| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 34, 32307-32312, August 22, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|| ¶
From the
Protein Structure Research Group and
||Molecular Adaptation Research Group, Institute
for Biological Resources and Functions, National Institute of Advanced
Industrial Science and Technology, 2-17-2-1 Tsukisamu-Higashi, Toyohira,
Sapporo 062-8517 and Division of Biological
Sciences, Graduate School of Science, Hokkaido University, N10W8 Kita, Sapporo
060-0810, Japan
Received for publication, April 28, 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
An elegant explanation for the non-colligative depression of Tf of AFP was made by "adsorption-inhibition mechanism," in which accumulations of AFPs onto the growing surfaces of ice crystal result in the creations of numbers of convex ice surfaces at the limited open spaces on the ice between the bound AFPs (5). The free energy of the convex ice surfaces becomes larger with the increase of the surface curvature (so-called Kelvin effect (6)). Therefore, the further binding of water molecules onto the convex ice surfaces becomes energetically unfavorable, leading to the suppression of the ice crystal growth by increasing the number of bound AFPs. This irreversible ice binding model was then modified to a semi-reversible model (7, 8) implicated in the kinetics of AFP-induced ice growth inhibition, which includes the following AFP adsorption steps (9): (i) attachment to the ice-water interface; (ii) rearrangement of adsorbed molecules by diffusion, reorientation, and/or conformational change; and (iii) detachment from the interface. It was assumed that such progressive steps of AFP adsorption are highly correlated with the local concentration of AFPs around the ice crystal surface (2).
Here we focused a 7-kDa small globular type III AFP (10) and its genetically synthesized dimer, trimer, and tetramer to investigate the AFP multimerization effects on the Tf depression and the ice crystal morphology. The type III AFP has been studied extensively for the species identified from Macrozoarces americanus (Atlantic ocean pout) and appeared to possess a diversity in amino acid sequences by multiple gene encodings (10, 11). Determinations of the three-dimensional x-ray structures (12, 13) and the high precision NMR structure (14) revealed that type III AFP constructs a globular shape characterized by "2-fold symmetric motif" or a "pretzel fold." It also revealed that the putative ice-binding residues (e.g. Gln9, Asn14, Thr15, Thr18, and Gln44), which were identified by observing the effect of site-directed mutations of these residues on the TH activity, are located in the planar surface of the type III AFP with tight atom packing, so as to undergo a specific interaction with the ice prism plane (15). In 1995, RD3 was identified as an exceptional isomer of the type III AFP from Lycodichthys dearborni (Antarctic ocean pout) and appeared to comprise the two type III AFP domains connected in tandem through a 9-residue linker sequence (-Asp-Gly-Thr-Thr-Ser-Pro-Gly-Leu-Lys-) (16). The three-dimensional structure of RD3 was determined using multidimensional NMR techniques (17), which revealed the following: (i) the N- and C-domains construct the ordinary type III AFP structures that are laterally oriented with an angle of 32 ± 12°, and (ii) a flexible bend of the 9-residue linker sequence brings the two ice-binding planes of RD3 to an allowed range of orientation. Interestingly, it appeared that TH activity of RD3 is not a simple additive activity of the two type III AFP monomers but exhibits about 6-fold enhancements compared with the monomer (on a molar basis) in the lower concentration range (0–0.5 mM) (17). This significant TH enhancement of RD3 was ascribed to some cooperative effect originating from the two proximal domains; however, no further information has been obtained for considering detailed tandem repetition effects of the AFP domains on the thermal hysteresis and the ice crystal morphology.
