pre-
pared with UCSF Chimera
(Pettersen et al. 2004) using the
crystallographic structure of na-
tive BfrA as input (PDB code:
1UYP)
1211
in E. coli (Liebl et al. 1998; Muñoz-Gutiérrez et al. 2009). The
carbohydrate moiety had no influence on the exo-type activity,
the substrate preference, or the stabilities towards heat and pH.
In this sense, the yeast-produced BfrA kept the ability to
hydrolyze a broad range of substrates containing a terminal
fructose unit, including high DP fructans, which is a common
behavior of bacterial exo-β-fructofuranosidases (Wanker et al.
1995; Blatch and Woods 1993; Bezzate et al. 1994; Menéndez
et al. 2004).
An ideal biocatalyst for fast and continuous production of
invert sugar should stably operate at elevated temperatures
and high sucrose concentrations. The current commercial
invertase from S. cerevisiae is thermolabile and suffers from
substrate inhibition (Bowski et al. 1971). Glycosylated
BfrA, similarly to the enzyme produced in E. coli (Liebl et
al. 1998), is highly resistant to thermal denaturation and
fully hydrolyzed sucrose in concentrated solutions (60–
70 %, w/v), indicating that high levels of the released prod-
ucts glucose and fructose do not restrain substrate cleavage.
BfrA showed transfructosylation activity (synthesis of 1-
kestose) at sucrose concentrations above 20 % (w/v) but
the initial velocity of glucose release (overall activity)
decreased. The inhibitory effect of the substrate sucrose
has been documented in β-fructosidases of different
origins (Goosen et al. 2007; Kotwal and Shankar 2009). The
phenomenon can be attributed to the osmotic effect imposed
by sucrose, the lower activity of water, and the incorporation
of the sucrose molecule as fructosyl acceptor (Seo and Wei
2008; Ritsema et al. 2006)
The four copy clone PpBfrA(4×) developed in this study
constitutes a convenient source of secreted BfrA. The trans-
genic P. pastoris acquired the ability to utilize cane sugar as
β -propeller β-sandwich
1212
a cheap energy source for growth allowing high yield of
biomass (115 g/l dry weight; 402 g/l wet weight) and re-
markable BfrA activity in the culture supernatant (136 U/ml)
and the periplasmic space (264 U/g wet biomass) in a 72-
h fed-batch fermentation. The cells did not require permeabi-
lization for activity on sucrose, totally hydrolysed cane sugar
at the initial concentration of 60 % (w/v), and remained over
90 % active after the five assayed cycles of biomass reuse at
60 °C. Our current efforts are directed to the immobilization of
both the free enzyme and the whole cells with the aim of
developing thermostable biocatalysts for the commercial pro-
duction of invert syrup.
References
Alberto F, Bignon C, Sulzenbacher G, Henrissat B, Czjzek M (2004)
The three-dimensional structure of invertase β-fructosidase from
Thermotoga maritima reveals a bimodular arrangement and an
evolutionary relationship between retaining and inverting glyco-
sidases. J Biol Chem 279:18903–18910
Bezzate S, Steinmetz M, Aymerich S (1994) Cloning, sequencing, and
disruption of a levanase gene of Bacillus polymyxa CF43. J
Bacteriol 176:2177–2183
Blatch GL, Woods DR (1993) Molecular characterization of a fructa-
nase produced by Bacteroides fragilis BF-1. J Bacteriol
175:3058–3066
Bowski L, Saini R, Ryu DY, Vieth WR (1971) Kinetic modeling of the
hydrolysis of sucrose by invertase. Biotechnol Bioeng 13:641–
656
Bradford MM (1976) A rapid and sensitive method for quantification
of microgram quantities of protein utilizing the principle of
protein-dye binding. Anal Biochem 72:248–254
Bretthauer RK, Castellino FJ (1999) Glycosylation of Pichia pastoris-
derived proteins. Biotechnol Appl Biochem 30:193–200
Caragea C, Sinapov J, Silvescu A, Dobbs D, Honavar V (2007)
Glycosylation site prediction using ensembles of Support Vector
Machine classifiers. BMC Bioinforma 8:438
Carlson M, Botstein D (1982) Two differentially regulated mRNAs
with different 5′ ends encode secreted and intracellular forms of
yeast invertase. Cell 28:1455–154
Cereghino JL, Cregg JM (2000) Heterologous protein expression in the
methylotrophic yeast Pichia pastoris. FEMS Microbiol Rev
24:45–66
Cregg JM, Tschopp JF, Stillman C, Siegel R, Akong M, Craig WS,
Buckholz RG, Madden KR, Kellaris PA, Davies GR, Smiley BL,
