INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Many cold water marine fishes produce antifreeze proteins/polypeptides (AFPs)¹ or antifreeze glycoproteins (AFGPs) to prevent freezing in icy sea waters (1-3). Whereas AFGPs from different fish species are similar in structure, the AFPs are structurally diverse. Biochemical characterization has grouped the AFPs into four distinct groups: the type I -helical AFPs from righteye flounders and shorthorn sculpins; type II lectin-like AFPs from sea raven and herring; type III globular AFPs from eelpouts (4-6); and more recently, a type IV helix bundle AFP has been reported from longhorn sculpin, Myoxocephalus octodecimspinosis (7).

Most, if not all of the antifreeze proteins were isolated and characterized from the sera and synthesized in the liver. Recently, our laboratories have demonstrated the presence of new isoforms of type I AFPs in the winter flounder, Pleuronectes americanus, that are synthesized in the peripheral tissues such as the skin and gills (8). These skin-type AFPs are encoded by a distinct set of AFP genes that lack the signal peptide, which is indicative of an intracellular location. The presence of both extracellular and intracellular AFPs with differential tissue expression within a single fish species has raised questions about the relative roles of these AFP isoforms in freezing protection, their structure/function, and evolutionary relationships. These findings have prompted us to reexamine the presence of skin-type AFPs in other fishes. Shorthorn sculpin, like the winter flounder, has been found to produce type I -helical AFPs (9-11), and antifreeze peptides have been isolated from the skin of European shorthorn sculpin (12). The latter peptides, however, were not further characterized. Therefore, the present investigation attempts to clarify the occurrence, structure/function relationship, and tissue distribution of AFPs in the shorthorn sculpin.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Materials-- Tissues from shorthorn sculpin (Myoxocephalus scorpius) were collected from Conception Bay, Newfoundland, Canada. The tissues of freshly killed fish were collected and stored in liquid nitrogen, shipped on dry ice, and stored at 70 °C before use.

Library Construction and Screening-- Total skin RNA from a shorthorn sculpin collected on May 30, 1995 was isolated by acid guanidinium thiocyanate-phenol-chloroform extraction as described by Gong et al. (13). Skin Poly(A⁺) RNA was obtained using a Micro-FastTract^TM mRNA isolation kit (Invitrogen, San Diego, CA). A skin cDNA library was constructed and screened as described by the manufacturer using a ZAP-cDNA synthesis kit and Uni-ZAP^TM Cloning kit (Stratagene, La Jolla, CA). Nylon membranes (Colony/Plaque Screen^TM , Biotechnology Systems/NEN Life Science Products, Boston, MA) were hybridized in a buffer of: 4× SET (0.6 M NaCl, 120 mM Tris, pH 8.0, 8 mM EDTA), 0.4% NaPP_i, 25 mM PB buffer, 0.5% SDS, 9% dextran sulfate, 50% formamide, 250 mg/ml denatured calf thymus DNA and ~0.5 × 10⁶ cpm/ml of probe, at 42 °C for 15 h. Probe was labeled with [-³²P]dATP or dCTP using a Random Primers DNA Labeling System kit (Life Technologies, Gaithersburg, MD). The final wash was performed in 0.2× SSC (30 mM NaCl, 30 mM sodium citrate, pH 7.0), and 0.5% SDS, at 50 °C for 15 min. A 260-bp clone corresponding to winter flounder skin AFP (8) was used as a probe to screen approximately 6.0 × 10⁴ clones of the primary library. DNA sequencing was performed using a double-stranded Nested Deletion Kit (Pharmacia Biotech, Baie d'Urfé, Quebec).

Northern Analysis-- Total RNA from various tissues of three shorthorn sculpin collected on May 30, 1995, March 6, 1996 and September 4, 1997 were isolated using TRIzol^TM Reagent as described by the manufucturer (Life Technologies, Inc.). Five µg of total RNA from each tissue were separated in a denaturing (37% formaldehyde) 1.3% agarose gel, transferred to Hybond^TM -N (Amersham Pharmacia Biotech) nylon membranes, and UV-cross-linked. The hybridization solution consisted of 40% formamide, 5% dextran sulfate, 4× SSC, 7 mM Tris, pH 7.5, 1% SDS, 1× Denhardt's buffer, 100 mg/ml denatured calf thymus DNA, and ~1.0 × 10⁶ cpm/ml of labeled probe. A PstI/SmaI-digested fragment of the s3-2 3'-UTR (437 bp), that included the last 15 bp of the ORF but lacked the poly(A⁺) tail, was used as the probe. Both calf thymus DNA and probe were heat-denatured before use. Hybridization was performed at 60 °C with overnight incubation. Washing conditions began at room temperature and finished at 72 °C in solutions ranging from 1× SSC-1% SDS to 0.1× SSC-0.1% SDS-with 15-min incubations in each solution. Signal was identified by autoradiography.

