Fully active QAE isoform confers thermal hysteresis activity on a defective SP isoform of type III antifreeze protein Manabu Takamichi 1,2 , Yoshiyuki Nishimiya 1 , Ai Miura 1 and Sakae Tsuda 1,2 1 Functional Protein Research Group, Research Institute of Genome-based Biofactory, National Institute of Advanced Industrial Science and Technology (AIST), Sapporo, Japan 2 Division of Biological Sciences, Graduate School of Science, Hokkaido University, Sapporo, Japan Antifreeze proteins (AFPs) function to inhibit the growth of naturally generated hexagonal ice crystals in supercooled water by specific accumulation onto a set of oxygen atoms constructing specific planes of the crystals [1]. The vacant narrow spaces on an ice plane between the bound AFPs can undergo ice growth to form a convex ice front, which is energetically unfavor- able for the further incorporation of water molecules (the ‘adsorption inhibition’ model [2–4]). The tempera- ture at which ice growth is initiated is commonly referred to as the hysteresis freezing point (T f ). A tem- perature difference between the melting point (T m ) and T f observed for an ice crystal in the presence of AFPs has been defined as thermal hysteresis (TH) [5], which is now generally recognized as a measure of the potency of antifreeze activity. AFPs have been found in various organisms, such as fish, insects, plants, fungi and bacteria, adapted to subzero temperature environments [1]. Of these, fish express type I–IV AFPs and antifreeze glycoprotein (AFGP), which are structurally diverse and form the ice-binding surface in different ways. Significantly, all types of AFP and AFGP are expressed as mixtures of several isoforms [1,4], and a cooperative effect between Keywords cooperative effect; ice growth inhibition; notched-fin eelpout; thermal hysteresis; type III antifreeze protein 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, Sapporo 062-8517, Japan Fax: +81 11 857 8983 Tel: +81 11 857 8983 E-mail: s.tsuda@aist.go.jp (Received 15 November 2008, revised 29 December 2008, accepted 5 January 2009) doi:10.1111/j.1742-4658.2009.06887.x Type III antifreeze protein is naturally expressed as a mixture of sulfopro- pyl-Sephadex (SP) and quaternary aminoethyl-Sephadex (QAE)-binding isoforms, whose sequence identity is approximately 55%. We studied the ice-binding properties of a SP isoform (nfeAFP6) and the differences from those of a QAE isoform (nfeAFP8); both of these isoforms have been identified from the Japanese fish Zoarces elongatus Kner. The two isoforms possessed ice-shaping ability, such as the creation of an ice bipyramid, but nfeAFP6 was unable to halt crystal growth and exhibited no thermal hysteresis activity. For example, the ice growth rate for nfeAFP6 was 1000-fold higher than that for nfeAFP8 when measured for 0.1 mm protein solution at 0.25 °C below the melting point. Nevertheless, nfeAFP6 exhib- ited full thermal hysteresis activity in the presence of only 1% nfeAFP8 (i.e. [nfeAFP8] ⁄ [nfeAFP6] = 0.01), the effectiveness of which was indistin- guishable from that of nfeAFP8 alone. We also observed a burst of ice crystal growth from the tip of the ice bipyramid for both isoforms on low- ering the temperature. These results suggest that the ice growth inhibitory activity of an antifreeze protein isoform lacking the active component is restored by the addition of a minute amount of the active isoform. Abbreviations AFGP, antifreeze glycoprotein; AFP, antifreeze protein; IGM, ice growth modifier; nfeAFP6 and 8, SP and QAE isoforms of type III AFP from notched-fin eelpout; QAE, quaternary aminoethyl; SP, sulfopropyl; T f, freezing point; TH, thermal hysteresis; T m, melting point. FEBS Journal 276 (2009) 1471–1479 ª 2009 The Authors Journal compilation ª 2009 FEBS 1471 isoforms with respect to the TH value has been identi- fied [6–8]. For example, an AFGP based on a repeti- tive polypeptide consisting of Thr–Ala–Ala tripeptide units changes the level of T f depression by the addition of peptides of other lengths [6,7]. Insect AFP from Dendroides canadensis larva, an isoform mixture of repetitive b-helical polypeptides, interacts among the isoforms, which affects the observed TH value [8]. Thus, the characterization of such isoforms and their cooperative effects on antifreeze activity will aid in our understanding of the natural production of various AFP isoforms. Type III AFP is a 7 kDa globular protein that exhibits high sequence identity with the C-terminal domain of human sialic acid synthase [9,10]. Type III AFP was first discovered in the ocean pout Macrozoar- ces americanus as a mixture