Báo cáo khoa học: Comparison of functional properties of two fungal antifreeze proteins from Antarctomyces psychrotrophicus and Typhula ishikariensis ppt

Comparison of functional properties of two fungal
antifreeze proteins from Antarctomyces psychrotrophicus
and Typhula ishikariensis
Nan Xiao
1,2
, Keita Suzuki
1,2
, Yoshiyuki Nishimiya
1
, Hidemasa Kondo
1
, Ai Miura
1
, Sakae Tsuda
1,2
and Tamotsu Hoshino
1,2
1 Research Institute of Genome-based Biofactory, National Institute of Advanced Industrial Science and Technology (AIST), Toyohira-ku,
Sapporo, Japan
2 Division of Biological Sciences, Graduate School of Science, Hokkaido University, Kita-ku, Sapporo, Japan
Introduction
Many living organisms, including fungi, have biochem-
ical and ecological strategies to protect themselves
from freezing. Antifreeze protein (AFP) is regarded as
a singular substance in such strategies, which provides
freeze tolerance in several psychrophiles, such as prok-
aryotes and poikilothermic eukaryotes, in a sub-zero
temperature environment [1,2]. AFPs bind to the sur-
face of seed ice crystals generated in an AFP-contain-
ing fluid and inhibit the growth of these crystals [3].
This inhibition causes thermal hysteresis (TH), which

Keywords
fungal AFP; ice growth inhibition;
psychrophile; recrystallization inhibition;
thermal hysteresis
Correspondence
S. Tsuda, Functional Protein Research
Group, Research Institute of Genome-Based
Biofactory, National Institute of Advanced
Industrial Science and Technology (AIST),
2-17-2-1 Tsukisamu-Higashi, Toyohira-ku,
Sapporo 062-8517, Japan
Fax: +81 11 857 8983
Tel: +81 11 857 8983
E-mail:
(Received 8 August 2009, revised
6 November 2009, accepted 10 November
2009)
doi:10.1111/j.1742-4658.2009.07490.x
Antifreeze proteins are structurally diverse polypeptides that have thermal
hysteresis activity and have been discovered in many cold-adapted organ-
isms. Of these, fungal antifreeze protein has been purified and partially
characterized only in a species of psychrophilic basidiomycete, Typhula
ishikariensis. Here we report a new fungal antifreeze protein from another
psychrophile, Antarctomyces psychrotrophicus. We examined its biochemical
properties and thermal hysteresis activity, and compared them with those
of the T. ishikariensis antifreeze protein. The antifreeze protein from
A. psychrotrophicus was purified and identified as an extracellular protein
of approximately 28 kDa, which halved in size following digestion with gly-
cosidase. The A. psychrotrophicus antifreeze protein generated bipyramidal
ice crystals and exhibited thermal hysteresis activity (for example thermal

hysteresis = 0.42 °C for a 0.48 mm solution) similar to that of fish anti-
freeze proteins, while a unique rugged pattern was created on the facets of
the ice bipyramid. The thermal hysteresis activity of the A. psychrotrophicus
antifreeze protein was maximized under alkaline conditions, while that of
the T. ishikariensis antifreeze protein was greatest under acidic conditions.
The T. ishikariensis antifreeze protein exhibited a bursting ice growth nor-
mal to the c-axis of the ice crystal and high thermal hysteresis activity
(approximately 2 °C), as in the case of insect hyperactive antifreeze pro-
teins. From these results, we speculate that the A. psychrotrophicus anti-
freeze protein is very different from the T. ishikariensis antifreeze protein,
and that these two psychrophiles have evolved from different genes.
Abbreviations
AFP, antifreeze protein; AnpAFP, AFP from Antarctomyces psychrotrophicus; ITS, internal transcribed spacer; nfeAFP, AFP from notched-fin
eelpout; PDA, potato dextrose agar; PDB, potato dextrose broth; RI, recrystallization inhibition; T
f,
freezing point; TH, thermal hysteresis;
TisAFP, AFP from Typhula ishikariensis; TisAFP8, an isoform of TisAFP exhibiting a high TH activity; T
m,
melting point.
394 FEBS Journal 277 (2010) 394–403 ª 2009 The Authors Journal compilation ª 2009 FEBS
is the noncolligative depression of the freezing point
(T
f
) of a solution containing ice below its melting point
(T
m
) [4,5]. Within the hysteresis temperature gap,
AFPs modify the ice crystal habit, in that the AFP-
saturated ice crystal forms a unique shape, such as a
hexagonal bipyramid [6]. Recrystallization inhibition

(RI) also results from the adsorption of AFP to ice
crystals [4,7]. In the RI assays the size of the ice crystal
shows hardly any change at temperatures close to
0 °C.
AFPs have been identified in bacteria, plants, inver-
tebrates and fish, and were characterized according to
their structures and TH values [1,2,8]. Fish AFPs have
been grouped into five types (AFPI–IV and AFGP),
and insect AFPs have been grouped into three types
(right- and left-handed b-helices, and a glycine-rich
repeat) [8]. Although more structural information is
needed for grouping plant and bacterial AFPs, they
presumably have structural variations [9–12]. Insect
AFPs are termed ‘hyperactive AFPs’ because their
maximal TH activity is 5–6 °C, which is much higher
than that of fish AFPs (0.5–1 °C) [6]. An observation
of crystal burst, which is normal to the c-axis of the
ice crystal, is another characteristic of hyperactive
AFP [6]. In contrast, most of the plant and bacterial
AFPs exhibit very weak TH (0.01–0.1 °C), although
they possess RI activity [12,13].
Fungal AFPs have been discovered in snow molds
that have pathogenic activities against dormant
plants under snow cover [14–17]. Snow molds include
two major fungal taxa of ascomycetes and basidio-
mycetes and one pseudofungal taxon of oomycetes.
Among them, AFP was only identified in the basid-
iomycetes Coprinus psychromorbidus [15] and Typhu-
la ishikariensis (TisAFP) [16]. Hoshino et al. purified
the TisAFP from the culture medium and cloned the

