Antifreeze Glycoproteins from Antarctic Fish INACTIVATION BY BORATE*

A titration curve of pH M borate in the in

Actiuities-Preparation of partially purified Osage orange lectin as well as the hemagglutination assay was carried out as previously described with sheep red cells (7,8). Lectin solution (7 mg/ml in 0.9% NaCl) was used to prepare a series of dilutions using 0.9% NaCl solution.
One drop of the lectin solution was mixed with 1 drop of 0.5 M borate or 0.5 M phosphate buffer in the pH range of 7.0 to 9.0. Borate solutions in the concentration range of 0.025 to 0.4 M were also used. The hemagglutination assay was then continued as before (7,8 "The extent of oxidation is 70% for galactose and 12% for N-acetylgalactosamine. The antifreeze activity of the antifreeze glycoprotein polyaldehyde was 82.l%,of the native. b The freezing temperature of the antifreeze glycoprotein polyaldehyde (5 mg/ml) in water was P0.400", which was used as 100% antifreeze activity. Glycoprotein-There are many sites of interaction between carbohydrate and borate in the hemagglutination test and many more in the hemagglutination inhibition tests. The latter include the glycoproteins on the surface of the red blood cells, the lectin, and the antifreeze glycoprotein (the inhibitor).
The hemagglutination activity of Osage orange lectin decreased as the pH of the borate buffer increased (Fig. 2). There was no effect of pH on lectin activity in the same pH range when phosphate buffer was used instead of the borate. Varying the borate concentration as well as the pH resulted in loss of lectin activity as the borate concentration increased. Antilectin activity of antifreeze glycoprotein decreased as the pH of the borate buffer increased. In the same range of pH, but in the presence of phosphate buffer, there was no effect of pH on the antilectin activity of antifreeze glycoprotein. Since there was inhibition by both the borate and the antifreeze glycoprotein, only an approximate value can be estimated for the effect of borate on the antifreeze glycoprotein.
Borate appeared to decrease the effects due to the antifreeze glycoprotein to approximately one-sixth of the decrease in the absence of borate. The effect of pH and borate on antilectin activity of the glycoprotein polyaldehyde was approximately the same as on native glycoprotein (Fig. 3). DISCUSSION Although the complexing of borate to hydroxyl groups with the necessary stereochemical relationships on carbohydrates might be considered a part of classical carbohydrate chemistry, the need for much more quantitative information in such critical areas as equilibria and kinetics of interaction has recently been noted by Acree (15). Even less information is apparently available on the interactions of borate in biological systems and the properties of the complexes. Techniques such as the use of magnetic resonance are now beginning to provide some of this information (14).
When tetraborate salts (e.g. Na,B,O,) are added to solutions of aliphatic organic polyhydroxyl compounds which satisfy certain spatial requirements, the pH of the solution decreases and borate ion is bound to the hydroxyl groups (15). In acidic solutions the product is described as a borate anion complexed to two hydroxyl groups (Fig. 4A). In more basic solutions, the amount of complex increases and one borate, still remaining as an anion, can complex with four hydroxyl groups (Fig. 4B) and thus form cross-linkages (14,15).
The antifreeze glycoprotein is inactive when two borate anions are complexed per glycotripeptide.
We did not show, however, where these borate anions were bound. Inspection of ball and stick models (Dreiding) of the carbohydrate side chain constructed in our laboratory indicated that the borate might bind to the 3,4 or 4,6 hydroxyl groups of the terminal galactose residues and to the 4,6 hydroxyl groups of the N-acetylgalactosamine residues. The C-6 hydroxyl groups of the galactose residues do not appear to be essential for activity as shown by the retention of antifreeze activity by the polyaldehyde antifreeze glycoprotein in which oxidation had occurred at approximately 80% of the C-6 hydroxyl groups of the galactose residues and only 20% of the C-6 hydroxyl groups of the N-acetylgalac-