Cruze J, Torregrossa R, Velicelebi G, Thill GP (1987) High-level
expression and efficient assembly of hepatitis B surface antigen in
the methylotrophic yeast Pichia pastoris. Bio/Technology 5:479–
485
Goosen C, Yuan XL, van Munster JM, Ram AF, van der Maarel
MJ, Dijkhuizen L (2007) Molecular and biochemical charac-
terization of a novel intracellular invertase from Aspergillus
niger with transfructosylating activity. Eukaryot Cell 6:674–
681
Hasal P, Vojtisek V, Cejkova A, Kleezek P, Kofronova O (1992) An
immobilized whole yeast cell biocatalyst for enzymatic sucrose
hydrolysis. Enzyme Microb Technol 14:221–229
Kotwal SM, Shankar V (2009) Immobilized invertase. Biotechnol Adv
27:311–322
Laemmli UK (1970) Cleavage of structural proteins during the assem-
bly of the head of bacteriophage T4. Nature 227:680–685
Appl Microbiol Biotechnol (2013) 97:1201–1212
Lafraya A, Sanz-Aparicio J, Polaina J, Marín-Navarro J (2011)
Fructo-oligosaccharide synthesis by mutant versions of Sac-
charomyces cerevisiae invertase. Appl Environ Microbiol
77:6148–6157
Liebl W, Brem A, Gotschlich A (1998) Analysis of the gene for
β-fructosidase (invertase, inulinase) of the hyperthermophilic
bacterium Thermotoga maritma, and characterization of the
enzyme expressed in Escherichia coli. Appl Microbiol Bio-
technol 50:55–64
Menéndez C, Hernández L, Banguela A, País J (2004) Functional
production and secretion of the Gluconacetobacter diazotrophicus
fructose-releasing exo-levanase (LsdB) in Pichia pastoris.
Enzyme Microb Technol 34:446–452
Miller GL (1959) Use of dinitrosalicylic acid reagent for determination
of reducing sugar. Anal Chem 31:426–428
Muñoz-Gutiérrez I, Rodríguez-Alegría ME, López-Mungía A (2009)
Kinetic behavior and specificity of β-fructosidases in the hydro-
lysis of plant and microbial fructans. Process Biochem 44:891–
898
Paifer E, Margolles E, Cremata J, Montesino R, Herrera L, Delgado JM
(1994) Efficient expression and secretion of recombinant alpha
amylase in Pichia pastoris using two different signal sequences.
Yeast 10:1415–1419
Petrescu AJ, Milac AL, Petrescu SM, Dwek RA, Wormald MR (2004)
Statistical analysis of the protein environment of N-glycosylation
sites: implications for occupancy, structure, and folding. Glyco-
biology 14:103–114
Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM,
Meng EC, Ferrin TE (2004) UCSF Chimera—a visualization
system for exploratory research and analysis. J Comput Chem
25:1605–1612
Reddy A, Maley F (1996) Studies on identifying the catalytic role of
Glu-204 in the active site of yeast invertase. J Biol Chem
271:13953–13958
Ritsema T, Hernández L, Verhaar A, Altenbach D, Boller T, Wiemken
A, Smeekens S (2006) Developing fructan-synthesizing capability
in a plant invertase via mutations in the sucrose-binding box.
Plant J 48:228–237
Seo SK, Wei A (2008) Probing osmotic effects on invertase with L-(−)-
sucrose. Org Biomol Chem 6:3362–3365
Sturm A (1999) Invertases. Primary structures, functions, and roles in
plant development and sucrose partitioning. Plant Physiol 121:1–
7
Trimble RB, Atkinson PH, Tschopp JF, Townsend RR, Maley F (1991)
Structure of oligosaccharides on Saccharomyces SUC2 invertase
secreted by the methylotrophic yeast Pichia pastoris. J Biol Chem
266:22807–22817
Trujillo LE, Arrieta JG, Dafhnis F, García J, Valdés J, Támbara Y,
Pérez M, Hernández L (2001) Fructo-oligosaccharides production
by the Gluconacetobacter diazotrophicus levansucrase expressed
in the methylotrophic yeast Pichia pastoris. Enzyme Microb
Technol 28:139–144
Trujillo LE, Gómez R, Banguela A, Soto M, Arrieta JG, Hernández L
(2004) Catalytical properties of N-glycosylated Gluconaceto-
bacter diazotrophicus levansucrase produced in yeast. Electron J
Biotechnol 7:115–123
Tschopp JF, Sverlow G, Kosson R, Craig W, Grinna L (1987) High
level secretion of glycosylated invertase in the methylotrophic
yeast Pichia pastoris. Bio/Technology 5:1305–1308
Wang Y, Zhang YHP (2009) Overexpression and simple purification of
the Thermotoga maritima 6-phosphogluconate dehydrogenase in
Escherichia coli and its application for NADPH regeneration.
Microb Cell Fact 8:30
Wanker E, Klimgsbichel E, Schwab H (1995) Efficient secretion of
Bacillus subtilis levanase by Saccharomyces cerevisiae. Gene
161:45–49