In the present study, we attempted to design and generate new intramolecular dimer, trimer, and tetramer of the type III AFP, for which we utilized the primary sequences of either N- or C-domain of RD3. For example, RD3NNC denotes a protein comprising the tandem connections of the two successive N-domain sequences plus the C-domain sequence of RD3. We succeeded in generating RD3NN, RD3NNC, RD3NCN, RD3NCC, and RD3NCNC, for which connections of the AFP domains utilize the 9-residue linker sequence of the native RD3. For these artificial AFP multimers, we measured and compared the TH activity and the ice crystal morphology. Significantly, these AFP multimers exhibit different levels of enhanced TH activity compared with the type III AFP monomer at a lower AFP concentration range on molar and domain bases. It further appeared that these AFP multimers affect the ice crystal morphology in different ways. We believe that these results will significantly expand our understandings of the ice-binding mechanism of AFP and its effect on the TH activity and the ice crystal morphology, which will lead to the design and the production of an activity-modified or -enhanced variant of AFP.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sma/Kpn). By using this as a template, we
further prepared a gene that encodes 6–9 residues of the linker plus RD3
(denoted Linker-RD3) by PCR using a forward primer
(5'-CGCGGATCCTATTCGTAGTTTTTAACCATG-3') and a backward primer
(5'-GGAACTGCAGCCCGGGTCTGAAATCCGTTGTTGCTAACCAG-3'). It is noted
that the gene encoding the part of the linker is designed so as to include
SmaI restriction site. Likewise, a gene encoding RD3 plus linker
(RD3-Linker) was constructed using a forward primer
(5'-GTCCCCCGGGGAGGTGGTGCCGTCTTCGTAGTTTTTAACCATGTCCGGCATCAGGGTCTG-3')
and a backward primer (5'-GAGCTGCAGTTAACTTTAAG-3'). In this case,
the gene encoding 1st to 6th residues of the linker was prepared to include
only SmaI restriction site. In addition, the gene encoding a part of
the linker (including the SmaI restriction site) connected with that
encoding the N-domain of RD3 (denoted as Linker-RD3N) was amplified using a
forward (5'-CGCGGATCCTATTCGTAGTTTTTAACCATG-3') and a backward
primer (5'-GGAACTGCAGCCCGGGTCTGAAATCCGTTGTTGCTAACCAG-3') by using
the gene encoding the N-domain of RD3 as a template. Combined use of these
prepared DNA fragments (RD3Sma/Kpn, Linker-RD3, Linker-RD3N,
RD3-Linker, RD3, and RD3Nl) made it possible to construct the genes encoding
RD3NN, RD3NCC, RD3NNC, RD3NCN, and RD3NCNC as summarized in
Fig. 1, for which the DNA
sequences were occasionally digested with NdeI and BamHI
followed by ligations with a vector pET20b (Novagen).
|
Expressions and Purifications of the Recombinant Multimers—The plasmid DNAs were transformed into Escherichia coli strain BL 21 (DE3), which was grown at 28 °C in LB medium supplemented with 100 µg/ml ampicillin until the cell growth reaches to the early stationary phase. To induce the expression of any recombinant AFP multimer, 0.5 mM isopropyl-D-thiogalactopyranoside was added, and the cultures were grown at 28 °C for overnight. The purification of the multimers was commonly performed according to the methods described previously (19). The culture was centrifuged at 4,000 x g for 30 min at 4 °C, and then we sonicated the precipitated cell pellet containing the inclusion body of the AFP multimers. After the sonication, the fraction containing the AFP multimers was collected by centrifugation at 12,000 x g for 30 min at 4 °C, and then it was washed with 0.1% (v/v) Triton X-100, 1 mM EDTA for three times, and dissolved into 100 mM Tris-HCl (pH 8.5) containing 6 M guanidine hydrochloride at room temperature. Each dissolved AFP multimer was diluted with 50 mM K2HPO4 containing 100 mM NaCl (pH 10.7) at 4 °C followed by extensive dialysis against 50 mM sodium acetate (pH 3.7) at 4 °C. The precipitant formed during dialysis was removed occasionally by centrifugation. The supernatant containing AFP was loaded onto a high-S column (Bio-Rad), and the column-bound AFP was eluted with a linear NaCl gradient (0–0.5 M) using 50 mM sodium acetate buffer (pH 3.7). The fractions containing the isolated AFP were stored and dialyzed against 0.1 M ammonium bicarbonate (pH 7.9). The purity and the molecular weights (14 (dimer), 22 (trimers), and 29 kDa (tetramer)) of the samples were checked by 16% SDS-PAGE.