genes. TisAFP did not exhibit any similarity in pri-
mary structure with other AFPs and therefore it was
considered to be a representative AFP from eukary-
otic microorganisms. Kawahara et al. [18] recently
reported that seven strains of ascomycetes from Ant-
arctica produced extracellular substances that modify
ice crystal shape, although they were not identified
as AFPs. As ascomycota is the largest phylum of
fungi, it may be possible to identify some species
found in freezing environments that adapt by pro-
ducing AFPs.
In this study we performed assays of antifreeze
activity against culture media of a total of 23 species
of ascomycetes, and identified and purified AFP from
an ascomycete collected in Antarctica (AnpAFP). We
believe that comparison of the biochemical characteris-
tics between AnpAFP and TisAFP will provide crucial
information on the molecular diversity and the distri-
bution of AFPs in fungi.
Results
Antifreeze activity assay against ascomycetes
We performed antifreeze activity assays on 1 lL of the
culture medium from each of 23 psychrotrophic asco-
mycetes (Fig. 1). The assay was performed by observa-
tion of the ice-shaping ability (i.e. the formation of
bipyramidal or hexagonal ice crystals), which indicates
adsorption of AFPs to specific ice crystal planes. As
shown in Fig. 1, modification of the formation of ice
Antarctomyces psychrotrophicus
Aniptodera chesapeakensis

Aphanoascus terreus
Ascomycetes sp.
Ascosphaera apis
Cladosporium sp.
Diatrype stigma
Geomyces pannorum
Geotrichum candidum
Graphostroma platystoma
Morchella esculent
Monodictys austrina
Oidiodendron echinulatum
Penicillium camembertii
Pseudeurotium zonatum
Schizosaccharmyces japonicus
Sclerotinia spp.
Taphrina mume
Toly pocladium cylindrosporum
Trichoderma hamatum
Truncatella angustata
Cf.Verticillium sp. 254/HP3
Verticillium sp. olrim438
12
1
2+
+
Fig. 1. A total of 23 species of fungi were tested for ice modifica-
tion (antifreeze) activity. Strains exhibiting ice modification in the
culture medium are indicated by ‘+’, and those that did not exhibit
ice modification are indicated by ‘)’. Picture 1 is the modified ice
crystal of Antarctomyces psychrotrophicus and picture 2 is that of

Penicillium camemberti.
N. Xiao et al. Antifreeze protein from ascomycetous fungus
FEBS Journal 277 (2010) 394–403 ª 2009 The Authors Journal compilation ª 2009 FEBS 395
crystals was observed in two ascomycetes, namely
A. psychrotrophicus and P. camemberti. The bipyrami-
dal shape of the ice crystal formed in A. psychrotrophi-
cus is typical of that observed for fish AFPs [19]. After
2 months of culture of A. psychrotrophicus and P. cam-
emberti, the TH activities, measured in each culture
medium, were 0.3 °C and 0 °C, respectively. Protein
expression in the culture medium of A. psychrotrophi-
cus was monitored, during a 10-week period of culture,
using SDS ⁄ PAGE) (data not shown). Protein bands,
corresponding to AnpAFP, were detected from
2 weeks of culture onwards, and the concentration of
AnpAFP increased consistently with time. The former
result suggests that AFP is secreted into the extracellu-
lar space of A. psychrotrophicus.
Biochemical properties of AnpAFP
Figure 2 shows the biochemical properties of AnpAFP.
SDS ⁄ PAGE (Fig. 2A) followed by silver staining
showed that the molecular mass of AnpAFP is approxi-
mately 28 kDa. The AnpAFP was purified from the
culture medium (lane B) by successive application of
anion-exchange chromatography (lane C), affinity chro-
matography on hydroxyapatite (lane D) and size-exclu-
sion chromatography (lane E). Significantly, MALDI-
TOF ⁄ MS performed for the ‘single-band’ sample of
AnpAFP revealed the presence of nearly 10 different
polypeptides, of approximately 21–22 kDa (Fig. 2B).

Glycoprotein analysis was then performed on the
AnpAFP sample using a glycoprotein staining kit
(Fig. 2C). As shown in the figure, AnpAFP stained as
a single pink band (lane A). It was found that after
incubation with N-orO-glycosidase, AnpAFP
migrates to a position corresponding to nearly half of
the original molecular mass. These data indicate that
AnpAFP is a glycopeptide.
Figure 3 shows the results of amino acid sequence
analysis for AnpAFP, and a 20-residue sequence was
determined as representing its N-terminus. This
sequence showed no significant similarity to the corre-
sponding sequences of known AFPs from bacteria,
plants, invertebrates and fish, but showed slight similar-
ity to the sequence of TisAFP (Fig. 3A). As shown,
seven or eight residues of the 20 were identical, or were
of the same type of amino acid; therefore, it may be
assumed that these two fungal AFPs share sequence
identity to some extent. We examined the amino acid
composition of AnpAFP in further detail, and com-
pared it with that of TisAFP; the results are detailed in
Fig. 3B. In AnpAFP, the most abundant residue was
Asx (17.4%); however, Asx was present at a much
lower frequency (3.4%) in TisAFP. In both the AFPs,
Thr was the second most abundant residue (13.6% and
15.1%). In addition, it was estimated that both AFPs
had relatively low contents of Arg (1.3% and 0.9%),
Met (0.8% and 0.9%) and Tyr (1.9% and 2.4%).
Another significant feature of AnpAFP is the presence
of 2% Cys; however, Cys was not detected in TisAFP.

