Disaccharide Residues are Required for Native Antifreeze Glycoprotein Activity

Antifreeze glycoproteins (AFGPs) are able to bind to ice, halt its growth, and are the most potent inhibitors of ice recrystallization known. The structural basis for AFGP’s unique properties remains largely elusive. Here we determined the antifreeze activities of AFGP variants that we constructed by chemically modifying the hydroxyl groups of the disaccharide of natural AFGPs. Using nuclear magnetic resonance, two-dimensional infrared spectroscopy, and circular dichroism, the expected modifications were confirmed as well as their effect on AFGPs solution structure. We find that the presence of all the hydroxyls on the disaccharides is a requirement for the native AFGP hysteresis as well as the maximal inhibition of ice recrystallization. The saccharide hydroxyls are apparently as important as the acetyl group on the galactosamine, the α-linkage between the disaccharide and threonine, and the methyl groups on the threonine and alanine. We conclude that the use of hydrogen-bonding through the hydroxyl groups of the disaccharide and hydrophobic interactions through the polypeptide backbone are equally important in promoting the antifreeze activities observed in the native AFGPs. These important criteria should be considered when designing synthetic mimics.


Materials and Methods
Antifreeze glycoproteins (AFGP1-5, Mw = 22.1 kDa) were purified from the Antarctic toothfish Dissostichus mawsoni as described previously and are abbreviated as AFGPs throughout the manuscript 1 .
Then, 5 mg galactose oxidase, 1.4 mg catalase and 1.2 mg peroxidase were added and incubated for another 8 hours. The reaction mixture was then treated with 5% tricholoracetic acid to precipitate the enzymes and the supernatant containing the AFGP-ald was extensively dialyzed in distilled water at 4 °C and then lyophilized. The yield was approximately 45 mg.
Preparation of AFGP-car. 20 mg of AFGP-ald was dissolved in 3 mL milliQ water. Then, 200 mg CaCO3 was added to maintain the pH during oxidation with bromine water. 15 µL bromine water was added and gently swirled to disperse the bromine. The solution color was a homogenous bright yellow. After 2 hours of occasional shaking, most of the color had disappeared and another 10 µL of bromine water was added. After 2 hours the color of the solution was still yellow and 100 µL of 2% sodium thiosulfate was added to destroy the excess bromine. Approximately 2 mL of 2 M HCl was slowly added to dissolve the CaCO3. The oxidized AFGP was dialyzed overnight at 4 °C with two changes of 4 L of deionized water and then lyophilized. The yield of AFGP-car was approximately 20 mg.
Preparation of AFGP-ipp. 20 mg of dry AFGP was dissolved in 4 mL of N,N-dimethyl formamide (Fisher Chemical). After swirling to dissolve, 2 mL of N,N-dimethoxypropane (DMP) (Sigma 136808) was added along with a few crystals (~1 mg) of p-toluene-sulfonic acid as a catalyst. The cloudy solution was stirred over night at room temperature. After 12 hours the solution was clear and another 1 mL of DMP was added and stirred for 6 hours. The preparation was lyophilized until nearly dry and then dialyzed in Spectropore 3 (3.5 kDa mol wt cutoff,) at 4 °C with 3 changes of deionized water over 36 hours. The yield after lyophilization was approximately 20 mg.
Preparation of AFGP-ipp-ald. 10 mg of dry AFGP-ald was treated as in the preparation of AFGP-ipp above, except that it was dissolved in 2 ml of DMF and 1 ml of DMP. The remainder of the preparation was identical as detailed in the AFGP-IPP preparation. The yield was 10 mg.

