- Email: [email protected]
Isolation and characterization of type I antifreeze proteins from Atlantic snailfish (Liparis atlanticus) and dusky snailfish (Liparis gibbus)
Biochimica et Biophysica Acta 1547 (2001) 235^244 www.bba-direct.com
Isolation and characterization of type I antifreeze proteins from Atlantic snail¢sh (Liparis atlanticus) and dusky snail¢sh (Liparis gibbus) Robert P. Evans *, Garth L. Fletcher Ocean Sciences Centre, Memorial University of Newfoundland, St. John's, NF A1C 5S7, Canada Received 10 August 2000; received in revised form 1 March 2001; accepted 15 March 2001
Abstract Antifreeze proteins (AFPs) were isolated from the blood plasma of Atlantic snailfish Liparis atlanticus and dusky snailfish Liparis gibbus, which belong to the Teleost family Cyclopteridae, a close relative of sculpins. Using a combination of gel filtration chromatography and reversed-phase HPLC, proteins were purified to individual peaks. Atlantic snailfish plasma contained two different proteins (MW = 9344, 9415) while dusky snailfish plasma contained five protein isoforms (MW = 9514^9814), as determined by mass spectrometry. Further characterization revealed that these proteins are rich in alanine ( s 50 mol%), and have K-helical secondary structure that can undergo reversible thermal denaturation. Thermal hysteresis activities of these proteins were similar to each other but lower than the major type I AFPs from winter flounder. Results of this study have indicated that although the AFPs from snailfish are significantly larger than previously described type I AFPs, they share enough characteristics to be classified in this group. ß 2001 Elsevier Science B.V. All rights reserved. Keywords: Type I; Antifreeze; Protein; Plasma ; Snail¢sh; Liparis
1. Introduction Teleost ¢sh that inhabit icy seawater in winter, at temperatures as low as 31.8³C, synthesize antifreeze proteins/polypeptides (AFPs) or antifreeze glycoproteins (AFGPs) for protection against freezing. Species from diverse taxonomic groups produce AFPs that are classi¢ed as four distinct classes based on their physical characteristics, termed type I, II, III and IV. However, regardless of protein or polypeptide structure, all classes of ¢sh AFPs lower the
* Corresponding author. Fax: +1-709-737-3220; E-mail: [email protected]
freezing point of plasma non-colligatively by binding to certain surfaces of ice crystals, modifying their structure and inhibiting further growth. Extensive reviews have been published recently that deal with general structure/function characteristics, diversity, physiology and molecular genetics of AFPs and AFGPs [1^4]. Of the four classes of AFPs discovered to date, the smallest and simplest class has been the most extensively studied. Type I AFPs are found in certain sculpins (e.g. Myoxocephalus) and unrelated right eye £ounders (Pleuronectes). Structurally, they are usually polypeptides that have high alanine content ( s 60 mol%), have an amphipathic K-helical secondary structure, and are usually small (approx. 3.3^4.5 kDa). The winter £ounder Pleuronectes americanus
0167-4838 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 8 3 8 ( 0 1 ) 0 0 1 9 0 - X
BBAPRO 36423 31-5-01
236
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
has provided an excellent working model for developing an understanding of AFP function and regulation. Winter £ounder synthesize AFP in the liver, which is then secreted into the blood to protect the ¢sh from extracellular freezing. Additionally, this species produces a type I protein in its skin that lacks a peptide signal sequence that would indicate it remains intracellular [5]. The di¡erent sites of action suggest an alternative biological function for serum and skin AFPs. The snail¢shes belong to a large family of benthic and pelagic marine ¢shes, Cyclopteridae, that inhabit northern regions of the Atlantic, Paci¢c and Arctic oceans. This family is closely related to the sculpins, which belong to a di¡erent family of the same order Scorpaeniformes. The Atlantic snail¢sh, Liparis atlanticus, is a very small species, with mature adults ranging in length from 10 to 95 mm. They inhabit inshore waters of the northwest Atlantic Ocean from Ungava Bay in northern Quebec and south to New York [6]. During the winter months, January to March, ¢sh move further inshore where females spawn, in depths less than 2 m. Dusky snail¢sh Liparis gibbus is a larger species, averaging 114 mm in length. Their habitat extends from the Canadian Arctic Ocean, across to Greenland and down to the coast of Newfoundland where they live mainly on muddy bottoms at depths up to 200 m [6]. Because their harsh natural environments would frequently expose both of these snail¢sh species to ice laden seawater, they were considered excellent candidates for possession of AFPs. It was originally hypothesized that the genus Liparis might produce either type I or type II AFPs, since closely related species from the order Scorpaeniformes produce both classes of AFPs. Shorthorn sculpin, Myoxocephalus scorpius, and grubby sculpin, Myoxocephalus aenaeus, have type I AFP while sea raven, Hemitripterus americanus, has type II AFP in their blood plasma. Furthermore, it has recently been shown that another Scorpaeniforme, the longhorn sculpin Myoxocephalus octodecimspinosis, produces an entirely unrelated antifreeze protein in its blood plasma ^ type IV [7^9]. For these reasons, we have puri¢ed and characterized AFPs from snail¢sh to help elucidate their evolutionary origins and to clarify some of the relationships between the antifreezes from ¢sh in the order Scorpaeniformes.
2. Materials and methods 2.1. Sample collection Sixty-¢ve Atlantic snail¢sh, L. atlanticus, were collected by divers near Logy Bay, Newfoundland in the winters of 1995 and 1996. Three specimens of dusky snail¢sh, L. gibbus, were collected from Placentia Bay, Newfoundland during the winter of 1995. The live ¢sh were brought into the laboratory and placed into holding tanks supplied with ambient temperature seawater prior to blood collection. Fish were anaesthetized using MS-222 and blood was collected from the caudal vein using a heparin containing syringe and 23-gauge needle. Blood samples were centrifuged at 2000 rpm for 10 min and the plasma was removed with a pipette. 2.2. Isolation and puri¢cation of plasma AFPs Pooled plasma samples from Atlantic snail¢sh or from individual dusky snail¢sh (1 ml each) were applied to a Sephadex G-150 gel ¢ltration column (0.9U60 cm) and eluted with 0.1 M NH4 HCO3 . Protein fractions that exhibited antifreeze activity were collected and lyophilized. These were redissolved in 0.1 M NH4 HCO3 and applied a second time to the G-150 column. Following dialysis in 0.1 M NH4 HCO3 and lyophilization, samples were further puri¢ed by reverse-phase HPLC. The partially puri¢ed AFPs were separated on a Nucleosil C8 column (0.46U25 cm). A gradient of 40^62% acetonitrile (solvent A) and 0.1% tri£uoroacetic acid (solvent B) was used with a £ow rate of 1 ml/min. Individual peaks were collected, lyophilized and redissolved in 0.01 M NH4 HCO3 for activity measurements. The puri¢ed proteins were separated on 15% polyacrylamide gels in a Tris-Tricine bu¡er system to SDSPAGE and stained with 0.1% Coomassie brilliant blue R-250. Amino acid analysis and amino terminal sequencing were performed by the Biotechnology Service Centre, Hospital for Sick Children, Toronto, and mass spectrometry (MS) was performed by the Carbohydrate Centre, University of Toronto. 2.3. Measurement of antifreeze activity Antifreeze activity was measured as thermal hyste-
BBAPRO 36423 31-5-01
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