Ice crystal morphologies of AnpAFP and TisAFP
When a seed ice crystal is formed in water with a
temperature lower than 0 °C, it simply expands to
attain a rounded hexagonal shape when cooled at the
31
21
45
66.2
97.4
116.25
220
A
A
C
B
BCDE
28 kDa
Positive
control
AnpAFP
Digested by
glycosidase.
(N) (O)
ABCD
(kDa)
Fig. 2. (A) SDS–PAGE (12.5%) after each purification step of
AnpAFP. Lane A, molecular mass marker; lane B, culture medium;
lane C, sample after anion exchange chromatography (High-Q col-
umn); lane D, sample after affinity chromatography (hydroxyapa-
tite); lane E, sample after size exclusion chromatography

(sephacryl-100). AnpAFP was purified as a single band on SDS–
PAGE using these purification methods. (B) Mass spectra of puri-
fied AnpAFP. (C) Glycoprotein staining: lane A, AnpAFP (28 kDa);
lane B, positive control (horseradish peroxidase, 45 kDa); lane C,
AnpAFP after incubation with N-glycosidase; lane D, AnpAFP after
incubation with O-glycosidase. After incubation with glycosidase,
AnpAFP migrated to the pink band position that is nearly half of the
original molecular mass of that before incubation with glycosidase.
Antifreeze protein from ascomycetous fungus N. Xiao et al.
396 FEBS Journal 277 (2010) 394–403 ª 2009 The Authors Journal compilation ª 2009 FEBS
rate of 0.05 °CÆmin
)1
between )0.2 and )0.3 °C, as
shown in the photomicroscope images A–D of
Fig. 4I. A similar, but not identical, expansion of ice
crystals was observed with a 0.04 mm solution of An-
pAFP and the same temperature gradient. In this
case, a rugged pattern was observed at the edge of
the rounded hexagonal ice crystal (Fig. 4II). In a con-
centrated solution of AnpAFP (0.48 mm) (Fig. 4III),
the ice crystal was modified into a bipyramidal shape
that is typically observed for moderately active AFPs,
such as fish type I–III AFPs [19]. The only difference
is that the facets of the ice bipyramid have a rugged
pattern. The ice bipyramid formed almost stably on a
downward temperature gradient (Fig. 4III, A–B) and,
finally, exhibited a rapid elongation along the c-axis
below the nonequilibrium freezing temperature
(Fig. 4III, D). Figure 4IV shows the change in the
crystal shape in the presence of 0.5 mm type III AFP

from notched-fin eelpout (nfeAFP) when cooled in
the hysteresis gap. The rapid ice crystal elongation
observed following the slight crystal growth is also
typical of moderately active AFPs (Fig. 4IV, C), and
was ascribed to the binding of AFP to the ice crystal
surface [18]. These data indicate that AnpAFP has an
antifreeze activity similar to that found in fish when
compared on a weight basis.
Figure 4V, A–D, are photomicroscope images of an
ice crystal formed in a 0.05 mm solution of a recom-
binant protein of TisAFP (TisAFP8) when cooled at
0.05 °CÆmin
)1
from )0.2 to )0.3 °C. These images
show the process of rapid growth of the ice crystal,
which is completely different from the elongation of
AnpAFP and fish AFP crystals shown in Fig. 4III
and IV, respectively, but is similar to the bursting
pattern observed for insect hyperactive AFPs [6]. A
dendritic growth pattern is observed in Fig. 4V B–D,
which implies that this ice crystal bursts explosively
in six directions (i.e. ±a1–a3) normal to the c-axis.
Observations of the same pattern of ice crystal burst
have been well documented for hyperactive AFP from
snow fleas, Hypogastrura harveyi [6] and bacterial
AFP from Marinomonas primoryensis [6,20], which is
indicative of the binding of these hyperactive AFPs
to both pyramidal and basal planes of a seed hexago-
nal ice crystal.
Comparison of TH activity between AnpAFP and

TisAFP8
Figure 5A shows the results of a comparison of the
concentration dependence of TH activity among
AnpAFP, TisAFP and fish type III AFP. It also shows
the concentration dependence of TH of a recombinant
TisAFP isoform (denoted TisAFP8), whose amino acid
sequence was recently determined by our group. As
shown, the TH activity of AnpAFP was comparable to
that of fish type III AFP. A maximal TH, of approxi-
mately 0.42 °C, was obtained for a 0.48 mm solution
of AnpAFP. Wild-type TisAFP- showed the same TH
value as AnpAFP but at less than one-tenth of the
AnpAFP concentration. Notably, the TH value of the
recombinant TisAFP8 that produced crystal bursting
(Fig. 4V) was nearly twofold higher than that of the
wild-type TisAFP. The maximum TH value of
recombinant TisAFP8 was 1.9 °C and that of the
wild-type TisAFP was 1.1 °C.
Figure 5B shows the pH-dependence of the TH
value examined for AnpAFP and TisAFP (wild type).
As shown, AnpAFP exhibited increased TH activities
AGLDLGAASX FGALAFEGVA
AGPSAVPLGT AGNYVI LAST
AGPTAVPLGT AGNYAI LAST
AGPTAVPLGT AGNYAI LASA
AGLDLGA ASX FGALAFEGVA
1
20
10
1

20
10
AnpAFP
TisAFP
AnpAFP
Asx
17.4 3.4
Glx
5.0
5.4
Ser
10.8
8.8
Gly
8.6
14.1
His
2.2
0.0
Arg
1.3
0.9
Thr
13.6
15.1
Ala
9.1
16.1
Pro
3.8

4.9
Tyr
1.9
2.4
Val
7.6
7.3
Met
0.8
0.9
Cys
2.0
0.0
Ile
4.0
7.8
Leu
5.4
9.3
Phe
3.6
3.4
Lys
3.9
4.9
A.Acid
AnpAFP TisAFP
A
B
Fig. 3. (A) Alignment of the N-terminal sequences of AnpAFP with

those of three TisAFP isoforms. The 10th residue from the N-termi-
nus of AnpAFP (marked X) could not be conclusively identified
using Edman degradation. The blue and yellow residues of AnpAFP
were present in all TisAFP sequences, and the green residues were
identified in some sequences. (B) Comparison of the amino acid
composition of AnpAFP and TisAFP (see the text).
N. Xiao et al. Antifreeze protein from ascomycetous fungus
FEBS Journal 277 (2010) 394–403 ª 2009 The Authors Journal compilation ª 2009 FEBS 397
in alkaline conditions; the optimal value was obtained
at pH 9.3. By contrast, TisAFP exhibited higher TH
activities in acidic conditions; the optimal value was
obtained at pH 5.8.
RI of AnpAFP and TisAFP
The ability to inhibit recrystallization was assessed
using photomicroscopic observation of ice crystals in
AFP solutions annealed at )6 °C for 3 h, as shown
in Fig. 6. In this experiment, all samples were dissolved
in 100 mm ammonium bicarbonate (pH 7.9) containing
30% (w ⁄ w) sucrose (control solution), which were
immediately frozen entirely by applying a fast cooling
rate (55 °CÆmin
)1
). The samples were then warmed up
to )6 °C and incubated at that temperature for 3 h,
which enabled us to observe the time-dependent change
of the ice crystals in the sample. Figure 6A shows a
photomicroscope image of the control solution before
3 h of incubation. After the 3 h incubation period, the
ice crystals formed in the control showed significant
growth (Fig. 6B). By contrast, the sizes of the ice crys-