RT-PCR-- One µg of total RNA from various tissues was combined with 0.5 µg of a primer, 5'-AGCTCCGGTCTGAACTTCAA-3', complimentary to a region in the 3'-UTR of s3-2. Reverse transcription was performed using Moloney murine leukemia virus RT as described by the manufacturer (Life Technologies). One tenth of the RT mixture was PCR amplified using the same primer plus a second primer, 5'-TGCGTAGCAGTGTCTCCGTA-3', corresponding to the 3'-UTR region adjacent to the ORF. The amplification consisted of 30 cycles at 94 °C (60 s), 70 °C (90 s), and 72 °C (105 s). PCR products were resolved in a 1.2% agarose gel with EtBr staining. The identity of the products was confirmed by Southern blotting and hybridization with the 437-bp 3'-UTR probe.

Primer Extension-- A 22-mer primer complimentary to the 5'-end of s3-2, 5'-CAACAACTTCCTGATGAGTCAC-3', was synthesized and labeled with [-³²P]dATP by T4 polynucleotide kinase (New England Biolabs, Beverly, MA) to a specific activity of ~1.0 × 10⁵ cpm/ng. Primer extension was performed as described by Boorstein and Craig (14) using Moloney murine leukemia virus RT (Life Technologies) Primer extension products were precipitated with EtOH and sodium acetate, redissolved in formamide loading buffer, separated in a denaturing polyacrylamide gel, and visualized by autoradiography.

Construction of pET-15b-s3-2-- PCR primers were designed in accordance with the s3-2 cDNA sequence and the cloning sites of pET System expression vector pET-15b (Novagen, Madison, WI); 5'-GCGGCAGCCATATGGCGGCGGCGGCGAAG-3' (upper strand), 5'-GCAGCCGGATCCTCGAGACACTGCTACGC-3' (lower strand), where the underlined regions represent the NdeI and BamHI/XhoI restriction sites. PCR was performed using pfu DNA polymerase (Stratagene). The amplification procedure consisted of 20 cycles at 98 °C (45 s), 66 °C (20 s), and 78 °C (60 s). The resultant 313-bp PCR-synthesized fragment and pET-15b vector were both double digested with NdeI and BamHI, recovered from low melting temperature agarose gel (FMC BioProducts, Rockland, ME), and ligated using T4 DNA ligase (Amersham Pharmacia Biotech). Initial cloning work was performed in DH5 cells (Escherichia coli), and sequence of the insert was confirmed by using a ^T7Sequencing^TM kit (Amersham Pharmacia Biotech).

Expression and Purification of His-sssAFP-2 Recombinant Protein-- Induced expression of the recombinant protein (His-sssAFP-2) was performed in 1 liter of BL21(DE3) cells as described by the manufacturer (Novagen, Madison, WI). Bacteria cells were harvested by centrifugation at 5000 × g for 5 min, and the cell pellet was resuspended in 4 ml of ice-cold binding buffer (5 mM imidazole, 0.5 M NaCl, and 80 mM Tris-HCl, pH 7.9)/100 ml of cell paste. The bacterial pellet was subjected to ultrasound sonication, and cell debris was removed by centrifugation. The supernatant was loaded onto a prepacked and equilibrated Ni²⁺ column. Recombinant protein was then purified and eluted according to the manufacturer's instructions (Novagen). Following dialysis in 0.1 M NH₄HCO₃ and lyophilization, the sample was further purified by C₁₈ reverse-phase high performance liquid chromatography. The purified protein was subjected to SDS-PAGE, amino acid analysis, protein sequencing, and mass spectroscopy. Amino acid analysis and protein sequencing were performed by the Biotechnology Service Center, Hospital for Sick Children, Toronto, and mass spectrometry was performed by the Carbohydrate Center, University of Toronto, Toronto.

Circular Dichroism Spectroscopy-- Lyophilized His-sssAFP-2 was dissolved in 0.01 M PB buffer, pH 7.0. CD measurement was carried out at 0 °C using a cuvette of 0.1-cm path length. The final CD spectrum is an average of the mean residue ellipticities as calculated for three concentrations: 0.145, 0.098, and 0.065 mg/ml.