resis using a Clifton Nanolitre Osmometer (Clifton Technical Physics, Hartford, NY), following the procedure of Kao et al. [10]. Thermal hysteresis is de¢ned as the di¡erence between the melting and freezing temperatures (in ³C) of a test solution. L. atlanticus AFP (designated La-AFP) and L. gibbus AFP (designated Lg-AFP) were dissolved in 0.01 M NH4 HCO3 and centrifuged before use. For each sample, measurements were made in triplicate, and the average value taken. Video recordings were made of ice crystals for analysis of crystal morphologies. In some cases, individual frames were captured and the digital images were used to make photographs. 2.4. Circular dichroism spectroscopy Circular dichroism (CD) spectra were recorded on an Aviv Circular Dichroism Spectrometer model 62DS (Lakewood, NJ) in the Department of Medical Biophysics, University of Toronto. Spectra were obtained from 200 to 260 nm, with a 0.5 nm step, 1 nm bandwidth, and 20 s collection time per step. For thermal denaturation pro¢les, ellipticity was measured at 222 nm and data were collected at a rate of 1 s/point for 300 data points. Samples were cooled at a rate of 1³C every 16 s. Lyophilized protein was dissolved in 0.01 M NH4 HCO3 (pH 8.5). The protein concentration was approx. 0.35 mg/ml (La-AFP) and 0.15 mg/ml (Lg-AFP1). CD measurement was carried out using a cuvette of 0.1 cm path length. The following equation was used to calculate the predicted mean residue ellipticity for 100% helix: k a n a r 13 1 n Where n is the chain length and k is a wavelengthdependent factor (2.57 at 222 nm) and [a]r = 339 500 degree cm2 dmol31 . The percentage of helix for AFPs was determined from the average mean residue ellipticity at 222 nm. 3. Results 3.1. Puri¢cation and analysis of Liparis AFPs Analysis of blood plasma from both snail¢sh species indicated that ¢sh contained antifreeze activity
237
during the winter months from January to March. Atlantic snail¢sh had an average thermal hysteresis measurement of 0.73 þ 0.15³C (n = 65) while dusky snail¢sh plasma was measured as 0.92 þ 0.20³C (n = 3). Additionally, a seasonal cycle was observed for Atlantic snail¢sh antifreeze activity since it peaked during February and subsequently decreased during the spring to where it was essentially nonexistent in August (0.05³C). The antifreeze activity found in blood plasma was ¢rst isolated and then puri¢ed using a variety of standard techniques. Partially puri¢ed protein isolated from plasma by Sephadex G-150 gel ¢ltration chromatography could be further resolved by HPLC (Figs. 1 and 2). A single peak from L. atlanticus (designated as La-AFP) was isolated and two separate peaks were resolved for L. gibbus (designated as Lg-AFP1 and Lg-AFP2). These peaks were collected and analysed by SDS-PAGE. Based on these results, it appeared that the collected fractions were puri¢ed to homogeneity since there was a single band on the gel for each of these three protein peaks (Figs. 1 and 2). The minor peaks observed on some HPLC pro¢les were also collected and analysed for the presence of antifreeze activity. Initially, these small peaks appeared to contain some possible antifreeze activity but this activity was subsequently lost when the fractions were lyophilized and re-dissolved in 0.1 M NH4 HCO3 . Additionally, these fractions did not alter ice crystal structure and were not consistently observed between HPLC runs whereas the major AFP peaks were always seen. For these reasons it appears that all AFPs are contained within the major peaks collected and that the small ones were probably due to the presence of salts and other small blood solutes. On SDS-PAGE, the relative molecular weights of La-AFP and Lg-AFP1, Lg-AFP2 were 6.2 kDa (Figs. 1 and 2). Also based on SDS-PAGE, it appeared that the HPLC peaks collected during protein puri¢cation contained single proteins. However, upon further analysis by mass spectrometry this was concluded not to be accurate. It was determined that the single band from La-AFP on the gel was actually two di¡erent proteins having molecular weights of 9344 (major) and 9415 (minor). Amino terminal sequencing of the ¢rst 17 amino acid residues, by Edman degradation (see Table 2), indicated
BBAPRO 36423 31-5-01
238
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
of the individual proteins, possibly during signal peptide cleavage. Analysis of Lg-AFP1 and Lg-AFP2 by SDS-PAGE also indicated that the single bands were individual proteins. However, analysis by mass spec-
Fig. 1. Outline of the puri¢cation of L. atlanticus AFPs. Panel A shows a typical Sephadex G-150 pro¢le of crude blood plasma with fractions that contained antifreeze activity indicated. (B) Sephadex G-150 puri¢ed antifreeze from plasma was separated on a Nucleosil C8 column (25U0.46 cm); £ow rate was set at 1 ml/min with a 40^62% acetonitrile, 0.1% tri£uoroacetic acid gradient. The sole AFP peak found is labelled as L. atlanticus AFP (La-AFP). Panel C is the SDS-PAGE separation of L. atlanticus AFP. A 15% polyacrylamide gel was used with the Tris-Tricine bu¡er system. Lane 1, pooled G-150 column fractions of L. atlanticus AFPs (approx. 150 Wg of total protein); lane 2, HPLC puri¢ed protein (approx. 75 Wg of protein).
that the 71 Da di¡erence was due to a single alanine residue that was missing from the major protein form at the amino terminus. The sequence suggests that the two di¡erent forms of La-AFP are probably due to di¡erences in post-translational processing
Fig. 2. Outline of the puri¢cation of L. gibbus AFPs. Panel A shows a typical Sephadex G-150 pro¢le of crude blood plasma with fractions that contained antifreeze activity indicated. (B) Sephadex G-150 puri¢ed antifreeze from plasma was separated on a Nucleosil C8 column (25U0.46 cm); £ow rate was set at 1 ml/min with a 40^62% acetonitrile, 0.1% tri£uoroacetic acid gradient. The AFP peaks found are labelled as L. gibbus peak 1 (Lg-AFP1) and L. gibbus peak 2 (Lg-AFP2). Panel C is the SDS-PAGE separation of L. gibbus AFPs. A 15% polyacrylamide gel was used with the Tris-Tricine bu¡er system. Lane 1, pooled G-150 column fractions of L. gibbus AFPs (approx. 200 Wg of total protein); lane 2, HPLC puri¢ed Lg-AFP1 (approx. 100 Wg of protein); lane 3, HPLC puri¢ed Lg-AFP2 (approx. 80 Wg of protein).
BBAPRO 36423 31-5-01
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
239
trometry revealed that it was more complicated than initially surmised. Lg-AFP1 actually contained three isoforms: 9646 (major) and two minor isoforms of 9514 and 9814. Additionally, the Lg-AFP2 band actually contained a major isoform weighing 9573 and a minor one at 9742. The observed discrepancy in Mr between electrophoretic and mass spectrometry data may be a direct consequence of the amino acid content of these proteins. The large number of alanine residues would give rise to a larger than average number of SDS molecules associated with the proteins which would a¡ect gel mobility [11]. Repeated attempts to separate the protein mixtures into their individual protein isoforms by HPLC or SDS-PAGE proved to be unsuccessful. 3.2. Amino acid composition The amino acid compositions of snail¢sh AFPs (La-AFP, Lg-AFP1 and Lg-AFP2) are given in Table 1. These proteins are typical of all type I AFPs in that they have a high alanine content, which accounts for between 51 and 59 mol%. The abundance of alanine residues is similar to winter £ounder liver/ skin AFPs which are between 60 and 62% alanine [5] but is somewhat lower than shorthorn sculpin skin AFP which is approx. 70% alanine [12]. An interesting aspect of these antifreeze proteins is their proline content which is 2.5 mol% for La-AFP and approx. 4.2 mol% in Lg-AFP1 and 2 (Table 1). This would correspond to between three and ¢ve proline residues for each of these AFPs, which may disrupt the helix content of the protein backbone since proline is a known helix breaker [13]. 3.3. Antifreeze activity and ice crystal morphology of snail¢sh AFPs All of the snail¢sh AFPs tested here (La-AFP, LgAFP1 and Lg-AFP2) display concentration-dependent thermal hysteresis activities that are not signi¢cantly di¡erent from each other. As shown in Fig. 3, the thermal hysteretic ability of snail¢sh AFPs is intermediate on a molar basis, with activity lower than that of winter £ounder liver HPLC-6 (w£AFP-6) but higher than winter £ounder skin (wfsAFP-2). Owing to the fact that snail¢sh AFPs are nearly 3 times larger than w£AFP-6, their activ-
Fig. 3. Comparison of thermal hysteresis activity curves on (A) molar basis and (B) weight basis. HPLC puri¢ed AFPs from L. atlanticus (La-AFP) and L. gibbus peak 1 (Lg-AFP1) and 2 (Lg-AFP2). Data for winter £ounder plasma (HPLC-6) AFP were taken from [16] while winter £ounder skin AFP (wfsAFP2) data were taken from [5]. Shorthorn sculpin skin (sssAFP-2) data are from [12].