of 12 isoforms that can be grouped into 11 sulfopropyl (SP)- and one quaternary aminoethyl (QAE)-Sephadex-binding species [11]. The SP and QAE isoforms show approximately 55% sequence identity [12]. Immunological cross-reactivity studies have shown a significant difference between the two isoforms [9], and detailed structural determina- tions by X-ray crystallography and NMR spectroscopy have indicated that the SP and QAE isoforms con- struct a very similar tertiary fold characterized by a two-fold symmetric motif [13–20], which provides a large, flat, amphipathic ice-binding surface [21–24]. Further, ice etching experiments [18,25] and computer simulations [18,23,26] have shown that type III AFP can undergo complementary binding to a set of oxygen atoms located on the {10  1 0} prism plane. We have examined the cooperative effects between several SP and QAE isoforms of type III AFP using a commercial freezing point osmometer [27]. This device determines the T f value by automatic measurement of the ice–water equilibrium temperature of a sample solution. We found that the SP isoform has no appre- ciable TH activity by itself. However, we recently reported that a recombinant SP isoform can modify the shape of an ice crystal into a hexagonal bipyramid below T m , suggesting the possibility that the SP iso- form itself possesses some ability to control the growth of ice crystals, although the freezing point osmometer was unable to detect this ability. These considerations led us to examine the details of ice growth inhibition by the SP isoform and the difference between its activ- ity and that of the QAE isoform. In this study, we observed the morphological change in an ice crystal prepared in solutions of the SP iso- form, the QAE isoform and their mixture employing our custom-made photomicroscope system [28] to evaluate the abilities of these isoforms to inhibit the growth of ice (i.e. TH activity). We used a recombi- nant nfeAFP6 as the SP isoform and nfeAFP8 as the QAE isoform, both of which were identified from Zoarces elongatus Kner [27]. We discuss the ability of ice growth inhibition of the SP and QAE isoforms on the basis of a detailed analysis of the morphological change in the ice crystal. Results Recombinant type III AFP isoforms from notched- fin eelpout, nfeAFP6 and nfeAFP8, were used as the SP and QAE isoforms, respectively. The primary sequences of the two isoforms are described in Materi- als and methods. We first examined the ice growth inhibitory ability of each recombinant, as well as that of a negative control (lysozyme), using a photomicro- scope system developed previously [28]. In the nfeAFP8 solution, the morphology of a hexagonal ice nucleus was modified into a bipyramidal shape, as shown in Fig. 1 (photograph and illustration a). When the crystal growth of this ice bipyramid was measured at a cooling rate of 0.20 °CÆmin )1 , the growth was strongly inhibited at temperatures below T m (Fig. 1, photograph and illustrations a–c). On further cooling, a bursting crystal elongation (i.e. crystal growth) occurred suddenly and rapidly from the tip of the ice bipyramid (Fig. 1, photograph and illustration d). Here, we define this bursting temperature from the bipyramidal tip as T burst . As no significant ice growth occurred for nfeAFP8 until T burst , this temperature was identified as the ice growth initiation temperature, or simply the hysteresis freezing point (T f ) of this solu- tion [28]. In the case of nfeAFP8, TH = |T m ) T f | could hence be evaluated with T m and T burst . We also observed a bipyramidal ice crystal for nfeAFP6 at a cooling rate of 0.20 °CÆmin )1 (Fig. 1, photograph and illustration e). This ice bipyramid, however, expanded rapidly and continuously with retention of the a-toc-axis ratio (1 : 3) (Fig. 1, photo- graph and illustrations e–g), followed by a bursting elongation of the ice crystal suddenly from the tip of the ice bipyramid (Fig. 1, photograph and illustra- tion h). Such a bursting elongation was similarly observed for nfeAFP8 (Fig. 1, photograph d). The only difference was that the bursting elongation of nfeAFP6 accompanied rapid crystal expansion, whereas that of nfeAFP8 did not. We can hence evalu- ate, for nfeAFP6, the burst initiation temperature of an ice crystal as T burst . Rapid crystal expansion was observed even at a temperature only slightly below T m (i.e. T m ) 0.05 °C), which was closely similar to the observation for the negative control (Fig. 2C); the only Function of SP isoform of type III AFP M. Takamichi et al. 1472 FEBS Journal 276 (2009) 1471–1479 ª 2009 The Authors Journal compilation