tals in a 0.05 mgÆmL
)1
AnpAFP solution remained
small (Fig. 6C); they were much smaller compared with
the control (Fig. 6B). Such a tendency was further
emphasized by increasing the AnpAFP concentration
to 0.1 mgÆmL
)1
(Fig. 6D). These data imply that
AnpAFP possesses RI activity. Figure 6E–G shows the
photomicroscopic observations of RI activity for
TisAFP at various concentrations (0.01–0.1 mgÆmL
)1
).
As shown, more effective RI activity compared with
AnpAFP was indicated by the smaller size of the crys-
tal, which was reduced in size with increasing concen-
trations of TisAFP. To our knowledge, these are the
first RI data obtained for fungal AFPs.
Discussion
The Fungi kingdom has two major divisions: ascomy-
cetes and basidomycetes. AFP (i.e. TisAFP) has been
c
ABCD
I
II
III
IV
V
c

VI
Fig. 4. Photomicroscope images of an ice
crystal in AFP solutions, showing initiation
of rapid growth or bursting in a
0.05 °CÆmin
)1
temperature gradient from
)0.2 to )0.3 °C. I, ice crystal expansion
without AFP. II, ice crystal modified by
0.05 m
M AnpAFP. III, ice crystal burst in
0.48 m
M AnpAFP. IV, ice crystal burst in
moderately active fish type III AFP from
Notched-fin eelpout (NfeAFP) at a concen-
tration of 0.04 m
M. V, ice crystal burst
in 0.05 m
M of the wild-type TisAFP.
VI, a model of ice growth inhibition of
AnpAFP. The hexagonal ice plates are
thought to be stacked with rotations, which
will create a rugged pattern on the facets of
the ice crystal.
Antifreeze protein from ascomycetous fungus N. Xiao et al.
398 FEBS Journal 277 (2010) 394–403 ª 2009 The Authors Journal compilation ª 2009 FEBS
identified only in basidomycetes [2,14–16]. We searched
for new fungal AFPs in the psychrophilic ascomycetes
listed in Fig. 1, and found that only two ascomycete
species, A. psychrotrophicus and P. camemberti, had

antifreeze activities in the culture media. The A. psych-
rotrophicus strains were isolated from the soils of the
maritime and continental areas of Antarctica, suggest-
ing a wide distribution of this fungal species in Antarc-
tica. By contrast, P. camemberti is a common
ascomycete distributed in cold climate regions through-
out the world. These AFPs showed ice-binding activi-
ties, which may contribute to the cold-adaptation
capability of these psychrophilic fungi.
Purified AnpAFP was identified (by SDS ⁄ PAGE) as
a 28 kDa protein and was assumed to be a mixture of
approximately 10 peptides according to the MALDI-
TOF ⁄ MS analysis (Fig. 2). As most of the natural
AFPs from fish, insects, plants and microorganisms
have been reported to consist of five to 10 isoforms
[12,21–25], it could be assumed that AnpAFP consists
of approximately 10 isoforms. Here, AnpAFP, TisAFP
and AFPIII (Fig. 5) consist of a mixture of AFP iso-
forms in A. psychrotrophicus, Typhula ishikariensis and
Zoarces elongatus, respectively. We attempted to deter-
mine the primary sequence of the isoforms of
AnpAFP, and our preliminary Edman-degradation
experiments showed that at least Ala1, Gly2 and Leu5
are highly conserved between the isoforms (data not
shown).
pH
0
0.05
0.10
0.15

0.20
0.25
02468101214
2.5
2.0
1.5
1.0
0.5
0
TisAFP AnpAFP
TH (
°
C)
TH (
°
C)
1.5
1.0
0
0 0.12 0.24 0.36 0.48
TH (
°
C)
2.0
TisAFP8
TisAFP wild type
type III AFP(NfeAFP)
AnpAFP
0.5
A

B
Fig. 5. (A) TH activity of AnpAFP compared with wild-type TisAFP,
recombinant TisAFP8 and fish type III AFP as a function of concen-
tration (m
M). The AFPs include: recombinant TisAFP isoform 8
(square); wild-type TisAFP (triangle); natural AnpAFP (diamond,
dashed line); and natural type III fish AFP (circle, solid line). (B)
Effect of pH on the TH activities of natural AnpAFP (filled circle, solid
line) and TisAFP (open circle, dashed line). The vertical line at the left
shows the TH activity of 50 l
M TisAFP. The vertical line at the right
shows the TH activity of AnpAFP at the same concentration.
A
B
C
D
E
F
G
Fig. 6. Photomicroscope images showing RI activity of AFPs at )6 °C. All samples were dissolved in 100 mM ammonium bicarbonate (pH 7.9)
containing 30% (w ⁄ w) sucrose (control solution), which were immediately frozen entirely by applying a fast cooling rate (55 °CÆmin
)1
). The sam-
ples were then warmed up to )6 °C and incubated at that temperature for 3 h. Panel A shows a photomicroscope image of the control solution
before the 3 h incubation. The other panels are the images after the 3 h incubation observed for (B) the control solution, (C) 0.05 mgÆmL
)1
of
AnpAFP, (D) 0.1 mgÆmL
)1
of AnpAFP, (E) 0.01 mgÆmL