ity on a weight basis is substantially lower than this winter £ounder AFP, as would be expected. However, the activity of snail¢sh AFPs is also lower on a weight basis when compared to recombinant shorthorn sculpin skin AFP (sssAFP-2) that has a similar molecular weight (MW = 9700) [12]. At low protein concentrations, snail¢sh AFPs impart the typical hexagonal bipyramidal shape to ice crystals that other type I AFPs give (Fig. 4A,B) and these are subsequently elongated/sharpened into spicule-like forms at high protein concentrations (Fig. 4C), a characteristic typical of active AFPs.
BBAPRO 36423 31-5-01
240
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
Fig. 4. Ice crystal morphology in the presence of AFPs. Crystals were videotaped at temperatures below their melting point after they had remained a constant size for at least 10 s. Individual frames were captured from tape and converted to image ¢les. (A) La-AFP used at a concentration of approx. 1.25 mM (thermal hysteresis 0.16³C). (B) Lg-AFP2 used at a concentration of approx. 1.5 mM (thermal hysteresis 0.2³C) or (C) at 5 mM concentration (thermal hysteresis 0.6³C).
C
3.4. Secondary structure of Liparis AFPs CD spectral analyses of La-AFP and Lg-AFP2 (Fig. 5A,C) show strong minima at 208 and 222 nm, which are typical of an K-helical secondary structure. Calculation of helix content indicates that La-AFP is 79% K-helical and Lg-AFP2 is 83% Khelical when measured at 0³C. The helix content begins to disappear as the temperature increases and the proteins are fully denatured after 40³C but will regain their full helix content when cooled back down to 0³C, indicating reversible thermal denaturation is occurring. These AFPs become irreversibly unfolded after 70³C. The midpoint of thermal denaturation (Tm ) was calculated as 22³C for La-AFP and 28³C Lg-AFP2 which is similar to that of winter £ounder liver AFP [14]. The thermal denaturation pro¢le (Fig. 5B,D) for Liparis AFPs appears to be cooperative in nature as indicated by the CD spectral data. This property of these AFPs may be associated with the proline residues in the proteins that are known to disrupt helical proteins [13] and seems to indicate that these antifreeze proteins are unfolding as separately distinct helices. 4. Discussion The results of the present study clearly demonstrate that snail¢sh produce AFPs during winter. These proteins are similar to type I AFPs in their high alanine content and K-helical secondary structure, but are signi¢cantly larger than all type I AFPs isolated from any other source. Table 2 gives an overall comparison of the major type I AFPs investigated to date indicating the species from which they originate [5,11,12,26^32]. Fish from the Teleostei superorder Acanthopterygii contain all known types
BBAPRO 36423 31-5-01
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
241
Table 1 Amino acid composition (mol%) and Mr of Liparis AFPs Amino acids Asp Glu Ser Gly Arg Thr Ala Pro Val Ile Leu Phe Lys MWa a b
L. atlanticus La-AFP 3.63 2.95 2.80 4.56 1.57 10.26 58.79 2.50 5.57 1.28 2.55 0.12 3.43 9 344 (major) 9 415b
L. gibbus Lg-AFP1
Lg-AFP2
5.41 2.61 2.03 3.85 1.75 8.93 51.22 4.18 8.41 1.74 2.31 1.00 6.57 9 646 (major) 9 514; 9 814
5.46 2.63 2.05 3.88 0.87 9.01 51.68 4.21 8.48 1.75 2.33 1.01 6.63 9 573 (major) 9 742
Based on mass spectroscopic analysis. 15% of protein molecules contain one fewer Ala residue at the amino terminus.
of ¢sh AFPs/AFGPs and diverse species from two of its orders (Scorpaeniformes and Pleuronectiformes) produce all type I AFPs discovered thus far [4,15]. Here we have demonstrated that another family from the order Scorpaeniformes, namely Cyclopteridae,
produces type I AFPs that circulate in snail¢sh blood plasma. The type I AFPs found in snail¢sh plasma are interesting since although there is high similarity amongst snail¢sh species they are unlike closely re-
Table 2 Species origin and physical properties of type I AFPs
1
15% of protein molecules contain one fewer Ala residue at the amino terminus. MW of native protein. Reported a MW 9.7 kDa for recombinant protein. 3 Activity at 7 mg/ml protein. 2
BBAPRO 36423 31-5-01
242
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
Fig. 5. CD spectra and thermal denaturation pro¢les of L. atlanticus and L. gibbus AFPs. (A) CD spectrum and (B) thermal denaturation curve for L. atlanticus AFP (La-AFP). The protein concentration was approx. 0.35 mg/ml in 0.1 M NH4 HCO3 (pH 8.5) and cell length was 0.1 cm. (C) CD spectrum and (D) thermal denaturation curve for L. gibbus peak 1 (Lg-AFP1). The protein concentration was approx. 0.15 mg/ml in 0.1 M NH4 HCO3 (pH 8.5) and cell length was 0.1 cm.