ª 2009 FEBS difference was in the ice crystal morphology. In other words, the ice crystal created in nfeAFP6 and negative control solution melted above T m and grew below T m . That is, each solution maintained an ice–water equilib- rium state around T m , for which no infinite value of the non-equilibrium freezing point could be defined. We therefore conclude that T m and T f are equal for nfeAFP6 solution, and that no TH activity could be evaluated for nfeAFP6. Table 1 shows the growth rate of the ice bipyramid estimated for each isoform; the a-axis length of the ice bipyramid was used for evaluation. The growth rates were measured at two annealing temperatures between T m and T burst (T m ) 0.05 °C and T m ) 0.25 °C). As shown in Table 1, rapid and constant rates of ice growth were evaluated for nfeAFP6, although the rates were slower than that for the negative control (lyso- zyme). It should be noted that, for nfeAFP6, the ice growth (i.e. expansion) rate was decelerated by incre- asing the protein concentration and accelerated by lowering the annealing temperature. The crystal expansion was observed even at a high concentration of nfeAFP6 (5.0 mm). We also found very slow growth (4 · 10 )2 lmÆmin )1 ) of the ice bipyramid for diluted nfeAFP8 solution, which agrees with a previous report by Deluca et al. [29]. When the nfeAFP8 concentration exceeded 0.1 mm, crystal growth was halted, making it difficult to measure the rate at T m ) 0.05 °C. We were able to compare the ice growth rates for the two isoforms for their 0.1 mm solutions at T m ) 0.25 °C; the observed value for nfeAFP6 was approximately 1000-fold faster than that for nfeAFP8. In addition, the growth rate of nfeAFP8 decreased slightly with time, with only slight crystal growth along the c-axis (data not shown). Figure 3A shows the ice bipyramid observed in a 1 : 1 mixture of nfeAFP6 (0.05 mm) and nfeAFP8 (0.05 mm) isoforms with a total concentration of 0.1 mm at T m ) 0.25 °C. Crystal growth was clearly inhibited between T m and T burst for 30 min, similar to that in the case of nfeAFP8 only. Figure 3B shows the growth rate of the ice bipyramid for some mix- tures of nfeAFP6 and nfeAFP8 at T m ) 0.05 °C, the total concentration being adjusted to 0.1 mm. The ice growth rate decreased dramatically from 6 to 0.1 lmÆmin )1 by the addition of only 1% nfeAFP8 a A B b c d e a b c d e f g h f g h Fig. 1. Morphological change in an ice crystal observed for 0.1 m M solutions of nfeAFP8 (QAE isoform) and nfeAFP6 (SP isoform). (A) Photomicroscope images for nfeAFP8 (a–d) and nfeAFP6 (e–h) were obtained at a cooling rate of 0.20 °CÆmin )1 . (a, e) T = T m ) 0.1 °C; (b, f) T = T m ) 0.2 °C; (c, g) T = T burst ; (d, h) T = T burst after 0.05 s. (B) Illustrations of crystal growth observed for solutions of nfeAFP8 (a–d) and nfeAFP6 (e–h) at different temperatures. M. Takamichi et al. Function of SP isoform of type III AFP FEBS Journal 276 (2009) 1471–1479 ª 2009 The Authors Journal compilation ª 2009 FEBS 1473 (i.e. [nfeAFP8] ⁄ [nfeAFP6] = 0.01), and reached a pla- teau (0.06 lmÆmin )1 ) at 12.5% nfeAFP8. At 25% nfeAFP8, the growth rate was indistinguishable from that at 100% nfeAFP8. Figure 4A shows time-course snapshots of a busting growth of an ice bipyramid observed in a 0.1 mm solu- tion of the 1 : 1 mixture of the two isoforms, detected at the temperature T burst . Bursting growth occurred from the tip of the ice bipyramid, as similarly observed for the two isoforms (Fig. 1Ad and h). We were able to estimate the TH value for this mixture, as the ice crystal did not expand between T m and T burst (Fig. 3A), but underwent bursting crystal growth at T burst (i.e. T burst = T f ). TH activities of nfeAFP6, nfeAFP8 and their 1 : 1 mixture plotted against their A B C Fig. 2. Photomicroscope images of an ice crystal observed for 0.1 m M solutions of nfeAFP8 and nfeAFP6 between T m and T burst . (A) An ice bipyramid observed for nfeAFP8 at T m ) 0.25 °C; the crystal growth was strongly inhibited for longer than 30 min. (B) Expansion of an ice bipyramid observed for nfeAFP6 at T m ) 0.25 °C. The snapshots were taken before and after 2 min of annealing time. (C) Snapshots of 0.1 m M solution of negative con- trol (lysozyme) annealed slightly below its T m value (T m ) 0.05 °C). The horizontal bars and arrows indicated in the photographs repre- sent a scale of 20 lm and the direction of the c-axis, respectively. The c-axis is vertical to the paper for the negative control. Table 1. Ice growth rates determined for solutions of nfeAFP6, nfeAFP8 and the negative control (lysozyme) between T m and T burst . Sample Concentration (m M) Annealing temperature (°C) Ice growth rate (lmÆmin )1 ) nfeAFP6 (SP) 0.1 T m ) 0.05 5.9 T m ) 0.25 13.7 1.0 T m ) 0.05 1.8 T m ) 0.25 3.8 5.0 T m ) 0.05 0.6 T m ) 0.25 1.8 nfeAFP8 (QAE) 0.01 T m ) 0.05 4 · 10 )2 0.1 T m ) 0.05 – T m ) 0.25 1 · 10 )2 Lysozyme (negative control) 0.1 T m ) 0.05 150 A B Fig. 3. (A) Photomicroscope images of an ice crystal observed for a 0.1 m M solution of a 1 : 1 mixture of nfeAFP6 and nfeAFP8 annealed at T m ) 0.25 °C. The crystal showed no significant change for 30 min. The horizontal scale bar represents 20 lm; the arrow indicates the direction of the c-axis, which is perpendicular to the a-axis. (B) Crystal growth rates of an ice bipyramid determined at T m ) 0.05 °C for nfeAFP6 in the presence of various amounts of nfeAFP8, with the total protein concentration adjusted to 0.1 m M. The x-axis indicates the percentage of nfeAFP8. For example, 25% indicates 0.075 m M nfeAFP6 plus 0.025 mM nfeAFP8. To evaluate the growth rate, we measured the length of the middle of the bipyramid along the a-axis. Function of SP isoform of type III AFP M. Takamichi et al. 1474 FEBS Journal 276 (2009) 1471–1479 ª 2009 The Authors Journal compilation ª 2009 FEBS concentrations are shown in Fig. 4B. Interestingly, the TH activity of the 1 : 1 mixture was closely equivalent to that of nfeAFP8 when plotted against the total AFP concentration. Figure 4C plots the TH activity of a 0.1 mm solution of the mixture against the percent- age content of nfeAFP8. Full activity was observed in the presence of only 1% of the QAE isoform, as sug- gested by Fig. 3B. These results indicate that nfeAFP6 exhibited full TH activity in the presence of only a minute amount of nfeAFP8. Figure 5 shows the value of |T m ) T burst | plotted against the protein concentration (mm) for each nfeAFP6 and nfeAFP8 isoform. Interestingly, the two profiles show closely similar hyperbolic curves, which largely overlap the whole concentration range. This suggests that the SP and QAE isoforms possess similar levels of growth inhibitory ability for the tip of the ice bipyramid. Discussion This study reveals that the SP isoform (nfeAFP6) lacks the ability to inhibit ice growth (Figs 2 and 3), similar A a b c d B Fig. 4. (A) Burst elongation (crystal growth) arising from the tip of an ice bipyramid in a 0.1 mM solution of a 1 : 1 mixture of nfeAFP6 and nfeAFP8. (a) An ice bipyramid just before initiation of burst growth. (b) Initiation of the crystal burst from the tip of the ice bipyramid. (c, d) The crystal burst rapidly proceeds and leads to complete freezing of the solution. Insets show expanded views of the growth initiation point. Horizontal bar represents 20 lm. Arrow shows the direction of the c-axis. (B) TH activities of nfeAFP6 (filled triangles), nfeAFP8 (filled circles) and a 1 : 1 mixture of the two isoforms (open circles) as a function of each total concentration. TH was evaluated as the difference between T f and T m (i.e. TH = |T m ) T f |). For nfeAFP6, T f = T m . For nfeAFP8 and the 1 : 1 mixture, T burst = T f . (C) TH activity of nfeAFP6 in the presence of various proportions of nfeAFP8, with the total protein concentration adjusted to 0.1 m M. 0 0 0.1 0.2 0.3 0.4 0.1 0.2 0.3 0.4 0.5 0.6 Concentration (mM) | T m - T burst | (ºC) nfeAFP6 (SP) nfeAFP8 (QAE) Fig. 5. Concentration dependence of |T m ) T burst | for nfeAFP6 (open circles) and nfeAFP8 (filled circles). T burst represents the tem- perature at which a burst of ice crystal growth occurs from the tip of the ice bipyramid. We were unable to define TH activity, but obtained a |T m ) T burst | value for nfeAFP6. For nfeAFP8, this value was identical to the TH activity (Fig. 4B). M. Takamichi et al. Function of SP isoform of type III AFP FEBS Journal 276 (2009) 1471–1479 ª 2009 The Authors Journal compilation ª 2009 FEBS 1475 to the negative control, although nfeAFP6 can specifi- cally interact with an ice nucleus to form an ice bipyra- mid (i.e. it has ice-shaping ability). Consequently, nfeAFP6 exhibits no TH activity and shows an ordin- ary ice–water equilibrium phenomenon (i.e. T m = T f ) (Fig. 2). It is interesting that, although this SP isoform was found to be a dominant component of a purified fish type III AFP [27], it cannot function as an AFP, but should rather be termed an ‘ice growth modifier’ (IGM), according to the nomenclature proposed by Harding et al. [30]. Similar observations have been reported in studies of artificial AFP mutants [31–34]. For example, a flounder type I AFP lost 90% of its inherent TH activity and allowed continuous growth of ice bipyramids when