Measurements of Thermal Hysteresis and Ice Crystal
Morphology— The Tf measurements were
performed for the recombinant AFP multimers dissolved in 0.1 M
ammonium bicarbonate (pH 7.9) by using an osmometer (model OM 802, Vogel). The
observation of ice crystal morphology and the Tm
determination were performed using an in-house photomicroscope system
consisting of Leica DMLB 100 photomicroscope equipped with a Linkam LK600
(liquid nitrogen-type) temperature controller and a CCD camera. The
measurements of Tf and
Tm were repeated three times using the fresh
samples, and the averaged values were used for determination of the TH
activity (TH = Tf –
Tm). The TH values of the multimers were examined
as a function of the protein concentration ranging from 0.025 to 0.6
mM. For the observation of ice crystal morphology, a droplet
(
1.0 µl) of the sample solution was once frozen and subsequently
heated until a single ice crystal was observed separately in the solution by
manipulation of the temperature controller. The morphological change of the
single ice crystal into a hexagonal bipyramid caused by the accumulations of
AFP on the ice surfaces was then observed with the cooling rate of 0.01 °C
per min. The ice crystal morphology observed under different temperatures,
cooling rates, and elapses of time were recorded as a movie file on a personal
computer. The sample volume and the cooling rate were kept constant through
the measurements.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Tf (°C)) of the AFP
multimers in the molar concentration range between 0 and 0.6 mM. As
can be seen, the activity of the monomer (RD3Nl) is dramatically improved in
the dimers (RD3 and RD3NN), and the activities of the dimers are further
improved in the trimers (RD3NNC, RD3NCC, and RD3NCN). No significant
difference in the TH activity was detected between the trimers and a tetramer
(RD3NCNC), although the tetramer exhibits the highest activity. As a
consequence, the activity curves of Fig.
2a were largely divided into three groups corresponding
to monomer, dimers, and the others. Interestingly, it appeared that the TH
activity curves of the multimers are sigmoid in shape, which might be ascribed
to a kinetic icebinding process of the multimers.
|
The differences between the activity curves of
Fig. 2a are distinctly
identified in Fig. 2b,
which plots the TH activity ratio between
Tf (multimers) and
Tf (monomer). For the two recombinant
dimers, it appeared that RD3NN is less active than RD3 in the concentration
range between 0.02 and 0.2 mM. The maximum ratios of RD3 and RD3NN
were estimated to be 8.4 and 6.5, respectively. Among the three trimers,
RD3NNC and RD3NCC possessed almost identical activity through the whole
concentration range, whereas RD3NCN did not catch up to the activity levels of
the other two in the concentration range between 0.02 and 0.15 mM.
The maximum ratios of RD3NNC, RD3NCC, and RD3NCN were estimated to be 22.6,
22.0, and 16.8, respectively. A small enhancement of the relative activity was
distinctly detected between RD3NCNC and the others for a small range of AFP
concentrations between 0.02 and 0.1 mM. The maximum ratio of
RD3NCNC was estimated to be 24.6. Overall, the
Tf
(multimers)/Tf (monomer) ratio increases
in keeping with the number of domain multiplications, although the two species
that comprise the N-domain at the C-terminal end (i.e. RD3NN and
RD3NCN) tends to lower the activity. Another significant observation of
Fig. 2b is that the
top position of each curve shifts to the lower concentration side with
increasing the size of the multimers, suggesting that a larger size of the
multimer exerts the maximum activity at a lower AFP concentration. It is noted
that we have reported the detection of maximum 5.9-fold activity enhancements
of RD3 compared with the monomer
(17). It appeared that our
previous estimation is slightly insufficient presumably due to an imperfect
dissolution of the protein before the TH measurements, and should be revised
to maximum 8.4-fold enhancements in the present study.
Fig. 2c shows the
TH activity of the multimer plotted as a function of the "domain
concentration," in which the Tf
values of the dimer, trimer, and tetramer are plotted against the molar
concentration (mM) times 2, 3, and 4, respectively. This implies
that the longitudinal axis of Fig.
2c represents a "per domain
Tf," so that the observed difference
between the curves of Fig.
2c reflects the tandem repetition-induced change of the
TH activity of an AFP domain comprised in a certain multimer. As can be seen
in Fig. 2c, the per
domain Tf values of every multimer are
notably improved when compared with that of the monomer (RD3Nl). For dimers,
it detected a difference in the profile between RD3 and RD3NN; the activity of
the former is highly improved compared with the latter. For trimers, the
highest and similar profiles of the per domain
Tf curves are obtained for RD3NCC and
RD3NNC, whereas less improvement is detected for RD3NCN. An interesting
finding is that the activity profile of a tetramer RD3NCNC is almost identical
to that of RD3, which implies that the TH activity of the tetramer is not
significantly enhanced at a per domain level. It should be noted that in
Fig. 2c the curves of
the multimers are sigmoid in shape, and the species that comprise the N-domain
at the C-terminal end tend to lower the activity as already described.