lated sculpin plasma AFPs in their amino terminal sequence and their large size. This indicates that the snail¢sh AFPs derived from the same progenitor protein prior to the di¡erentiation of these species but that the sculpin AFPs derived from another one. However, we also have evidence that snail¢sh skin AFP sequences are very similar to those from shorthorn sculpin skin AFPs both at the DNA and amino acid levels (R.P. Evans and G.L. Fletcher, unpublished data). Taken together, this information suggests that type I AFPs have emerged on more than one occasion from a common ancestral protein. Only
after an ancestral protein (or proteins) that is common to all type I AFPs is discovered will we know for certain. Furthermore, it would also be helpful to determine the primary nucleotide sequence of snail¢sh plasma AFPs in order to con¢rm the presence of any repeat regions and to identify possible ice binding sites or surfaces to see how they compare to others. The activity of snail¢sh plasma AFPs is quite low when measured on a weight basis (mg/ml) compared to w£-AFP6 and even sssAFP-2 (Fig. 3). However, when their molecular weight is taken into account
BBAPRO 36423 31-5-01
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
when measurements are made (mM), the ability of these AFPs to lower solution freezing points is intermediate compared to other type I AFPs [10,16] but still lower than sssAFP-2 (Fig. 3). These observations suggest that although the snail¢sh AFPs are very large they may possess fewer ice binding surfaces or motifs than do the sssAFP-2 and w£-AFP6. This relationship appears to be consistent for the large type I AFPs from snail¢sh and shorthorn sculpin since these proteins have lower thermal hysteresis activities than their smaller type I counterparts [12]. One puzzling fact that arises from these results is the apparent inability of plasma AFPs to protect snail¢sh from freezing in ice-cold seawater that they would be exposed to in winter. Even at very high concentrations of 20 mg/ml, puri¢ed proteins would provide approx. 0.25³C of freeze protection, which does not meet required levels (up to 0.7³C). This suggests that some other factor is augmenting the plasma AFPs that protect these ¢sh from freezing in extreme environmental conditions. We have evidence that physiological concentrations of typical plasma salts can signi¢cantly enhance the thermal hysteresis activity of puri¢ed type I AFPs from winter £ounder (G.L. Fletcher and S.V. Goddard, unpublished data). These data have clear implications for studies that attempt to quantify the `true' antifreeze capability of a particular type I AFP and when comparisons are made between AFPs isolated by different groups. Other studies have demonstrated that some common blood solutes (i.e. sorbitol, glycerol, alanine etc.) can signi¢cantly enhance the activity of beetle, Dendroides canadensis, AFP [17] while rainbow smelt Osmerus mordax synthesize glycerol as an additional, colligative antifreeze in blood plasma [18,19]. More work needs to be done to determine which mechanism(s) is (are) involved to explain what is happening in snail¢sh. The results of this research, and that of others, suggest that it may be necessary to simplify the definition of type I antifreeze proteins. Since their discovery, this class of AFPs has been de¢ned as small helical proteins/polypeptides (less than 4.5 kDa) that have a speci¢c 11 amino acid repeat motif (Thr-X2 Asn/Asp-X7 ), where X is usually alanine or may be another amino acid that favours helix formation [2^ 4]. It was also proposed that the repeating nature of type I AFPs matched the ice-crystal lattice structure
243
in such a way that polar residues interacted through hydrogen bonding while the hydrophobic side was exposed to the water molecules which prevented further crystal growth [2^4]. However, recent studies have suggested that non-polar interactions may also play important roles in ice binding [20,21] while other groups have examined the roles of putative ice binding motifs (IBM) and what properties are necessary to allow an AFP to function correctly [22,23]. It has also been speculated that motifs that were originally thought to be important for ice binding are actually functioning in protein solubility, and a new site for ice binding has been proposed that involves hydrophobic interactions between the ice surface and the protein hydrophobic surface (i.e. alanine residues) [4,24,25]. The research presented here and other work on shorthorn sculpin skin AFP [12] have shown that both plasma and skin type I AFP can be signi¢cantly larger than previously thought and not necessarily contain the 11 amino acid repeat sequence. Essentially, this class of antifreeze proteins is di¡erentiated by its high alanine content and K-helical secondary structure. Acknowledgements We thank M. King and Dr. M. Shears at the OSC for technical assistance and the OSC divers for sample collection. We also thank Dr. C.L. Hew at University of Toronto for use of lab facilities and helpful guidance in some of the protein analysis and Dr. A. Chakrabartty at Princess Margaret Hospital where CD analysis was performed. This work was supported by a grant from NSERC. References [1] G.L. Fletcher, C.L. Hew, P.L. Davies, Annu. Rev. Physiol. 63 (2001) 359^390. [2] G.L. Fletcher, S.V. Goddard, P.L. Davies, Z. Gong, K.V. Ewart, C.L. Hew, in: H.O. Po«rtner, R.C. Playle (Eds.), New Insights into Fish Antifreeze Proteins: Physiological Signi¢cance and Molecular Regulation, Cambridge University Press, New York, 1998, pp. 240^265. [3] K.V. Ewart, Q. Lin, C.L. Hew, Cell Mol. Life Sci. 55 (1999) 271^283. [4] M.M. Harding, L.G. Ward, A.D. Haymet, Eur. J. Biochem. 264 (1999) 653^665.
BBAPRO 36423 31-5-01
244
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
[5] Z. Gong, K.V. Ewart, Z. Hu, G.L. Fletcher, C.L. Hew, J. Biol. Chem. 271 (1996) 4106^4112. [6] W.B. Scott, M.G. Scott, Atlantic Fishes of Canada, University of Toronto Press, Toronto, 1988. [7] G. Deng, D.W. Andrews, R.A. Laursen, FEBS Lett. 402 (1997) 17^20. [8] Z. Zhao, G. Deng, Q. Lui, R.A. Laursen, Biochim. Biophys. Acta 1382 (1998) 177^180. [9] G. Deng, R.A. Laursen, Biochim. Biophys. Acta 1388 (1998) 305^314. [10] M.H. Kao, G.L. Fletcher, N.C. Wang, C.L. Hew, Can. J. Zool. 64 (1986) 578^582. [11] C.L. Hew, G.L. Fletcher, V.S. Ananthanarayanan, Can. J. Biochem. 58 (1980) 377^383. [12] W.K. Low, M. Miao, K.V. Ewart, D.S. Yang, G.L. Fletcher, C.L. Hew, J. Biol. Chem. 273 (1998) 23098^23103. [13] A. Chakrabartty, R.L. Baldwin, Adv. Protein Chem. 46 (1995) 141^176. [14] D. Wen, R.A. Laursen, J. Biol. Chem. 267 (1992) 14102^ 14108. [15] C.H. Cheng, Curr. Opin. Genet. Dev. 8 (1998) 715^720. [16] K.V. Ewart, G.L. Fletcher, Can. J. Zool. 68 (1990) 1652^ 1658. [17] N. Li, C.A. Andorfer, J.G. Duman, J. Exp. Biol. 201 (1998) 2243^2251. [18] W.R. Driedzic, J.L. West, D.H. Sephton, J.A. Raymond, Fish Physiol. Biochem. 18 (1998) 125^134. [19] J.A. Raymond, W.R. Driedzic, Comp. Biochem. Physiol. 118B (1997) 000^000.
[20] H. Chao, M.E. Houston Jr., R.S. Hodges, C.M. Kay, B.D. Sykes, M.C. Loewen, P.L. Davies, F.D. Sonnichsen, Biochemistry 36 (1997) 14652^14660. [21] A. Cheng, K.M. Merz Jr., Biophys. J. 73 (1997) 2851^2873. [22] Q. Lin, K.V. Ewart, Q. Yan, W.K. Wong, D.S. Yang, C.L. Hew, Eur. J. Biochem. 264 (1999) 49^54. [23] Q. Lin, K.V. Ewart, D.S. Yang, C.L. Hew, FEBS Lett. 453 (1999) 331^334. [24] J. Baardsnes, L.H. Kondejewski, R.S. Hodges, H. Chao, C. Kay, P.L. Davies, FEBS Lett. 463 (1999) 87^91. [25] M.C. Loewen, H. Chao, M.E. Houston Jr., J. Baardsnes, R.S. Hodges, C.M. Kay, B.D. Sykes, F.D. Sonnichsen, P.L. Davies, Biochemistry 38 (1999) 4743^4749. [26] J.G. Duman, A.L. DeVries, J. Exp. Zool. 190 (1974) 89^98. [27] H. Chao, R.S. Hodges, C.M. Kay, S.Y. Gauthier, P.L. Davies, Protein Sci. 5 (1996) 1150^1156. [28] G.K. Scott, P.L. Davies, M.A. Shears, G.L. Fletcher, Eur. J. Biochem. 168 (1987) 629^633. [29] C.A. Knight, C.C. Cheng, A.L. DeVries, Biophys. J. 59 (1991) 409^418. [30] C.L. Hew, S. Joshi, N.C. Wang, M.H. Kao, V.S. Ananthanarayanan, Eur. J. Biochem. 151 (1985) 167^172. [31] H.M. Reisman, G.L. Fletcher, M.H. Kao, M.A. Shears, Environ. Biol. Fishes 18 (1987) 295^301. [32] G.L. Fletcher, R.F. Addison, D. Slaughter, C.L. Hew, Arctic 35 (1982) 302^306.