)1
of TisAFP, (F) 0.05 mg mL
)1
of TisAFP and (G) 0.1 mgÆmL
)1
of TisAFP.
N. Xiao et al. Antifreeze protein from ascomycetous fungus
FEBS Journal 277 (2010) 394–403 ª 2009 The Authors Journal compilation ª 2009 FEBS 399
For AnpAFP, glycosylation was suggested by
SDS ⁄ PAGE. The size of the molecule was halved after
incubation with glycosidase (Fig. 2C). AnpAFP
reacted with both N- and O-glycosidases. These results
suggest that half of the molecular weight of AnpAFP
is glycan and that both N- and O-linked glycosylation
occur in AnpAFP. If the isoforms of AnpAFP are dif-
ferently glycosylated, several peaks might be obtained
in the MALDI-TOF ⁄ MS spectrum. We could not
obtain any information about the ice-binding ability of
the glycan part of AnpAFP. Direct involvement of gly-
can in ice binding has been suggested only for fish
AFGP consisting of an -Ala-Thr-Ala- repeating unit
that links to a disaccharide, b-d-galactosyl-(1,3)-a-
N-acetyl-d-galactosamine [26]. N-linked glycosylation
was also suggested for AFP from carrots [27] and for
aCa
2+
-dependent species of fish type II AFP [28].
There is no involvement of glycans in ice binding of
these proteins because no significant change in the ice-
binding activity was detected when recombinant ver-

sions of these proteins (without glycan) were analyzed.
Both AnpAFP and TisAFP exhibited a high content
of Thr residues (Fig. 3B). Thr is generally a key resi-
due in the ice-binding ability of AFP. Insect b-helical
AFP is composed of a -Thr-X-Thr- repeat motif,
where the OH groups in the motif are arranged in line
to bind with the ice crystal [29]. In fish AFGP, Thr is
conjugated with a disaccharide that is directly involved
in ice binding [26]. Participation of Thr in ice binding
was also suggested for fish type I–III AFPs; TH activ-
ity was diminished or lost when Thr was replaced with
other amino acids [22,23,28,30–33]. We hence speculate
that Thr residues are involved in the ice-binding site of
AnpAFP and TisAFP. Regarding the other amino
acids, the contents of Ser, Met, Val and Phe are simi-
lar between AnpAFP and TisAFP, while the contents
of Gly, Ala, Leu and Ile differ (Fig. 3B). It is worth
noting that AnpAFP contains 2% Cys, while TisAFP
does not. In SDS ⁄ PAGE with b-mercaptoethanol, a
large molecular mass band (> 100 kDa, data not
shown) was seen together with the normal 28 kDa
band, which is ascribable to the polymerization of this
peptide. This result suggests that Cys residues natively
form an intramolecular disulfide bond in AnpAFP,
and that they are formed between the molecules in the
presence of a reductant.
The concentration dependence of TH was similar
between AnpAFP and AFPIII (Fig. 5). AnpAFP fur-
ther showed the ability to inhibit recrystallization
(Fig. 6), depending on the concentration of this pep-

tide. These data suggest that AnpAFP can inhibit ice
growth at a level similar to that of fish type III AFP.
As A. psychrotrophic survives freezing and thawing,
AnpAFP might provide freeze tolerance through the
effective RI activity. From the comparison of pH
dependence of TH activity between AnpAFP and
TisAFP8 (Fig. 5B), it was found that the function of
AnpAFP is maximized at an alkaline pH (approxi-
mately pH 9), while that of TisAFP8 was greatest
under acidic conditions (approximately pH 5). A
plausible explanation for this result is that either the
ice-binding site or the whole molecule of each protein
is denatured and loses its activity at different pH
ranges. A large difference was found in the content of
Asx between AnpAFP (17%) and TisAFP (4%); how-
ever, the content of Glx was similar (approximately
5%) (Fig. 3B). AFPIII exhibited almost the same
degree of TH activity as AnpAFP in the pH range
2–13 [23]. A similar result was reported for insect
hyperactive AFP [Rhagium inquisitor (Ri)AFP] [34]. It
seems that AnpAFP functions across a relatively wide
range of pH compared with TisAFP, although its TH
activity is lower. TisAFP also lost TH activity after
incubation at 30 °C, while AnpAFP did not (prelimin-
ary results; data not shown). All of these results sug-
gest a large difference between AnpAFP and TisAFP
in their basic biochemical properties.
In a dilute solution of AnpAFP (0.04 mm), a unique
indistinct pattern (a rounded hexagonal shape) was
observed on the edge of the seed ice crystal (Fig. 4II).

The ice crystal expanded slightly, retaining its mor-
phology when cooled, similar to ordinary ice crystals
in the absence of AFP (Fig. 4I). In a more highly con-
centrated solution of AnpAFP (0.48 mm), the crystal
grew into a bipyramid, similar to that observed in the
presence of fish AFPs, but differed in that it had a
unique rugged pattern on its facets (Fig. 4III and VI).
A plausible explanation for the AFP-induced ice bipyr-
amid formation has been described, with illustrations,
in Takamichi et al. [35]. Briefly, binding of AFP to the
prism planes, and the generation of a smaller ice
nucleus on the basal planes of a hexagonal seed ice
crystal, are thought to cause the successive stacking of
hexagonal ice plates on the basal plane. As a conse-
quence, pyramidal planes are created by the adsorption
of AFPs, and the 12 equivalent planes form the bipy-
ramidal ice crystal. It is highly likely that a similar
type of ice binding occurs in AnpAFP because a simi-
lar ice bipyramid formed (Fig. 4III) and its TH value
was comparable to that of AFPIII (Fig. 5A). We
assume that an exceptional feature of ice growth inhi-
bition of AnpAFP is that the hexagonal ice plates are
stacked with rotations, as illustrated in Fig 4.VI. This
hypothesis explains the observed indistinct pattern at
the edge of the hexagonal ice crystal seed (Fig. 4II), as
well as the formation of a rugged pattern on the facets
Antifreeze protein from ascomycetous fungus N. Xiao et al.
400 FEBS Journal 277 (2010) 394–403 ª 2009 The Authors Journal compilation ª 2009 FEBS
of the ice bipyramid (Fig. 4IV). Obviously, additional
experiments and consideration will be necessary to