(Eq. 1)

1

Measurement of Antifreeze Activity-- Antifreeze activity was measured as thermal hysteresis following the procedure of Chakrabartty et. al. (15), using a Clifton Nanolitre Osmometer (Clifton Technical Physics, Hartford, NY). Both control (wflAFP-6, winter flounder liver type AFP) and His-sssAFP-2 were dissolved in 0.1 M NH₄HCO₃ and centrifuged and diluted to desired concentrations before use. For each dilution, measurements were made from three wells and the average value taken.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Isolation of the Skin-type AFP cDNA Clone-- To investigate the presence of AFPs in the skin of shorthorn sculpin, a skin cDNA library was constructed and screened under low stringency conditions (0.6 M NaCl, 42 °C) using a winter flounder skin-type AFP clone (8). Several putative positives were identified; however, only two clones, s3-2 (1027 bp) and s17-12 (991 bp), contained a complete ORF of alanine-rich peptides. Sequence comparisons of s17-12 and s3-2 showed that the two clones were identical, with s3-2 containing an additional 36 bp of the 5'-untranslated region.

View larger version (37K): [in this window] [in a new window]   Fig. 1.   Sequence of clone s3-2 that encodes a 92-residue skin-type AFP (sssAFP-2) from shorthorn sculpin. The ORF begins at the ATG start signal at position 66 and ends at the stop codon TAG at position 342, indicated by the star. The deduced amino acid sequence is given above the ORF of the cDNA sequence. The complete cDNA clone is approximately 1.1 kb. Possible 11-residue repeats of the form Pr-X2-Pr-X7, where Pr is any polar residue and X is any nonpolar residue, usually alanine, are enclosed in boxes.

Tissue Distribution and Seasonal Variation of Skin-type AFP mRNA-- Total RNA from skin, liver, brain, and the dorsal fin of three shorthorn sculpin collected on May 30, 1995, March 6, 1996, and September 4, 1997, were investigated by Northern blot analysis (Fig. 2A). Hybridization with a 437-bp fragment of the 3'-UTR of s3-2 produced bands in the skin, brain, and dorsal fin samples of approximately 1.1 kb in size. Notably absent are positive signals from all three liver samples (Fig. 2A, lanes 1-3). Probing of the same blot with the ORF of s3-2 produces similar results, i.e. ~1.1-kb bands in skin, brain, and dorsal fin and absence in liver (data not shown). Strongest signal in all tissues (except liver) was observed from the fish collected in March (Fig. 2A, lanes 4, 7, and 10). Weaker signals were observed for skin and dorsal fin in May, with lowest levels found in the September samples. In the brain, no signal is detected in May, whereas a weak signal is present in September. These findings indicate that the skin-type AFP mRNA levels exhibit significant seasonal variation. A wider range of tissues were further investigated by RT-PCR (Fig. 2B). In Fig. 2B, positive signals were clearly detected from 1 µg of gill filament, skin, dorsal fin, brain, and stomach total RNA. Furthermore, weaker positive signals were seen in the kidney and muscle samples. However, the signal, as expected, was absent in the liver. The identity of the RT-PCR products was further confirmed by Southern analysis (Fig. 2B, lower panel). Primer extension studies were carried out to determine the complete length of the s3-2 clone and to further confirm its tissue-specific expression (Fig. 3). As shown in Fig. 3, there were three major primer extension products at approximately 91-95 nucleotides in length, which indicates that the entire s3-2 clone would be approximately 1090 bp in length, which in turn correlated well with the predicted size of 1.1 kb determined in Northern analysis. The three bands were easily detected in as little as 0.5 µg of skin total RNA (lane 1), but not in 20 µg of liver RNA (lane 4), further reconfirming that sssAFP-2 is not produced in the liver. Furthermore, the three bands were detectable in 5 µg of brain total RNA. Together, the above results indicate that sssAFP-2 mRNA expression appears to be seasonally regulated and widely expressed in many tissues but is clearly not present in the liver.

View larger version (46K): [in this window] [in a new window]   Fig. 2.   Seasonal variation and tissue distribution of sssAFP-2 mRNA as determined by Northern analysis and RT-PCR. Total RNA was isolated from the indicated tissues of fish collected on May 30, 1995, March 6, 1996, and September 4, 1997. Arrows indicate bands of interest. A, Northern blot demonstrating seasonal variation in skin, brain, and dorsal fin, and lack of expression in liver. Total RNA was probed with a fragment of the s3-2 3'-UTR. B, RT-PCR analysis of sssAFP mRNA in various tissues. All tissues indicated are from the fish collected on March 6, 1996. Lower panel displays the results of Southern hybridization of the RT-PCR products using the 3'-UTR probe.

View larger version (58K): [in this window] [in a new window]   Fig. 3.   Primer extension studies. Total RNA from skin (0.5-5 µg, lanes 1-4), liver (20 µg), and brain (1 and 5 µg) were subjected to primer extension reactions using a primer complimentary to the 5'-end of s3-2. Arrow indicates bands of interest.