It was indicated that both two ice-binding planes constructed in RD3 are able to interact with the prismic planes of a single hexagonal ice crystal, which results in the morphological change of the ice crystal into a unique bipyramidal shape (17). In the present study, the ice crystal morphology was examined for each water solution of the multimer under two different protein concentrations (0.003 and 0.05 mM). As shown in the photomicrographs of Fig. 3, every multimer changes the morphology of a single ice crystal into a diamond shape, which is closely similar to the ordinary hexagonal bipyramid observed for RD3 and various AFPs. Differently from the ordinary bipyramid, the ice crystals created by the multimers commonly lack the acute corners around the pyramidal junction that slightly protrudes to the a axis direction. It is also noted that the pyramidal plane acquires a slight curvature. As for RD3Nl, it does not create a bipyramidal ice crystal at the concentration of 0.003 mM, and it creates an incomplete bipyramid without acute tips when its concentration reaches 0.02 mM (Fig. 3). It is interesting to note that the c/a axial ratio is about 2 for RD3NN at the protein concentration of 0.003 mM, which slightly decreases in proportion with the size of the multimer, and is estimated to be 1.5 for RD3NCNC. Another significant observation is that the ice crystal size becomes dramatically small (1/3 to 1/4) only in the case of multimers by increasing their protein concentration from 0.003 to 0.05 mM. All of these data imply that a molecule comprising the multiple AFP domains connected in tandem acquires a unique manner of the ice binding.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In order to consider the present detected domain multiplication effects on
the TH activity and ice crystal morphology from the three-dimensional
structural viewpoint, we constructed model structures of the trimers and a
tetramer based on our previous determinations of the NMR structure of RD3 (PDB
code 1C8A
[PDB]
) (17).
Fig. 4 illustrates one of the
model structures for RD3NCNC, in which (i) two molecules of RD3 are connected
using the coordinates of the native RD3 linker sequence and (ii) the
N-terminal three residues (-Asn-Lys-Ala-) are deleted from the second NC
molecule, because these three residues do not exist in the second domain of
native RD3 (16,
17). The previous study
demonstrated that the N- and C-domains of RD3 are laterally oriented with an
angle of 32 ± 12°, and the ice binding plane of the N-domain is
located 3.5 Å behind that of the C-domain in an averaged structure
(17). Furthermore, it
demonstrated that the linker portion is flexible and has an ability to bring
the two ice-binding planes of RD3 to an allowed range of orientation; the two
planes are presumably aligned with respect to the ice surface in the ice-bound
form (17). These structural
features are preserved in the model structure of RD3NCNC, leading to a twisted
formation of the ice binding planes (indicated by ellipses in
Fig. 4). For clarity, we rename
this tetramer RD3N1C2N3C4 so as to
indicate the ordinal numbers of the comprised domains. Suppose that this
tetramer undertakes the manner of ice binding previously assumed for RD3
(17), the C4-domain
of RD3N1C2N3C4 located in front
presumably binds first to the ice prism plane in the [0001] direction. The
N3-domain may cause the second ice binding to the same prism plane
in the 
direction, which is
predicted by (i) position matches of the hydrophilic
(15) and hydrophobic
(22) atoms constructing the
putative ice binding planes of the N3-domain with the spacing of
water atoms of the ice prism plane and (ii) a 32° difference in the
lateral alignment of the ice binding planes of the N3- and
C4-domains (Ih: c axis = 7.361
Å; a axis = 4.507 Å; Ref.
23). Similarly, the 3rd ice
binding of the C2-domain of
RD3N1C2N3C4 is assumed against the
same prism plane in the
direction, which is predicted by a total of 64° lateral alignment of
the C2-domain against the C4-domain. The 4th ice binding
of the N1-domain of
RD3N1C2N3C4 might be further
assumed against the same prism plane in the
direction.