Ala21 was replaced with Leu [31]. Recombinant type I AFP from shorthorn sculpin that lacks N-terminal blocking (denoted rSS3) as well as that of the lysine mutant of Ala25 also failed to inhibit the growth of ice bipyramids [32,33]. For type II AFP from longsnout poacher, continuous growth of an ice bipyramid was observed by mutation of Ile58, a residue located within a planar ice-binding surface [34]. HPLC12, a QAE isoform of type III AFP with 94% sequence identity to nfeAFP8, permits con- tinuous growth of an ice bipyramid that expands rap- idly on amino acid replacement of Ala16, a residue located at the center of the ice-binding surface [29]. It has been observed that the bipyramidal ice crystal grows continuously with retention of the c-:a-axis ratio at approximately 3.3 for flounder type I AFP mutants that accumulate onto a {20  21} pyramidal plane. Significantly, a similar continuous growth of the ice bipyramid with retention of the c-:a-axis ratio of approximately 3 was observed for nfeAFP6 (Fig. 2), suggesting that nfeAFP6 binds to the pyramidal plane. In contrast, significant ice growth along the c-axis was identified for ordinary type III AFP and sculpin type I AFP, which bind to the prism plane. nfeAFP8 also permitted the growth of an ice bipyramid along the c-axis at a low concentration (0.005 mm), suggesting that nfeAFP8 binds to the prism plane. It should be noted that a recent study on ice etching revealed that type III AFP can bind to several ice planes, including the {10  10} prism and the {20  21} pyramidal plane [18]. The preliminary X-ray structure of nfeAFP6 (data not shown) obtained was indistinguishable from that of ordinary type III AFP (i.e. HPLC12) [14,18]. There is a 41% sequence difference between nfeAFP6 and AFP8, which includes Leu19 and Val20 locating at the edge of the putative ice-binding region. Hence, we assume that amino acid replacements of Leu19 and Val20 differentiate the manner of ice binding between nfeAFP6 and nfeAFP8. The fast and slow ice growth rates evaluated for nfeAFP6 and nfeAFP8 reveal a difference in their growth inhibitory function. An extremely slow growth rate was found for 0.01 mm nfeAFP8 (4 · 10 )2 lmÆmin )1 , Table 1) and for the mixture of approximately 0.01 mm nfeAFP8 plus 0.09 mm nfeAFP6 (10% nfeAFP8) (Fig. 3B). As nfeAFP6 could not halt ice growth by itself, it may be assumed that the two iso- forms act cooperatively for ice growth inhibition. Indeed, TH activity of the 1 : 99 mixture containing approximately 0.001 mm nfeAFP8 (TH = 0.33 °C) (Fig. 4C) was higher than that of 0.05 mm nfeAFP8 alone (TH = 0.26 °C) (Fig. 4B). We examined the 1 H- 15 N heteronuclear single quantum coherence spec- trum of 15 N-labeled nfeAFP6 in the absence and pres- ence of 20% of non-labeled nfeAFP8, and found that the two spectra were virtually identical (Fig. S1). Hence, we can assume no significant protein–protein interaction between the two isoforms, which is in good agreement with the proposed independent ice-binding model of AFP [35,36]. Kristiansen and Zachariassen [4] proposed a two-step irreversible adsorption of AFPs to the ice surface; i.e. following the initial adsorption controlled by hydrophobic forces, perma- nent adsorption occurs on this plane by hydrophilic forces. This proposition may account for the weak ice growth inhibition of nfeAFP6. That is, nfeAFP6 can undergo initial adsorption to a pyramidal ice plane, but fails to undergo the secondary permanent adsorp- tion on this plane. As nfeAFP8 presumably possesses an ability to execute the two-step irreversible adsorp- tion to the prism plane, nfeAFP6 may irreversibly bind to the prism plane in the presence of nfeAFP8. In the case of nfeAFP8, we observed that crystal bursting was initiated from the tapered tip of the ice bipyramid at T f (= T burst ) (Fig. 1Ad), implying that the tip is the weakest point. Significantly, even nfeAFP6 can inhibit crystal bursting from the tip between T m and T burst (Fig. 1Ae,f), similar to nfeAFP8 (Fig. 1Aa–c). The TH activities obtained for nfeAFP8 and the 1 : 1 mixture of nfeAFP6 and nfeAFP8 were indistinguishable (Fig. 4B). The level of depression of T burst for nfeAFP6 was also indistinguishable from that for nfeAFP8 (Fig. 5). These results suggest that the functions of nfeAFP6 and nfeAFP8 are equivalent with regard to the inhibition of growth from the tips of the ice bipyramid. To our knowledge, there is little documentation about such growth inhibition of the bipyramidal tip by type III AFP. We offer a plausible explanation below. The origin of an ice bipyramid is a hexagonal ice unit (i.e. ice nucleus) generated