BBAPRO 36423 31-5-01
Isolation and characterization of type I antifreeze proteins from Atlantic snail¢sh (Liparis atlanticus) and dusky snail¢sh (Liparis gibbus) Robert P. Evans *, Garth L. Fletcher Ocean Sciences Centre, Memorial University of Newfoundland, St. John's, NF A1C 5S7, Canada Received 10 August 2000; received in revised form 1 March 2001; accepted 15 March 2001
Abstract Antifreeze proteins (AFPs) were isolated from the blood plasma of Atlantic snailfish Liparis atlanticus and dusky snailfish Liparis gibbus, which belong to the Teleost family Cyclopteridae, a close relative of sculpins. Using a combination of gel filtration chromatography and reversed-phase HPLC, proteins were purified to individual peaks. Atlantic snailfish plasma contained two different proteins (MW = 9344, 9415) while dusky snailfish plasma contained five protein isoforms (MW = 9514^9814), as determined by mass spectrometry. Further characterization revealed that these proteins are rich in alanine ( s 50 mol%), and have K-helical secondary structure that can undergo reversible thermal denaturation. Thermal hysteresis activities of these proteins were similar to each other but lower than the major type I AFPs from winter flounder. Results of this study have indicated that although the AFPs from snailfish are significantly larger than previously described type I AFPs, they share enough characteristics to be classified in this group. ß 2001 Elsevier Science B.V. All rights reserved. Keywords: Type I; Antifreeze; Protein; Plasma ; Snail¢sh; Liparis
1. Introduction Teleost ¢sh that inhabit icy seawater in winter, at temperatures as low as 31.8³C, synthesize antifreeze proteins/polypeptides (AFPs) or antifreeze glycoproteins (AFGPs) for protection against freezing. Species from diverse taxonomic groups produce AFPs that are classi¢ed as four distinct classes based on their physical characteristics, termed type I, II, III and IV. However, regardless of protein or polypeptide structure, all classes of ¢sh AFPs lower the
* Corresponding author. Fax: +1-709-737-3220; E-mail: [email protected]
freezing point of plasma non-colligatively by binding to certain surfaces of ice crystals, modifying their structure and inhibiting further growth. Extensive reviews have been published recently that deal with general structure/function characteristics, diversity, physiology and molecular genetics of AFPs and AFGPs [1^4]. Of the four classes of AFPs discovered to date, the smallest and simplest class has been the most extensively studied. Type I AFPs are found in certain sculpins (e.g. Myoxocephalus) and unrelated right eye £ounders (Pleuronectes). Structurally, they are usually polypeptides that have high alanine content ( s 60 mol%), have an amphipathic K-helical secondary structure, and are usually small (approx. 3.3^4.5 kDa). The winter £ounder Pleuronectes americanus
0167-4838 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 8 3 8 ( 0 1 ) 0 0 1 9 0 - X
BBAPRO 36423 31-5-01
236
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
has provided an excellent working model for developing an understanding of AFP function and regulation. Winter £ounder synthesize AFP in the liver, which is then secreted into the blood to protect the ¢sh from extracellular freezing. Additionally, this species produces a type I protein in its skin that lacks a peptide signal sequence that would indicate it remains intracellular [5]. The di¡erent sites of action suggest an alternative biological function for serum and skin AFPs. The snail¢shes belong to a large family of benthic and pelagic marine ¢shes, Cyclopteridae, that inhabit northern regions of the Atlantic, Paci¢c and Arctic oceans. This family is closely related to the sculpins, which belong to a di¡erent family of the same order Scorpaeniformes. The Atlantic snail¢sh, Liparis atlanticus, is a very small species, with mature adults ranging in length from 10 to 95 mm. They inhabit inshore waters of the northwest Atlantic Ocean from Ungava Bay in northern Quebec and south to New York [6]. During the winter months, January to March, ¢sh move further inshore where females spawn, in depths less than 2 m. Dusky snail¢sh Liparis gibbus is a larger species, averaging 114 mm in length. Their habitat extends from the Canadian Arctic Ocean, across to Greenland and down to the coast of Newfoundland where they live mainly on muddy bottoms at depths up to 200 m [6]. Because their harsh natural environments would frequently expose both of these snail¢sh species to ice laden seawater, they were considered excellent candidates for possession of AFPs. It was originally hypothesized that the genus Liparis might produce either type I or type II AFPs, since closely related species from the order Scorpaeniformes produce both classes of AFPs. Shorthorn sculpin, Myoxocephalus scorpius, and grubby sculpin, Myoxocephalus aenaeus, have type I AFP while sea raven, Hemitripterus americanus, has type II AFP in their blood plasma. Furthermore, it has recently been shown that another Scorpaeniforme, the longhorn sculpin Myoxocephalus octodecimspinosis, produces an entirely unrelated antifreeze protein in its blood plasma ^ type IV [7^9]. For these reasons, we have puri¢ed and characterized AFPs from snail¢sh to help elucidate their evolutionary origins and to clarify some of the relationships between the antifreezes from ¢sh in the order Scorpaeniformes.
2. Materials and methods 2.1. Sample collection Sixty-¢ve Atlantic snail¢sh, L. atlanticus, were collected by divers near Logy Bay, Newfoundland in the winters of 1995 and 1996. Three specimens of dusky snail¢sh, L. gibbus, were collected from Placentia Bay, Newfoundland during the winter of 1995. The live ¢sh were brought into the laboratory and placed into holding tanks supplied with ambient temperature seawater prior to blood collection. Fish were anaesthetized using MS-222 and blood was collected from the caudal vein using a heparin containing syringe and 23-gauge needle. Blood samples were centrifuged at 2000 rpm for 10 min and the plasma was removed with a pipette. 2.2. Isolation and puri¢cation of plasma AFPs Pooled plasma samples from Atlantic snail¢sh or from individual dusky snail¢sh (1 ml each) were applied to a Sephadex G-150 gel ¢ltration column (0.9U60 cm) and eluted with 0.1 M NH4 HCO3 . Protein fractions that exhibited antifreeze activity were collected and lyophilized. These were redissolved in 0.1 M NH4 HCO3 and applied a second time to the G-150 column. Following dialysis in 0.1 M NH4 HCO3 and lyophilization, samples were further puri¢ed by reverse-phase HPLC. The partially puri¢ed AFPs were separated on a Nucleosil C8 column (0.46U25 cm). A gradient of 40^62% acetonitrile (solvent A) and 0.1% tri£uoroacetic acid (solvent B) was used with a £ow rate of 1 ml/min. Individual peaks were collected, lyophilized and redissolved in 0.01 M NH4 HCO3 for activity measurements. The puri¢ed proteins were separated on 15% polyacrylamide gels in a Tris-Tricine bu¡er system to SDSPAGE and stained with 0.1% Coomassie brilliant blue R-250. Amino acid analysis and amino terminal sequencing were performed by the Biotechnology Service Centre, Hospital for Sick Children, Toronto, and mass spectrometry (MS) was performed by the Carbohydrate Centre, University of Toronto. 2.3. Measurement of antifreeze activity Antifreeze activity was measured as thermal hyste-
BBAPRO 36423 31-5-01
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