verify this hypothesis. Nevertheless, we believe that
our data and hypothesis significantly contribute to the
understanding of the detailed functioning of AFP, as
these observations have never been reported for any
other species of AFP.
In the solution of recombinant TisAFP8, a seed ice
crystal maintains its size and shape upon cooling in the
hysteresis gap (Fig. 4V, A–B) and undergoes a crystal
burst below the nonequilibrium freezing temperature
(Fig. 4V, C–D). The dendritic growth pattern suggests
that the direction of the crystal burst is normal to the c-
axis, which is typical of hyperactive AFPs from insects
(e.g. snow flea, spruce budworm, etc.) and bacteria
(M. primoryensis) [6]. These observations imply that a
crystal burst always occurs from the prism plane with
hyperactive AFPs and is ascribed to the binding of
hyperactive AFPs to both the prism and the basal plane.
For most fish AFPs, the burst occurs from the tip of the
ice bipyramid (basal plane, see Fig. 4IV) because of a
lack of binding of fish AFPs to the basal plane. The
exceptional level of TH activity (Fig. 5A), and the
strong ability to inhibit ice growth (Fig. 4V), support
the exceptional antifreeze activity of TisAFP8, which is
comparable to that of hyperactive insect antifreezes. It
should be noted that the TH value of the TisAFP natu-
ral product (i.e. isoform mixture) was approximately
half of that of the TisAFP8 isoform, suggesting a very
low percentage of TisAFP8 in the TisAFP sample. Ti-
sAFP has no cysteines and no -Thr-X-Thr- repetitions,
and therefore possesses no similarity in amino acid

sequence to insect hyperactive AFPs [36]. 3D structural
determinations and site-directed mutagenesis experi-
ments of TisAFP8 are currently in progress, and should
be helpful in revealing the mechanism of binding of this
exceptionally strong ice growth inhibitor.
In summary, we discovered a new fungal AFP
(AnpAFP) from a psychrophile, A. psychrotrophicus.
AnpAFP is an extracellular protein of 28 kDa, whose
size is halved following digestion with glycosidase.
AnpAFP generates bipyramidal ice crystals and exhib-
its TH activity that is maximized under alkaline con-
ditions, and a unique rugged pattern appeared on the
facets of the ice bipyramid. AnpAFP also has the
ability to inhibit recrystallization. There is similarity
in the N-terminal residues between AnpAFP and
another fungal AFP from a basidomycete (TisAFP).
However, TisAFP uniquely exhibited a high TH value
and a pattern of ice crystal bursting similar to that of
the insect hyperactive AFPs. These two AFPs from
basidiomycetes and ascomycetes might have evolved
independently.
Experimental procedures
Preparations of fungal strains and media
Five species of fungi were isolated from various terrestrial
materials (mosses, soils and algal mats), which were
collected in 1996 near Great Wall station on King George
Island, South Shetland Islands, and from Zhongshan sta-
tion in Larsemann Hills, Prydz Bay, East Antarctica. Fungi
were also collected in 2007 near Soya coast, Lutzow-Holm
Bay, East Antarctica. All samples were stored at )20 °C,

and were transferred to 4 °C on potato dextrose agar plates
(PDA; Difco) containing 0.1 mgÆmL
)1
of ampicillin for cul-
tivation, and were then stored for 2 months at 4 °C. Fungal
colonies created on the surface of PDA were cultured at
)1 °C in potato dextrose broth (PDB). The preliminary
assay showed that A. psychrotrophicus NBRC 105511
(KS-1), NBRC 105512 (Z-23) and NBRC 105513 (Syw-1)
had extracellular antifreeze activities, and these strains were
chosen for further experiments. We collected the strain
KS-1 from the soils near Great Wall station on King
George Island, Z-23, near Zhongshan station in Larsemann
Hills, and collected Syw-1 from the soils in Kizahashi-
hama, Skarvsnes, Soya coast.
DNA was extracted from the A. psychrotrophicus strains
KS-1, Z-23 and Syw-1 using the ISOPALNT II protocol
(Nippon Gene Co., Ltd., Tokyo, Japan). The internal
transcribed spacer (ITS) region of genomic recombinant
DNA was amplified using the primer pairs ITS1-F
(5¢-CTTGGTCATTTAGAGGAAGTAA) and ITS4-B
(5¢-CAGGAGACTTGTACACGGTCCAG), according to
Gardens and Bruns (1993) [37], with modifications; they
used KOD plus polymerase (Toyobo Co., Ltd., Tokyo,
Japan) instead of Taq polymerase. The PCR product was
purified using the QIAquick PCR Purification Kit (Qiagen
GmbH, Hilden, Germany); and the amino acid sequence
was determined on an ABI PRISM 3100 Genetic Analyzer
(Applied Biosystems, CA, USA) using the primer ITS1-F.
Determination of antifreeze activity

The antifreeze activity of the culture medium of each strain
was examined by observation of ice crystal morphology
using our photomicroscope system [38] with a Leica DMLB
100 photomicroscope (Leica Microsystems AG, Wetzlar,
Germany) equipped with a Linkam LK600 temperature
controller (Linkam, Surrey, UK). For measuring TH activ-
ity, the purified sample of AnpAFP was dissolved in water,
and that of TisAFP was dissolved in 100 mm ammonium
bicarbonate (pH 7.9). The culture medium or AFP solution
was momentarily frozen by lowering it to )25 °C, and then
it was warmed up to almost 0 °C on the sample stage to
create an ice crystal seed in the solution. This sample
solution was then cooled down or warmed up slightly to
N. Xiao et al. Antifreeze protein from ascomycetous fungus
FEBS Journal 277 (2010) 394–403 ª 2009 The Authors Journal compilation ª 2009 FEBS 401
observe growth initiation or melting of ice crystals to deter-
mine the nonequilibrium T
f
and T
m
values, respectively. In
detail, T
f
was considered as the temperature at which ice
crystal growth occurs from the bipyramidal tip for
AnpAFP, and from the edges of the ice disc for TisAFP.
The difference between T
m
and T
f