in supercooled water under a pressure of 1013 hPa [37]. A scheme of transi- Function of SP isoform of type III AFP M. Takamichi et al. 1476 FEBS Journal 276 (2009) 1471–1479 ª 2009 The Authors Journal compilation ª 2009 FEBS tion from an ice nucleus to an ice bipyramid in the presence of AFP (Fig. 6), which was proposed many years ago [38], is still widely accepted with no signifi- cant revisions. Inherently, the ice growth rate on the six prism planes is much faster than that on the basal plane [37,39]. When AFPs are present, they accumulate on the six prism planes and inhibit their growth along the a 1 - , a 2 - and a 3 -axes (first layer in Fig. 6), but can- not inhibit the generation of a new ice nucleus on the first layer, namely ice growth along the c-axis [38]. AFPs further accumulate on the prism planes grown from the new ice nucleus and a hinge region between the second prism and the first basal plane [14], thereby creating a hexagonal ice plate that is smaller than the first layer. Repeated binding of AFP to the prism plane and the generation of a smaller ice nucleus cause successive stacking of hexagonal ice plates on the basal plane, leading to the formation of an ice bipyramid, as illustrated in Fig. 6. When pyramidal planes are cre- ated by the adsorption of AFPs, the 12 equivalent planes construct the bipyramidal ice crystal [40]. Hence, one can imagine that the tip of the ice bipyr- amid is the basal plane or a part of the basal plane, which is of ultimately small size. Therefore, such an extremely narrow space of the top plane scarcely allows elongation of the ice crystal along the c-axis between T m and T burst , thereby maintaining bipyrami- dal morphology (Fig. 1). Explosive ice growth along the c-axis may be induced by slight expansion of the narrow top plane through ice growth towards the other axes on lowering the temperature to T burst (Fig. 1). That is, an increase in isoform concentration (Fig. 5) may contribute to the stability of the top plane of the ice bipyramid by stabilizing its prism or pyra- midal planes, and this increases the value of |T m ) T burst |. These suppositions do not contradict the recent observations of Scotter et al. [41]. These authors showed that the crystal burst always occurred from the tip (basal plane) of the ice bipyramid for most fish AFPs, and occurred from the prism plane for insect hyperactive AFPs. They ascribed the former observa- tion to no binding ability of fish AFPs to the basal plane and the latter to the binding of hyperactive AFPs not only to the prism but also to the basal plane. We assume that the fully active isoform of type III AFP strongly binds to the prism plane, whereas a defective isoform weakly interacts with the pyramidal plane, but they possess similar abilities to inhibit ice growth from the tip of the ice bipyramid. In summary, we have found that a minute amount of the active QAE isoform of type III AFP confers TH activity on the SP isoform, which possesses no TH activity by itself. This may imply that the large amount of the SP isoform contained in the body fluid makes a significant contribution to ice growth inhibition with the help of the active isoform, thereby enabling the host fish to survive in seawater at subzero tempera- tures. The present findings may further suggest that any defective antifreeze analog (e.g. IGM) could be used as an effective TH substance by the addition of a minute amount of the fully active antifreeze substance. Materials and methods Sample preparation Recombinant proteins of the type III AFP isoforms nfeAFP6 and nfeAFP8 were prepared as described previously [27] with some modifications. After sonication of genetically trans- formed Escherichia coli BL21 (DE3), a soluble fraction con- taining a recombinant isoform was dialyzed against 50 mm citric acid buffer (pH 2.9). Cation exchange chromatography was then performed using an Econo-Pac High S cartridge (Bio-Rad, Hercules, CA, USA) with a linear NaCl gradient (0–0.5 m) with 50 mm citric acid buffer (pH 2.9). The amino acid sequences of nfeAFP6 and nfeAFP8 are as follows: nfeAFP6, G1ESVVATQLIPINTALTPAMMEGKVTNPS GIPFAEMSQIVGKQVNTPVAKGQTLMPGMVKTYVP AK66; nfeAFP8, N1QASVVANQLIPINTALTLVMMRA EVVTPMGIPAVDIPRLVSMQ VNRAVPLGTTLMPEM VKGYTPA65. The concentration of each purified sample was measured by UV absorption (280 nm) using a DU 530 spectrophotometer (Beckman Coulter, Fullerton, CA, USA). c-axis Basal plane Prism plane a 1 a 2 a 3 1 st layer 2 nd layer 3 rd layer Fig. 6. Left: illustration of the proposed transition scheme from ice nucleus to ice bipyramid based on previous ideas [38]. Right: illus- tration of an ice bipyramid observed in a solution of type III AFP. Detailed explanations are given in the text. M. Takamichi et al. Function of SP isoform of type III AFP FEBS Journal 276 (2009) 1471–1479 ª 2009 The Authors Journal compilation ª 2009 FEBS 1477 Measurement of ice growth rate and TH activity Ice crystal morphology was observed and the crystal growth rate was measured for solutions of nfeAFP6, nfeAFP8 and their mixtures using a custom-made photomicroscope system described in [28]. Detailed procedures for the evalua- tion of TH activity are also described in [28] using a cooling rate of 0.20°CÆmin )1 . The sample solution was placed in a capillary tube and frozen at approximately )30 °C; it was then warmed by manipulation of the temperature control device until a single ice crystal was observed ( 10–30 lm). The observed ice crystal was annealed slightly below its T m , and recorded using a Color-video 3CCD camera (Sony, Tokyo, Japan). The annealing period was 0.5–1 h for nfeAFP6 and 3 h for nfeAFP8. The ice growth rate was examined in five to ten still images captured at regular inter- vals; the length of the middle of the bipyramid along the a-axis was used for evaluations. In all solutions tested here, the value of T m was )0.3 °C. This was equivalent to that of the buffer solution (0.1 m ammonium bicarbonate, pH 7.9). Acknowledgements The authors thank Dr Hidemasa Kondo for providing them with a preliminary X-ray structure of nfeAFP6. References 1 Jia Z & Davies PL (2002) Antifreeze proteins: an unusual receptor–ligand interaction. Trends Biochem Sci 27, 101–106. 2 Raymond JA & DeVries AL (1977) Adsorption inhibition as a mechanism of freezing resistance in polar fishes. Proc Natl Acad Sci USA 74, 2589– 2593. 3 Knight CA & DeVries AL (1994) Effects of a poly- meric, nonequilibrium ‘antifreeze’ upon ice growth from water. J Cryst Growth 143, 301–310. 4 Kristiansen E & Zachariassen KE (2005) The mecha- nism by which fish antifreeze proteins cause thermal hysteresis. Cryobiology 51, 262–280. 5 DeVries AL (1986) Antifreeze glycopeptides and pep- tides: interactions with ice and water. Meth Enzymol 127, 293–303. 6 Osuga DT, Ward FC, Yeh Y & Feeney RE (1978) Cooperative functioning between antifreeze glyco- proteins. J Biol Chem 253, 6669–6672. 7 Schrag JD & DeVries AL (1983) The effect of freezing rate on the cooperativity of antifreeze glycopeptides. Comp Biochem Physiol 74, 381–385. 8 Wang L & Duman JG (2005) Antifreeze proteins of the beetle Dendroides canadensis enhance one another’s activities. Biochemistry 44, 10305–10312. 9 Hew CL, Slaughter D, Joshi SB, Fletcher GL & Anan- thanarayanan VS (1984) Antifreeze polypeptides from the Newfoundland ocean pout, Macrozoarces americ- anus: presence of multiple and compositionally diverse components. J Comp Physiol B Biochem Syst Environ Physiol 155, 81–88. 10 Baardsnes J & Davies PL (2001) Sialic acid synthase: the origin of fish type III antifreeze protein? Trends Biochem Sci 26, 468–469. 11 Hew CL, Wang NC, Joshi S, Fletcher GL, Scott GK, Hayes PH, Buettner B & Davies PL (1988) Multiple genes provide the basis for antifreeze protein diversity and dosage in the ocean pout, Macrozoarces americ- anus. J Biol Chem 263, 12049–12055. 12 Li XM, Trinh KY, Hew CL, Buettner B, Baenziger J & Davies PL (1985) Structure of an antifreeze polypeptide and its precursor from the ocean pout, Macrozoarces americanus. J Biol Chem 260, 12904–12909. 13 So ¨ nnichsen FD, Sykes BD, Chao H & Davies PL (1993) The nonhelical structure of antifreeze protein type III. Science 259, 1154–1157. 14 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. 15 So ¨ nnichsen FD, DeLuca CI, Davies PL & Sykes BD (1996) Refined solution structure of type III antifreeze protein: hydrophobic groups may be involved in the energetics of the protein–ice interaction. Structure 4, 1325–1337. 16 Yang DS, Hon WC, Bubanko S, Xue Y, Seetharaman J, Hew CL & Sicheri F (1998) Identification of the ice-binding surface on a type III antifreeze protein with a ‘flatness function’ algorithm. Biophys J 74, 2142– 2151. 17 Graether SP, DeLuca CI, Baardsnes J, Hill GA, Davies PL & Jia Z (1999) Quantitative and qualitative analysis of type III antifreeze protein structure and function. J Biol Chem 274, 11842–11847. 18 Antson AA, Smith DJ, Roper DI, Lewis S, Caves LS, Verma CS, Buckley SL, Lillford PJ & Hubbard RE (2001) Understanding the mechanism of ice binding by type III antifreeze proteins. J Mol Biol 305, 875–889. 19 Miura K, Ohgiya S, Hoshino T, Nemoto N, Suetake T, Miura A, Spyracopoulos L, Kondo H & Tsuda S (2001) NMR analysis of type III antifreeze protein intramolecular dimer. Structural basis for enhanced activity. J Biol Chem 276, 1304–1310. 