resis using a Clifton Nanolitre Osmometer (Clifton Technical Physics, Hartford, NY), following the procedure of Kao et al. [10]. Thermal hysteresis is de¢ned as the di¡erence between the melting and freezing temperatures (in ³C) of a test solution. L. atlanticus AFP (designated La-AFP) and L. gibbus AFP (designated Lg-AFP) were dissolved in 0.01 M NH4 HCO3 and centrifuged before use. For each sample, measurements were made in triplicate, and the average value taken. Video recordings were made of ice crystals for analysis of crystal morphologies. In some cases, individual frames were captured and the digital images were used to make photographs. 2.4. Circular dichroism spectroscopy Circular dichroism (CD) spectra were recorded on an Aviv Circular Dichroism Spectrometer model 62DS (Lakewood, NJ) in the Department of Medical Biophysics, University of Toronto. Spectra were obtained from 200 to 260 nm, with a 0.5 nm step, 1 nm bandwidth, and 20 s collection time per step. For thermal denaturation pro¢les, ellipticity was measured at 222 nm and data were collected at a rate of 1 s/point for 300 data points. Samples were cooled at a rate of 1³C every 16 s. Lyophilized protein was dissolved in 0.01 M NH4 HCO3 (pH 8.5). The protein concentration was approx. 0.35 mg/ml (La-AFP) and 0.15 mg/ml (Lg-AFP1). CD measurement was carried out using a cuvette of 0.1 cm path length. The following equation was used to calculate the predicted mean residue ellipticity for 100% helix: k a n a r 13 1 n Where n is the chain length and k is a wavelengthdependent factor (2.57 at 222 nm) and [a]r = 339 500 degree cm2 dmol31 . The percentage of helix for AFPs was determined from the average mean residue ellipticity at 222 nm. 3. Results 3.1. Puri¢cation and analysis of Liparis AFPs Analysis of blood plasma from both snail¢sh species indicated that ¢sh contained antifreeze activity
237
during the winter months from January to March. Atlantic snail¢sh had an average thermal hysteresis measurement of 0.73 þ 0.15³C (n = 65) while dusky snail¢sh plasma was measured as 0.92 þ 0.20³C (n = 3). Additionally, a seasonal cycle was observed for Atlantic snail¢sh antifreeze activity since it peaked during February and subsequently decreased during the spring to where it was essentially nonexistent in August (0.05³C). The antifreeze activity found in blood plasma was ¢rst isolated and then puri¢ed using a variety of standard techniques. Partially puri¢ed protein isolated from plasma by Sephadex G-150 gel ¢ltration chromatography could be further resolved by HPLC (Figs. 1 and 2). A single peak from L. atlanticus (designated as La-AFP) was isolated and two separate peaks were resolved for L. gibbus (designated as Lg-AFP1 and Lg-AFP2). These peaks were collected and analysed by SDS-PAGE. Based on these results, it appeared that the collected fractions were puri¢ed to homogeneity since there was a single band on the gel for each of these three protein peaks (Figs. 1 and 2). The minor peaks observed on some HPLC pro¢les were also collected and analysed for the presence of antifreeze activity. Initially, these small peaks appeared to contain some possible antifreeze activity but this activity was subsequently lost when the fractions were lyophilized and re-dissolved in 0.1 M NH4 HCO3 . Additionally, these fractions did not alter ice crystal structure and were not consistently observed between HPLC runs whereas the major AFP peaks were always seen. For these reasons it appears that all AFPs are contained within the major peaks collected and that the small ones were probably due to the presence of salts and other small blood solutes. On SDS-PAGE, the relative molecular weights of La-AFP and Lg-AFP1, Lg-AFP2 were 6.2 kDa (Figs. 1 and 2). Also based on SDS-PAGE, it appeared that the HPLC peaks collected during protein puri¢cation contained single proteins. However, upon further analysis by mass spectrometry this was concluded not to be accurate. It was determined that the single band from La-AFP on the gel was actually two di¡erent proteins having molecular weights of 9344 (major) and 9415 (minor). Amino terminal sequencing of the ¢rst 17 amino acid residues, by Edman degradation (see Table 2), indicated
BBAPRO 36423 31-5-01
238
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
of the individual proteins, possibly during signal peptide cleavage. Analysis of Lg-AFP1 and Lg-AFP2 by SDS-PAGE also indicated that the single bands were individual proteins. However, analysis by mass spec-
Fig. 1. Outline of the puri¢cation of L. atlanticus AFPs. Panel A shows a typical Sephadex G-150 pro¢le of crude blood plasma with fractions that contained antifreeze activity indicated. (B) Sephadex G-150 puri¢ed antifreeze from plasma was separated on a Nucleosil C8 column (25U0.46 cm); £ow rate was set at 1 ml/min with a 40^62% acetonitrile, 0.1% tri£uoroacetic acid gradient. The sole AFP peak found is labelled as L. atlanticus AFP (La-AFP). Panel C is the SDS-PAGE separation of L. atlanticus AFP. A 15% polyacrylamide gel was used with the Tris-Tricine bu¡er system. Lane 1, pooled G-150 column fractions of L. atlanticus AFPs (approx. 150 Wg of total protein); lane 2, HPLC puri¢ed protein (approx. 75 Wg of protein).
that the 71 Da di¡erence was due to a single alanine residue that was missing from the major protein form at the amino terminus. The sequence suggests that the two di¡erent forms of La-AFP are probably due to di¡erences in post-translational processing
Fig. 2. Outline of the puri¢cation of L. gibbus AFPs. Panel A shows a typical Sephadex G-150 pro¢le of crude blood plasma with fractions that contained antifreeze activity indicated. (B) Sephadex G-150 puri¢ed antifreeze from plasma was separated on a Nucleosil C8 column (25U0.46 cm); £ow rate was set at 1 ml/min with a 40^62% acetonitrile, 0.1% tri£uoroacetic acid gradient. The AFP peaks found are labelled as L. gibbus peak 1 (Lg-AFP1) and L. gibbus peak 2 (Lg-AFP2). Panel C is the SDS-PAGE separation of L. gibbus AFPs. A 15% polyacrylamide gel was used with the Tris-Tricine bu¡er system. Lane 1, pooled G-150 column fractions of L. gibbus AFPs (approx. 200 Wg of total protein); lane 2, HPLC puri¢ed Lg-AFP1 (approx. 100 Wg of protein); lane 3, HPLC puri¢ed Lg-AFP2 (approx. 80 Wg of protein).
BBAPRO 36423 31-5-01
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
239
trometry revealed that it was more complicated than initially surmised. Lg-AFP1 actually contained three isoforms: 9646 (major) and two minor isoforms of 9514 and 9814. Additionally, the Lg-AFP2 band actually contained a major isoform weighing 9573 and a minor one at 9742. The observed discrepancy in Mr between electrophoretic and mass spectrometry data may be a direct consequence of the amino acid content of these proteins. The large number of alanine residues would give rise to a larger than average number of SDS molecules associated with the proteins which would a¡ect gel mobility [11]. Repeated attempts to separate the protein mixtures into their individual protein isoforms by HPLC or SDS-PAGE proved to be unsuccessful. 3.2. Amino acid composition The amino acid compositions of snail¢sh AFPs (La-AFP, Lg-AFP1 and Lg-AFP2) are given in Table 1. These proteins are typical of all type I AFPs in that they have a high alanine content, which accounts for between 51 and 59 mol%. The abundance of alanine residues is similar to winter £ounder liver/ skin AFPs which are between 60 and 62% alanine [5] but is somewhat lower than shorthorn sculpin skin AFP which is approx. 70% alanine [12]. An interesting aspect of these antifreeze proteins is their proline content which is 2.5 mol% for La-AFP and approx. 4.2 mol% in Lg-AFP1 and 2 (Table 1). This would correspond to between three and ¢ve proline residues for each of these AFPs, which may disrupt the helix content of the protein backbone since proline is a known helix breaker [13]. 3.3. Antifreeze activity and ice crystal morphology of snail¢sh AFPs All of the snail¢sh AFPs tested here (La-AFP, LgAFP1 and Lg-AFP2) display concentration-dependent thermal hysteresis activities that are not signi¢cantly di¡erent from each other. As shown in Fig. 3, the thermal hysteretic ability of snail¢sh AFPs is intermediate on a molar basis, with activity lower than that of winter £ounder liver HPLC-6 (w£AFP-6) but higher than winter £ounder skin (wfsAFP-2). Owing to the fact that snail¢sh AFPs are nearly 3 times larger than w£AFP-6, their activ-
Fig. 3. Comparison of thermal hysteresis activity curves on (A) molar basis and (B) weight basis. HPLC puri¢ed AFPs from L. atlanticus (La-AFP) and L. gibbus peak 1 (Lg-AFP1) and 2 (Lg-AFP2). Data for winter £ounder plasma (HPLC-6) AFP were taken from [16] while winter £ounder skin AFP (wfsAFP2) data were taken from [5]. Shorthorn sculpin skin (sssAFP-2) data are from [12].