was determined to be the
TH value. All photomicroscope images and movies were
recorded using a color-video 3CCD camera (Sony, Tokyo,
Japan) and a personal computer.
Purification of AFP from A. psychrotrophicus
(AnpAFP)
A 1-L culture of A. psychrotrophicus was prepared by inoc-
ulating PDB with 10 mycelial discs (5 mm in diameter).
These were cut from the margin of an actively growing col-
ony on a PDA plate. The culture was maintained at )1 °C
for 2 months without shaking. We removed the mycelia by
filtration, and the resulting culture medium (500 mL) was
dialyzed against 25 mm Tris ⁄ HCl buffer (pH 8.0). The dial-
ysate was applied to an Econo-pac High Q column (Bio-
Rad, CA, USA) equilibrated with the same buffer. The
fraction that exhibited antifreeze activity was eluted with
the above buffer containing 1 m NaCl. The antifreeze active
fraction was dialyzed against 10 mm phosphate buffer (pH
7.2), and the dialysate was loaded onto an Econo-pac
CH-II column. The flow-through fraction was analyzed for
antifreeze activity and was concentrated to 1 mL using a
Microcon Ultracel YM-10 (Bedford, MA, USA). The con-
centrated sample was loaded onto Sephacryl-100 gel and
eluted using the same buffer (10 mm phosphate buffer,
pH7.2). All steps of the purification procedure were carried
out at 4 °C.
Purification of TisAFP from the culture medium was per-
formed as described previously [15]. cDNA encoding
TisAFP8 (accession number Q76CE8 in DDBJ ⁄ EMBL ⁄
GenBank) was inserted into the chromosome of a methylo-

trophic yeast (Pichia partoris) using pPICZ from the Easy-
SelectÔ Pichia Expression Kit (Invitrogen Co., CA, USA).
The methylotrophic yeast transformant obtained was cul-
tured with methanol (buffered minimal methanol-complex
medium) for 5 days at 25 °C. The culture product was pre-
pared by centrifugation (18 590 g, 30 min, 4 °C) and then
dialyzed against 50 mm Tris ⁄ HCl (pH 8.5) containing
0.1 mm phenylmethanesulfonyl fluoride. The dialysate was
applied to a 5-mL Ni-nitrilotriacetic acid Superflow column
(Qiagen GmbH) equilibrated with the same buffer, and
then eluted with 100 mm imidazole at 0.2 mLÆmin
)1
. The
fraction was dialyzed against 10 mm acetic acid buffer (pH
3.0) containing 1 m m EDTA and 0.1 mm phen-
ylmethanesulfonyl fluoride. The dialyzed protein was
applied onto an Econo-pac High S column and then eluted
with the same buffer containing 0.1 m NaCl. The fraction
containing TisAFP was dialyzed against 100 mm ammo-
nium bicarbonate and then TH activity was measured.
Protein analyses of AnpAFP
The molecular mass of AnpAFP was measured using
SDS ⁄ PAGE and MALDI-TOF ⁄ MS (Voyager DEÔ PRO;
Applied Biosystems). The N-terminal amino acid sequence
of purified AnpAFP was determined using an ABI 491 Pro-
tein Sequencer (Applied Biosystems). The N-terminal amino
acid sequence of AnpAFP was deposited in Uni-
protKB ⁄ Swiss-Prot (accession number P86268). The amino
acid analyses were performed by the Center of Instrumental
Analysis, Hokkaido University, Sapporo. Western blot

analyses of AnpAFP were carried out using rabbit poly-
clonal anti-TisAFP, which was prepared in our laboratory.
Purified fish type III AFP, prepared in our laboratory, was
used as the negative control for western blot analyses.
Glycans in AnpAFP were detected using a glycoprotein
staining kit (GelCodeÔ; TaKaRa Bio Inc., Tokyo, Japan)
after SDS ⁄ PAGE. Horseradish peroxidase and soybean
trypsin inhibitor were used as positive and negative
controls, respectively. Glycans were hydrolyzed by N-glyco-
sidase A and O-glycosidase (Roche Diagnostics GmbH,
Mannheim, Germany), under optimal conditions for each
enzyme, at 37 °C for 24 h.
References
1 Duman JG (2001) Antifreeze and ice nucleator pro-
teins in terrestrial arthropods. Annu Rev Physiol 63,
327–357.
2 Duman JA & Olsen TM (1993) Thermal hysteresis pro-
tein activity in bacteria, fungi and phylogenetically
diverse plants. Cryobiology, 30, 322–328.
3 Raymond JA & DeVries AL (1977) Adsorption inhibi-
tion as a mechanism of freezing resistance in polar
fishes. Proc Natl Acad Sci USA 74, 2589–2593.
4 Knight CA, Devries AL & Oolman LD (1984) Fish
antifreeze protein and the freezing and recrystallization
of ice. Nature, 308, 295–296.
5 Fletcher GL, Kao MH & Fourney RM (1986) Anti-
freeze peptides confer freezing resistance to fish. Can J
Zool 64, 1897–1901.
6 Scotter AJ, Marshall CB, Graham LA, Gilbert JA,
Garnham CP & Davies PL (2006) The basis for hyper-

activity of antifreeze proteins. Cryobiology 53, 229–239.
7 Knight CA, Hallett J & DeVries AL (1988) Solute
effects on ice recrystallization: an assessment technique.
Cryobiology 25, 55–60.
8 Jia Z & Davies PL (2002) Antifreeze proteins: an unu-
sual receptor-ligand interaction. Trends Biochem Sci 27,
101–106.
9 Barrett J (2001) Thermal hysteresis proteins. Int J
Biochem Cell Biol 33(2), 105–117.
10 Garnham CP, Gilbert JA, Hartman CP, Campbell
RL, Laybourn-Parry J & Davies PL (2008) A Ca
2+
-
dependent bacterial antifreeze protein domain has a
Antifreeze protein from ascomycetous fungus N. Xiao et al.
402 FEBS Journal 277 (2010) 394–403 ª 2009 The Authors Journal compilation ª 2009 FEBS
novel beta-helical ice-binding fold. Biochem J 411,
171–180.
11 Urrutia ME, Duman JG & Knight CA (1992) Plant
thermal hysteresis proteins. Biochim Biophys Acta 22,
199–206.
12 Griffith M & Yaish MW (2004) Antifreeze proteins in
overwintering plants: a tale of two activities. Trends
Plant Sci 9, 399–405.
13 Raymond JA, Fritsen C & Shen K (2007) An ice-bind-
ing protein from an Antarctic sea ice bacterium. FEMS
Microbiol Ecol 61, 214–221.
14 Snider CS, Hsiang T, Zhao G & Griffith M (2000) Role
of ice nucleation and antifreeze activities in pathogene-
sis and growth of snow molds. Phytopathology 90, 354–