20 Ko TP, Robinson H, Gao YG, Cheng CH, DeVries AL & Wang AH (2003) The refined crystal structure of an eel pout type III antifreeze protein RD1 at 0.62-A ˚ reso- lution reveals structural microheterogeneity of protein and solvation. Biophys J 84, 1228–1237. 21 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. Function of SP isoform of type III AFP M. Takamichi et al. 1478 FEBS Journal 276 (2009) 1471–1479 ª 2009 The Authors Journal compilation ª 2009 FEBS 22 DeLuca CI, Davies PL, Ye Q & Jia Z (1998) The effects of steric mutations on the structure of type III anti- freeze protein and its interaction with ice. J Mol Biol 275, 515–525. 23 Chen G & Jia Z (1999) Ice-binding surface of fish type - III antifreeze. Biophys J 77, 1602–1608. 24 Baardsnes J & Davies PL (2002) Contribution of hydro- phobic residues to ice binding by fish type III antifreeze protein. Biochim Biophys Acta 1601, 49–54. 25 Cheng CC & DeVries AL (1991) The role of antifreeze glycopeptides and peptides in the freezing avoidance of cold-water fish. In Life Under Extreme Conditions (diPrisco G, ed.), pp. 2–14. Springer Verlag, Berlin, Heidelberg. 26 Madura JD, Taylor MS, Wierzbicki A, Harrington JP, Sikes CS & So ¨ nnichsen FD (1996) The dynamics and binding of a Type III antifreeze protein in water and on ice. J Mol Struct 388, 65–77. 27 Nishimiya Y, Sato R, Takamichi M, Miura A & Tsuda S (2005) Co-operative effect of the isoforms of type III antifreeze protein expressed in Notched-fin eelpout, Zoarces elongates Kner. FEBS J 272, 482–492. 28 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 274, 6469–6476. 29 DeLuca CI, Chao H, So ¨ nnichsen FD, Sykes BD & Davies PL (1996) Effect of type III antifreeze protein dilution and mutation on the growth inhibition of ice. Biophys J 71, 2346–2355. 30 Harding MM, Ward LG & Haymet ADJ (1999) Type I ‘antifreeze’ proteins: structure–activity studies and mechanisms of ice growth inhibition. Eur J Biochem 264, 653–665. 31 Baardsnes J, Kondejewski LH, Hodges RS, Chao H, Kay C & Davies PL (1999) New ice-binding face for type I antifreeze protein. FEBS Lett 463, 87–91. 32 Fairley K, Westman BJ, Pham LH, Haymet ADJ, Har- ding MM & Mackay JP (2002) Type I shorthorn scul- pin antifreeze protein. J Biol Chem 277, 24073–24080. 33 Baardsnes J, Jelokhani-Niaraki M, Kondejewski LH, Kuiper MJ, Kay CM, Hodges RS & Davies PL (2001) Antifreeze protein from shorthorn sculpin: identification of the ice-binding surface. Protein Sci 10, 2566–2576. 34 Nishimiya Y, Kondo H, Takamichi M, Sugimoto H, Suzuki M, Miura A & Tsuda S (2008) Crystal structure and mutational analysis of Ca 2+ -independent type II antifreeze protein from longsnout poacher, Brachyopsis rostratus. J Mol Biol 382, 734–746. 35 Chao H, DeLuca CI & Davies PL (1993) Mixing anti- freeze protein types changes ice crystal morphology without affecting antifreeze activity. FEBS Lett 357, 183–186. 36 Deluca CI, Comley R & Davies PL (1998) Antifreeze proteins bind independently to ice. Biophys J 74, 1502– 1508. 37 Hobbs PV (1974) Ice Physics. Clarendon Press, Oxford. 38 Davies PL & Hew CL (1990) Biochemistry of fish anti- freeze proteins. FASEB J 4, 2460–2468. 39 Fletcher NH (1970) Crystal growth. In The Chemical Physics of Ice, Chapter 5, pp. 111–112. Cambridge University Press, London. 40 Houston ME Jr, Chao H, Hodges RS, Sykes BD, Kay CM, So ¨ nnichsen FD, Loewen MC & Davies PL (1998) Binding of an oligopeptide to a specific plane of ice. J Biol Chem 273, 11714–11718. 41 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. Supporting information The following supplementary material is available: Fig. S1. 500 MHz 1 H- 15 N heteronuclear single quan- tum coherence spectra of a recombinant 15 N-labeled protein of the defective isoform nfeAFP6 dissolved in water in the absence and presence of non-labeled nfeAFP8 (temperature, 4 °C; pH 6.7; [nfeAFP6] : [nfeAFP8] = 1 : 4; total concentration, 1 mm). This supplementary material can be found in the online version of this article. Please note: Wiley-Blackwell is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. M. Takamichi et al. Function of SP isoform of type III AFP FEBS Journal 276 (2009) 1471–1479 ª 2009 The Authors Journal compilation ª 2009 FEBS 1479 . Fully active QAE isoform confers thermal hysteresis activity on a defective SP isoform of type III antifreeze protein Manabu Takamichi 1,2 ,. the ice bipyramid. In summary, we have found that a minute amount of the active QAE isoform of type III AFP confers TH activity on the SP isoform, which