ity on a weight basis is substantially lower than this winter £ounder AFP, as would be expected. However, the activity of snail¢sh AFPs is also lower on a weight basis when compared to recombinant shorthorn sculpin skin AFP (sssAFP-2) that has a similar molecular weight (MW = 9700) [12]. At low protein concentrations, snail¢sh AFPs impart the typical hexagonal bipyramidal shape to ice crystals that other type I AFPs give (Fig. 4A,B) and these are subsequently elongated/sharpened into spicule-like forms at high protein concentrations (Fig. 4C), a characteristic typical of active AFPs.
BBAPRO 36423 31-5-01
240
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
Fig. 4. Ice crystal morphology in the presence of AFPs. Crystals were videotaped at temperatures below their melting point after they had remained a constant size for at least 10 s. Individual frames were captured from tape and converted to image ¢les. (A) La-AFP used at a concentration of approx. 1.25 mM (thermal hysteresis 0.16³C). (B) Lg-AFP2 used at a concentration of approx. 1.5 mM (thermal hysteresis 0.2³C) or (C) at 5 mM concentration (thermal hysteresis 0.6³C).
C
3.4. Secondary structure of Liparis AFPs CD spectral analyses of La-AFP and Lg-AFP2 (Fig. 5A,C) show strong minima at 208 and 222 nm, which are typical of an K-helical secondary structure. Calculation of helix content indicates that La-AFP is 79% K-helical and Lg-AFP2 is 83% Khelical when measured at 0³C. The helix content begins to disappear as the temperature increases and the proteins are fully denatured after 40³C but will regain their full helix content when cooled back down to 0³C, indicating reversible thermal denaturation is occurring. These AFPs become irreversibly unfolded after 70³C. The midpoint of thermal denaturation (Tm ) was calculated as 22³C for La-AFP and 28³C Lg-AFP2 which is similar to that of winter £ounder liver AFP [14]. The thermal denaturation pro¢le (Fig. 5B,D) for Liparis AFPs appears to be cooperative in nature as indicated by the CD spectral data. This property of these AFPs may be associated with the proline residues in the proteins that are known to disrupt helical proteins [13] and seems to indicate that these antifreeze proteins are unfolding as separately distinct helices. 4. Discussion The results of the present study clearly demonstrate that snail¢sh produce AFPs during winter. These proteins are similar to type I AFPs in their high alanine content and K-helical secondary structure, but are signi¢cantly larger than all type I AFPs isolated from any other source. Table 2 gives an overall comparison of the major type I AFPs investigated to date indicating the species from which they originate [5,11,12,26^32]. Fish from the Teleostei superorder Acanthopterygii contain all known types
BBAPRO 36423 31-5-01
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
241
Table 1 Amino acid composition (mol%) and Mr of Liparis AFPs Amino acids Asp Glu Ser Gly Arg Thr Ala Pro Val Ile Leu Phe Lys MWa a b
L. atlanticus La-AFP 3.63 2.95 2.80 4.56 1.57 10.26 58.79 2.50 5.57 1.28 2.55 0.12 3.43 9 344 (major) 9 415b
L. gibbus Lg-AFP1
Lg-AFP2
5.41 2.61 2.03 3.85 1.75 8.93 51.22 4.18 8.41 1.74 2.31 1.00 6.57 9 646 (major) 9 514; 9 814
5.46 2.63 2.05 3.88 0.87 9.01 51.68 4.21 8.48 1.75 2.33 1.01 6.63 9 573 (major) 9 742
Based on mass spectroscopic analysis. 15% of protein molecules contain one fewer Ala residue at the amino terminus.
of ¢sh AFPs/AFGPs and diverse species from two of its orders (Scorpaeniformes and Pleuronectiformes) produce all type I AFPs discovered thus far [4,15]. Here we have demonstrated that another family from the order Scorpaeniformes, namely Cyclopteridae,
produces type I AFPs that circulate in snail¢sh blood plasma. The type I AFPs found in snail¢sh plasma are interesting since although there is high similarity amongst snail¢sh species they are unlike closely re-
Table 2 Species origin and physical properties of type I AFPs
1
15% of protein molecules contain one fewer Ala residue at the amino terminus. MW of native protein. Reported a MW 9.7 kDa for recombinant protein. 3 Activity at 7 mg/ml protein. 2
BBAPRO 36423 31-5-01
242
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
Fig. 5. CD spectra and thermal denaturation pro¢les of L. atlanticus and L. gibbus AFPs. (A) CD spectrum and (B) thermal denaturation curve for L. atlanticus AFP (La-AFP). The protein concentration was approx. 0.35 mg/ml in 0.1 M NH4 HCO3 (pH 8.5) and cell length was 0.1 cm. (C) CD spectrum and (D) thermal denaturation curve for L. gibbus peak 1 (Lg-AFP1). The protein concentration was approx. 0.15 mg/ml in 0.1 M NH4 HCO3 (pH 8.5) and cell length was 0.1 cm.