Properties of the Recombinant Protein His-sssAFP-2-- To further study the structure/function of sssAFP-2, the 276-bp ORF of s3-2 was cloned into the expression vector pET-15b. The recombinant protein (designated as His-sssAFP-2) contained a 20-residue histidine tag with a single thrombin cleavage site at the N terminus of the full 92-residue sequence. The protein was soluble and purified from the cell extract using a Ni²⁺ column and reverse-phase HPLC. The reverse-phase HPLC profile and SDS-PAGE of the products from each purification step are shown in Fig. 4, A and B. The identity of the purified protein was confirmed by amino acid analysis, protein sequencing, and mass spectrometry. Mass spectrometry produced an estimated protein mass of 9723.46 with a standard deviation of 1.75, which is in agreement with the mass predicted from the amino acid sequence. In the CD spectrum (Fig. 6), His-sssAFP-2 displays the typical minimums at 208 and 222 nm that are indicative of an -helical structure, and the percent helix was calculated to be 74% at 0 °C.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Purification of recombinant His-sssAFP-2. Recombinant His-sssAFP-2 was produced in E. coli and isolated in a Ni2+ column prior to final purification by C18 reverse-phase HPLC. A, reverse-phase HPLC profile. B, SDS-PAGE of the purification products at each step of purification. Lane 1, low molecular weight size markers; lane 2, bacterial cell extract after sonication; lane 3, product of Ni2+ column purification; lane 4, purified His-sssAFP-2 after reverse-phase HPLC.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   His-sssAFP-2 is an active antifreeze peptide. A and B, shown is ice crystal morphology in the presence of His-sssAFP-2. A, 0.09 mM His-sssAFP-2; B, 1.12 mM His-sssAFP-2. C and D, thermal hysteretic activity of His-sssAFP-2 () and wflAFP-6 (), measured in molar concentration (C) and in mg/ml (D).

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

In the present study, we have demonstrated that the shorthorn sculpin, similar to the winter flounder, possesses skin-type AFPs along with serum (liver-type) AFPs as an intergral part of its defense strategy against freezing. The cDNA sequence of sssAFP-2 indicates that it lacks the pre- and prosequences, implying that it is intracellular and functionally analogous to the skin-type AFPs found in flounder. We suggest that an updated nomenclature of these proteins to denote the fish species and the nature of the AFP isoform is warranted, i.e. liver-type (l) versus the skin-type (s). Thus, the compliment of AFPs of the shorthorn sculpin includes the serum-type AFPs, ss-3 (renamed sslAFP-3 to denote Shorthorn Sculpin Liver-type AFP-3) and ss-8 (renamed as sslAFP-8), and the new skin-type AFP, sssAFP-2, identified here. Furthermore, evidence is provided that demonstrates sssAFP-2 possesses antifreeze activity comparable with winter flounder HPLC-6 (renamed wflAFP-6 for Winter Flounder Liver-type AFP-6). However, the amino acid composition of sssAFP-2 does not match that of either FPDP I or II, the two antifreeze peptides previously isolated from the skin of European shorthorn sculpin (12). In fact, FPDP I and II seem more closely related to sslAFP-3 and sslAFP-8, raising the possibility that these preparations may have been contaminated by serum-derived proteins. Furthermore, GenBank^TM searches using the s3-2 cDNA sequence or sssAFP-2 primary sequence did not indicate any identity with any known sequences.

Northern analysis, RT-PCR, and primer extension studies indicate that sssAFP-2 is produced in a wide variety of tissues, most notably the skin, dorsal fin, brain, and gill filament, but not in the liver. The relatively abundant sssAFP-2 mRNA level in brain is interesting, and its functional significance remains to be determined. It is also interesting to note that skin-type AFP genes are not expressed in the liver of sculpin, but are expressed in the liver of flounder (8). Furthermore, in contrast to the expression of wfsAFPs, which only show moderate seasonal fluctuation (17), the sculpin skin-type AFP mRNA levels show significant seasonal variation, which may reflect the different habitats of these two species.