|
Obviously, it not possible to predict accurately the target ice surfaces of
RD3NCNC especially for its third and fourth ice bindings, because it is
extremely difficult to know the atom positions of putative ice binding
residues in the AFP domains connected through the flexible linkers and their
position matches with the water atom spacing of the ice
prism plane. Nevertheless, it is
presumed that any domains of RD3NCNC possess the ability for specific ice
binding to strengthen the interaction between the protein and ice, which is
thought to be a key determinant to enhance the TH activity at molar and domain
bases. Such an enhancement of the ice binding ability of the multimer is also
identified by the change of the ice crystal morphology; the
c/a-axial ratio becomes smaller by increasing the size of
the multimer (Fig. 3). DeLuca
et al. (21) reported
that genetically expressed type III AFP connected with thioredoxin (12 kDa) or
maltose-binding protein (42 kDa) possessed only 2–3-fold higher TH
activity than the type III AFP monomer. This small enhancement was ascribed to
just a wider coverage of the ice surface pursued by the connected proteins
that exhibit no antifreeze activity by themselves. The fact that the tetramer
only slightly increases the molar basis TH activity
(Fig. 2b) therefore
suggests that there are significant imperfections of the ice binding between
the fourth AFP domain and the ice crystal surface. One may attribute such
imperfections to the structural and dynamic natures of the linker but not to
an inherent ice binding ability of the fourth AFP domain.
A notable observation is that the multimers containing the N-domain sequence at the C terminus (i.e. RD3NN and RD3NCN) significantly lower their TH activities (Fig. 2, b and c). One of the plausible explanations for this result is that the strength of ice binding is different between the N- and C-domains presumably due to their 20% sequence difference (16), although both of them construct a highly similar type III AFP structure (17) comprising the same numbers of the putative ice binding residues (Gln9, Asn14, Thr15, Thr18, and Gln44 for the N-domain and Gln79, Asn84, Thr85, Thr88, and Gln144 for the C-domain) (15). It should be noted that the averaged order parameter (S2) of the backbone dynamics of the N-domain is slightly smaller than that of the C-domain (17).
To summarize, we succeeded to synthesize intramolecular dimer, trimer, and
tetramer of the type III AFP, and we examined their antifreeze activities.
They exhibit a significant enhancement of thermal hysteresis on molar and
domain bases and also an enhanced ability to modify the morphology and size of
a single ice crystal. The present concept to design an activityenhanced
variant of AFP by tandem repetitions of AFP domains is potentially applicable
to any type of AFP (i.e. types I–IV and 
-helical AFPs).
Such approaches might realize a production of an artificial variant of AFP
that exerts an extraordinary function.
|
FOOTNOTES |
|---|
The on-line version of this article (available at
http://www.jbc.org)
contains a figure.
¶ To whom correspondence should be addressed. Tel.: 81-11-857-8912; Fax: 81-11-857-8983; E-mail: sakae.tsuda{at}aist.go.jp.
1 The abbreviations used are: AFP, antifreeze protein; RD3, native type III
AFP intramolecular dimer comprising the N- and C-domains connected in tandem
through a 9-residue linker sequence; RD3Nl, the N-domain of RD3 with the
linker; RD3NN, the intramolecular dimer comprising the two N-domains connected
in tandem by the linker; RD3NNC (RD3NCN, RD3NCC), the intramolecular trimer
comprising the N- and C-domains connected in tandem by the linker; RD3NCNC,
the intramolecular tetramer comprising the two RD3 molecules connected in
tandem by the linker; Tf, freezing temperature;
Tm, melting temperature; TH, thermal
hysteresis.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  N. B. Holland, Y. Nishimiya, S. Tsuda, and F. D. Sonnichsen Activity of a Two-Domain Antifreeze Protein Is Not Dependent on Linker Sequence Biophys. J., January 15, 2007; 92(2): 541 - 546. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. B. Marshall, A. Chakrabartty, and P. L. Davies Hyperactive Antifreeze Protein from Winter Flounder Is a Very Long Rod-like Dimer of {alpha}-Helices J. Biol. Chem., May 6, 2005; 280(18): 17920 - 17929. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Liu, Z. Jia, G. Chen, C. Tung, and R. Liu Systematic Size Study of an Insect Antifreeze Protein and Its Interaction with Ice Biophys. J., February 1, 2005; 88(2): 953 - 958. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Baardsnes, M. J. Kuiper, and P. L. Davies Antifreeze Protein Dimer: WHEN TWO ICE-BINDING FACES ARE BETTER THAN ONE J. Biol. Chem., October 3, 2003; 278(40): 38942 - 38947. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |
), RD3NN (
), RD3NNC (•), RD3NCN (
),
RD3NCC (
), RD3NCNC (x), and RD3Nl (
). For the per
domain 