lated sculpin plasma AFPs in their amino terminal sequence and their large size. This indicates that the snail¢sh AFPs derived from the same progenitor protein prior to the di¡erentiation of these species but that the sculpin AFPs derived from another one. However, we also have evidence that snail¢sh skin AFP sequences are very similar to those from shorthorn sculpin skin AFPs both at the DNA and amino acid levels (R.P. Evans and G.L. Fletcher, unpublished data). Taken together, this information suggests that type I AFPs have emerged on more than one occasion from a common ancestral protein. Only
after an ancestral protein (or proteins) that is common to all type I AFPs is discovered will we know for certain. Furthermore, it would also be helpful to determine the primary nucleotide sequence of snail¢sh plasma AFPs in order to con¢rm the presence of any repeat regions and to identify possible ice binding sites or surfaces to see how they compare to others. The activity of snail¢sh plasma AFPs is quite low when measured on a weight basis (mg/ml) compared to w£-AFP6 and even sssAFP-2 (Fig. 3). However, when their molecular weight is taken into account
BBAPRO 36423 31-5-01
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
when measurements are made (mM), the ability of these AFPs to lower solution freezing points is intermediate compared to other type I AFPs [10,16] but still lower than sssAFP-2 (Fig. 3). These observations suggest that although the snail¢sh AFPs are very large they may possess fewer ice binding surfaces or motifs than do the sssAFP-2 and w£-AFP6. This relationship appears to be consistent for the large type I AFPs from snail¢sh and shorthorn sculpin since these proteins have lower thermal hysteresis activities than their smaller type I counterparts [12]. One puzzling fact that arises from these results is the apparent inability of plasma AFPs to protect snail¢sh from freezing in ice-cold seawater that they would be exposed to in winter. Even at very high concentrations of 20 mg/ml, puri¢ed proteins would provide approx. 0.25³C of freeze protection, which does not meet required levels (up to 0.7³C). This suggests that some other factor is augmenting the plasma AFPs that protect these ¢sh from freezing in extreme environmental conditions. We have evidence that physiological concentrations of typical plasma salts can signi¢cantly enhance the thermal hysteresis activity of puri¢ed type I AFPs from winter £ounder (G.L. Fletcher and S.V. Goddard, unpublished data). These data have clear implications for studies that attempt to quantify the `true' antifreeze capability of a particular type I AFP and when comparisons are made between AFPs isolated by different groups. Other studies have demonstrated that some common blood solutes (i.e. sorbitol, glycerol, alanine etc.) can signi¢cantly enhance the activity of beetle, Dendroides canadensis, AFP [17] while rainbow smelt Osmerus mordax synthesize glycerol as an additional, colligative antifreeze in blood plasma [18,19]. More work needs to be done to determine which mechanism(s) is (are) involved to explain what is happening in snail¢sh. The results of this research, and that of others, suggest that it may be necessary to simplify the definition of type I antifreeze proteins. Since their discovery, this class of AFPs has been de¢ned as small helical proteins/polypeptides (less than 4.5 kDa) that have a speci¢c 11 amino acid repeat motif (Thr-X2 Asn/Asp-X7 ), where X is usually alanine or may be another amino acid that favours helix formation [2^ 4]. It was also proposed that the repeating nature of type I AFPs matched the ice-crystal lattice structure
243
in such a way that polar residues interacted through hydrogen bonding while the hydrophobic side was exposed to the water molecules which prevented further crystal growth [2^4]. However, recent studies have suggested that non-polar interactions may also play important roles in ice binding [20,21] while other groups have examined the roles of putative ice binding motifs (IBM) and what properties are necessary to allow an AFP to function correctly [22,23]. It has also been speculated that motifs that were originally thought to be important for ice binding are actually functioning in protein solubility, and a new site for ice binding has been proposed that involves hydrophobic interactions between the ice surface and the protein hydrophobic surface (i.e. alanine residues) [4,24,25]. The research presented here and other work on shorthorn sculpin skin AFP [12] have shown that both plasma and skin type I AFP can be signi¢cantly larger than previously thought and not necessarily contain the 11 amino acid repeat sequence. Essentially, this class of antifreeze proteins is di¡erentiated by its high alanine content and K-helical secondary structure. Acknowledgements We thank M. King and Dr. M. Shears at the OSC for technical assistance and the OSC divers for sample collection. We also thank Dr. C.L. Hew at University of Toronto for use of lab facilities and helpful guidance in some of the protein analysis and Dr. A. Chakrabartty at Princess Margaret Hospital where CD analysis was performed. This work was supported by a grant from NSERC. References [1] G.L. Fletcher, C.L. Hew, P.L. Davies, Annu. Rev. Physiol. 63 (2001) 359^390. [2] G.L. Fletcher, S.V. Goddard, P.L. Davies, Z. Gong, K.V. Ewart, C.L. Hew, in: H.O. Po«rtner, R.C. Playle (Eds.), New Insights into Fish Antifreeze Proteins: Physiological Signi¢cance and Molecular Regulation, Cambridge University Press, New York, 1998, pp. 240^265. [3] K.V. Ewart, Q. Lin, C.L. Hew, Cell Mol. Life Sci. 55 (1999) 271^283. [4] M.M. Harding, L.G. Ward, A.D. Haymet, Eur. J. Biochem. 264 (1999) 653^665.
BBAPRO 36423 31-5-01
244
R.P. Evans, G.L. Fletcher / Biochimica et Biophysica Acta 1547 (2001) 235^244
[5] Z. Gong, K.V. Ewart, Z. Hu, G.L. Fletcher, C.L. Hew, J. Biol. Chem. 271 (1996) 4106^4112. [6] W.B. Scott, M.G. Scott, Atlantic Fishes of Canada, University of Toronto Press, Toronto, 1988. [7] G. Deng, D.W. Andrews, R.A. Laursen, FEBS Lett. 402 (1997) 17^20. [8] Z. Zhao, G. Deng, Q. Lui, R.A. Laursen, Biochim. Biophys. Acta 1382 (1998) 177^180. [9] G. Deng, R.A. Laursen, Biochim. Biophys. Acta 1388 (1998) 305^314. [10] M.H. Kao, G.L. Fletcher, N.C. Wang, C.L. Hew, Can. J. Zool. 64 (1986) 578^582. [11] C.L. Hew, G.L. Fletcher, V.S. Ananthanarayanan, Can. J. Biochem. 58 (1980) 377^383. [12] W.K. Low, M. Miao, K.V. Ewart, D.S. Yang, G.L. Fletcher, C.L. Hew, J. Biol. Chem. 273 (1998) 23098^23103. [13] A. Chakrabartty, R.L. Baldwin, Adv. Protein Chem. 46 (1995) 141^176. [14] D. Wen, R.A. Laursen, J. Biol. Chem. 267 (1992) 14102^ 14108. [15] C.H. Cheng, Curr. Opin. Genet. Dev. 8 (1998) 715^720. [16] K.V. Ewart, G.L. Fletcher, Can. J. Zool. 68 (1990) 1652^ 1658. [17] N. Li, C.A. Andorfer, J.G. Duman, J. Exp. Biol. 201 (1998) 2243^2251. [18] W.R. Driedzic, J.L. West, D.H. Sephton, J.A. Raymond, Fish Physiol. Biochem. 18 (1998) 125^134. [19] J.A. Raymond, W.R. Driedzic, Comp. Biochem. Physiol. 118B (1997) 000^000.
[20] H. Chao, M.E. Houston Jr., R.S. Hodges, C.M. Kay, B.D. Sykes, M.C. Loewen, P.L. Davies, F.D. Sonnichsen, Biochemistry 36 (1997) 14652^14660. [21] A. Cheng, K.M. Merz Jr., Biophys. J. 73 (1997) 2851^2873. [22] Q. Lin, K.V. Ewart, Q. Yan, W.K. Wong, D.S. Yang, C.L. Hew, Eur. J. Biochem. 264 (1999) 49^54. [23] Q. Lin, K.V. Ewart, D.S. Yang, C.L. Hew, FEBS Lett. 453 (1999) 331^334. [24] J. Baardsnes, L.H. Kondejewski, R.S. Hodges, H. Chao, C. Kay, P.L. Davies, FEBS Lett. 463 (1999) 87^91. [25] M.C. Loewen, H. Chao, M.E. Houston Jr., J. Baardsnes, R.S. Hodges, C.M. Kay, B.D. Sykes, F.D. Sonnichsen, P.L. Davies, Biochemistry 38 (1999) 4743^4749. [26] J.G. Duman, A.L. DeVries, J. Exp. Zool. 190 (1974) 89^98. [27] H. Chao, R.S. Hodges, C.M. Kay, S.Y. Gauthier, P.L. Davies, Protein Sci. 5 (1996) 1150^1156. [28] G.K. Scott, P.L. Davies, M.A. Shears, G.L. Fletcher, Eur. J. Biochem. 168 (1987) 629^633. [29] C.A. Knight, C.C. Cheng, A.L. DeVries, Biophys. J. 59 (1991) 409^418. [30] C.L. Hew, S. Joshi, N.C. Wang, M.H. Kao, V.S. Ananthanarayanan, Eur. J. Biochem. 151 (1985) 167^172. [31] H.M. Reisman, G.L. Fletcher, M.H. Kao, M.A. Shears, Environ. Biol. Fishes 18 (1987) 295^301. [32] G.L. Fletcher, R.F. Addison, D. Slaughter, C.L. Hew, Arctic 35 (1982) 302^306.
BBAPRO 36423 31-5-01