Type I AFPs are alanine-rich, partially amphipathic -helical peptides. It has been proposed that type I AFPs present hydrophilic polar residues for interaction with the ice-crystal lattice through hydrogen bonding while presenting a hydrophobic surface to incoming water molecules, thereby preventing further crystal growth (18-20). As well, recent work has suggested that nonpolar interactions, such as van der Waals forces and hydrophobic effects, may have greater relevance in ice-binding than previously assumed (21, 22). The majority of type I AFPs that have been studied in the past possess 11-residue repeats that consist of Thr-X₂-Asn/Asp-X₇ where X can be any residue but is usually alanine (20, 23, 24). The Thr and Asn/Asp residues are located on the same face of the helix in a periodic nature which presumably leads to hydrogen-bonding with ice in a lattice-matching manner. One notable exception is the shorthorn sculpin serum AFP, sslAFP-8, which does not possess the typical 11-residue repeats (11, 25). In sssAFP-2, there are many putative 11-residue repeats (indicated by boxed residues in Fig. 1); however, none of these putative repeats match the established Thr-X₂-Asn/Asp-X₇ motif. The alanine content of sssAFP-2 is approximately 70%, the secondary structure of sssAFP-2 is predicted to be entirely helical (26) and was confirmed by CD spectroscopy (Fig. 6). Furthermore, a helical wheel presentation of sssAFP-2 suggests a partially amphipathic molecule (see Fig. 7). Thus, sssAFP-2 meets the criteria necessary to be classified as a type I AFP, and at 92 residues, it is the largest known naturally occurring type I AFP identified thus far.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   CD spectrum of His-sssAFP-2. The spectrum is the average of four measurements at 0 °C with His-sssAFP-2 concentrations of ~0.1 mg/ml. From the molar ellipticity at 222 nm, the -helical content of His-sssAFP-2 was calculated at approximately 74%.

View larger version (75K): [in this window] [in a new window]   Fig. 7.   Comparison of sslAFP-8 and sssAFP-2 helices. The key residues in ice-binding as identified by Wierzbicki et. al. (25) are identified in sslAFP-8. The sequence of sslAFP-8 is: MDGETPAQKAARLAAAAAALAAKTAADAAAKAAAIAAAAASA (25). The lysine residues of sssAFP-2 that may correspond to those of sslAFP-8 are also indicated.

The predominant 11-residue repeat within sssAFP-2 is Pr-X₂-Pr-X₇ (see Fig. 1), where Pr represents another polar residue and X is predominantly alanine. Three of the repeats highlighted in Fig. 1 begin with lysine. Although threonine is not present, the correct spacing of polar residues within the 11-residue repeat itself is maintained. Proper positioning of the polar residues in a lattice-matching manner is believed to be critical for AFP binding to ice (19, 27). This belief is supported by crystal and NMR solution structures determined for wflAFP-6 (20, 28), and molecular dynamics simulation techniques (22). Furthermore, the possibility of an 11-residue repeat beginning with lysine has been noted before (20).

Recently, a combination of molecular dynamics and modeling of the interaction of sslAFP-8 with the ice lattice has produced another possible mechanism of interaction (25). This approach involves two binding models: accommodation of binding surface residues within ice cages, and the inclusion of key lysine side chains into the ice lattice through their tetrahedral end groups (25). A comparison of the secondary structure of sssAFP-2 with sslAFP-8 (Fig. 7) shows that sssAFP-2 is similar to sslAFP-8 in that many lysine side chains project in a similar manner to one face of the helix. Furthermore, a comparison of the sequences for the two proteins reveals that many lysine residues within sssAFP-2 possess the correct spacing within the protein as defined by Wierzbicki et. al. (25). Nonetheless, more structural work will be required to resolve and define any similarities or differences in the binding mechanisms of the sculpin AFPs.

It appears that the defense against the dangers of freezing in ice-laden, sub-zero sea water of the shorthorn sculpin is identical to that of the winter flounder. Both species secrete AFP into the blood serum that serves to depress the freezing temperature of their extracellular fluids, and both produce skin-type AFPs that lack signal peptides, suggesting that they function as intracellular protectants. Precisely how these skin-type AFPs help to protect fish from cold and freezing is subject to some debate (for review, see Ref. 29). Valerio et. al. (30) have demonstrated that winter flounder skin is an effective barrier to ice propagation and that the effectiveness of this barrier can be increased by the addition of antifreeze proteins to the extracellular space. This suggests that despite the lack of a signal peptide, the skin-type AFPs may use an alternative pathway for secretion of the cells into the intercellular space, thereby acting to block ice propagation.

Recently, Murray et al.² have identified the gill epithelial cells of the winter flounder as a major site of skin-type AFP production utilizing in situ hybridization and immuno-cytochemical techniques. Because of their importance in gas exchange, gill epithelia are the thinnest of all external epithelial tissues in the fish, and the most likely site to come into intimate contact with potentially lethal ice crystals. Thus, it is possible that the skin-type AFP simply serves to lower the freezing temperature of the intracellular fluids and thus ensure that this essential layer of cells cannot freeze.