361.
15 Hoshino T, Kiriaki M & Nakajima T (2003) Novel
thermal hysteresis proteins from low temperature basid-
iomycete, Coprinus psychromorbidus. Cryo Letters 24,
135–142.
16 Hoshino T, Kiriaki M, Ohgiya S, Fujiwara M, Kondo
H, Nishimiya Y, Yumoto I & Tsuda S (2003) Antifreeze
proteins from snow mold fungi. Can J Bot 81,
1175–1181.
17 Hoshino T (2005) Ecophysiology of snow mold fungi.
Plant Biol 6, 27–35.
18 Kawahara H, Takemura T & Obata H (2006) Function
analysis and screening of antifreeze material from fungi.
Cryobio Cryotech 52, 151–155.
19 Davies PL & Hew CL (1990) Biochemistry of fish anti-
freeze proteins. FASEB J 4, 2460–2468.
20 Gilbert JA, Davies PL & Laybourn-Parry J (2005)
A hyperactive, Ca
2+
-dependent antifreeze protein in an
Antarctic bacterium. FEMS Microbiol Lett 245, 67–72.
21 Gourlie B, Lin Y, Price J, DeVries AL, Powers D &
Huang RC (1984) Winter flounder antifreeze proteins: a
multigene family. J Biol Chem 259, 14960–14965.
22 Liu Y, Li Z, Lin Q, Kosinski J, Seetharaman J,
Bujnicki JM, Sivaraman J & Hew CL (2007) Structure
and evolutionary origin of Ca
2+
-dependent herring type
II antifreeze protein. PLoS ONE 2, e548.

23 Chao H, So
¨
nnichsen FD, DeLuca CI, Sykes BD &
Davies PL (1994) Structure-function relationship in the
globular type III antifreeze protein: identification of a
cluster of surface residues required for binding to ice.
Protein Sci 3, 1760–1769.
24 Duman JG, Li N, Verleye D, Goetz FW, Wu DW,
Andorfer CA, Benjamin T & Parmelee DC (1998)
Molecular characterization and sequencing of antifreeze
proteins from larvae of the beetle Dendroides canaden-
sis. J Comp Physiol B 168, 225–232.
25 Graham LA, Qin W, Lougheed SC, Davies PL &
Walker VK (2007) Evolution of hyperactive,
repetitive antifreeze proteins in beetles. J Mol Evol 64,
387–398.
26 Harding MM, Anderberg PI & Haymet AD (2003)
‘Antifreeze’ glycoproteins from polar fish. Eur J Bio-
chem 270, 1381–1392.
27 Worrall D, Elias L, Ashford D, Smallwood M, Sidebot-
tom C, Lillford P, Telford J, Holt C & Bowles D (1998)
A carrot leucine-rich-repeat protein that inhibits ice
recrystallization. Science 282, 115–117.
28 Yasui M, Takamishi M, Miura A, Nishimiya Y, Kondo
H & Tsuda S (2008) Hydroxyl groups of threonines
contribute to the activity of Ca
2+
-depdendent type II
antifreeze protein. Cryobio Crotech, 54, 1–8.
29 Doucet D, Tyshenko MG, Kuiper MJ, Graether SP,

Sykes BD, Daugulis AJ, Davies PL & Walker VK
(2000) Structure-function relationships in spruce bud-
worm antifreeze protein revealed by isoform diversity.
Eur J Biochem 267, 6082–6088.
30 Wen D & Laursen RA (1992) A model for binding of
an antifreeze polypeptide to ice. Biophys J 63, 1659–
1662.
31 Haymet AD, Ward LG, Harding MM & Knight CA
(1998) Valine substituted winter flounder ‘antifreeze’:
preservation of ice growth hysteresis. FEBS Lett 430,
301–306.
32 Zhang W & Laursen RA (1998) Structure-function rela-
tionships in a type I antifreeze polypeptide. The role of
threonine methyl and hydroxyl groups in antifreeze
activity. J Biol Chem 273, 34806–34812.
33 Jia Z, DeLuca CI, Chao H & Davies PL (1996) Struc-
tural basis for the binding of a globular antifreeze pro-
tein to ice. Nature 384, 285–288.
34 Kristiansen E, Ramløv H, Hagen L, Pedersen SA,
Andersen RA & Zachariassen KE (2005) Isolation and
characterization of hemolymph antifreeze proteins from
larvae of the longhorn beetle Rhagium inquisitor (L.).
Comp Biochem Physiol B Biochem Mol Biol 142, 90–97.
35 Takamichi M, Nishimiya Y, Miura A & Tsuda S (2009)
Fully active QAE isoform confers thermal hysteresis
activity on a defective SP isoform of type III antifreeze
protein. FEBS J 276, 1471–1479.
36 Graether SP & Sykes BD (2004) Cold survival in
freeze-intolerant insects: the structure and function of
beta-helical antifreeze proteins. Eur J Biochem 271,

3285–3296.
37 Gardens M & Bruns TD (1993) ITS primers with
enhanced specificity for basidiomycetes–application to
the identification of mycorrhizae and rusts. Mol Ecol 2,
113–118.
38 Takamichi M, Nishimiya Y, Miura A & Tsuda S (2007)
Effect of annealing time of an ice crystal on the activity
of type III antifreeze protein. FEBS J. 272, 482–492.
N. Xiao et al. Antifreeze protein from ascomycetous fungus
FEBS Journal 277 (2010) 394–403 ª 2009 The Authors Journal compilation ª 2009 FEBS 403