Preparation of Borate-AFGP (AFGP-bor).
AFGP was dissolved in a 0.3 M sodium borate which was prepared by adjusting the pH of 0.3M boric acid to 9.0 using 4 M NaOH.
NMR measurements. NMR measurements were performed in a mixture of 10% D2O/90% H2O. For the 1 H-NMR and 13 C-NMR experiments (1D and 2D) and diffusion measurements (with water suppression), a 5 mm QXI 1 H/ 13 C/ 15 N/ 19 F probe equipped with a z-gradient on the 700 MHz Bruker AVANCE III system and a 5 mm TXI 1 H/ 13 C/ 15 N probe endowed with a z-gradient on the 850 MHz Bruker AVANCE III were used. The NMR samples of AFGP, AFGP-ald, AFGP-car, and AFGP-ipp were dissolved in 0.5 mL 10% D2O/90% H2O.
The AFGP-bor NMR sample was prepared by dissolving AFGP in 0.5 mL 10% D2O/90% H2O with 0.3 M borate (pH=9.0) and subsequent adjustment of the pH to 9.0 using NaOH. Additionally 1 H-NMR measurements were conducted with water suppression using watergate W5 pulse sequence with gradients and double echo 2  CD measurements. CD spectra were recorded at a 1 nm interval from 260 nm to 180 nm using a Jasco J-1500 spectrometer. CD measurements were performed in a rectangular cell with the optical path of 0.1 cm and at a concentration of 1 mg/mL H2O at 22 o C.
IR/2D-IR measurements. All linear IR absorption measurements were performed using a Bruker Vertex 80v FTIR spectrometer equipped with a liquid-nitrogen-cooled-mercury-cadmium-telluride (MCT) detector.
The spectra were recorded under nitrogen atmosphere at a wavelength resolution of 3 cm -1 . For every spectrum 100 scans were averaged. In all the measurements, a path length of 100 µm was used. The temperature-dependent FTIR measurements were performed using a Peltier-cooled temperature cell (Mid-IR Falcon, Pike technologies). The temperature was ramped from 20 to 5 ˚C at a rate of 1 ˚C/min. In all IR and 2D-IR experiments, the proteins were dissolved in heavy water and at a concentration of 2 wt%. The background measurements for pure D2O were performed using the same ramping parameters and at the same temperature.
We performed 2D-IR experiments by vibrationally exciting the samples with intense femtosecond midinfrared pulses centred at 1650 cm -1 , and probing them with femtosecond pulses centred at 1470 cm -1 . The details of the setup have been described elsewhere 6 . The excitation is performed with a mid-infrared pulse pair. This excitation pulse pair induces transient absorption changes that are monitored by a probe pulse that is delayed by a time Tw. After transmission through the sample, the probe pulse is sent into an infrared spectrograph and detected with an infrared mercury-cadmium-telluride (MCT) detector array, thus yielding the transient absorption spectrum as a function of the probe frequency. The dependence of the transient absorption spectrum on the excitation frequency is determined by measuring transient spectra for many different delay times between the two excitation pulses. By Fourier transformation of these spectra, we obtain the dependence of the transient absorption spectrum on the excitation frequency. By plotting the transient absorption spectrum as a function of the excitation and the probing frequency, we obtain a 2D-IR transient absorption spectrum for each delay time Tw. We measure 2D-IR spectra both for the case that the probe and pump beams have a parallel polarization, and the case where they have a perpendicular polarization. All measurements are performed under N2 atmosphere in a standard sample cell with a path length of 100 μm.
The temperature of the protein is kept constant by using a Peltier element with an active feedback loop. TH measurements. TH activity was determined at AFGP concentrations of 10 mg/mL in water using a Clifton Nanoliter Osmometer as described elsewhere 7 . The hysteresis measurements were performed with a cooling rate of 0.074 o C/min and without annealing. Measurements were preformed multiple times on independent samples 7 . IRI measurements. IRI activity was measured using the splat cooling method 8 instead of sucrose method 9-10 since borate can interact with sucrose which influences the AFGP-borate binding and makes the results unreliable. AFGP and the modified variants were dissolved in PBS buffer (Dulbecco's Phosphate-Buffered Saline, 1×, without calcium and magnesium) with a final protein concentration of 2 µg/mL. We chose 2 µg/mL in order to have a maximal IRI activity for the native AFGP 9, 11 . Figure S1. 1 H-NMR and 2D 1 H, 13 C-HSQC 1 J spectra of AFGP and AFGP-ald in the region from 7.5 to 9.4 ppm. (a) 1 H -NMR spectra shows the new signal at ~9.2 ppm (peak c14) is assigned to proton of the aldehyde group. (b) 2D 1 H, 13 C-HSQC 1 J spectra proves peak c14 is from AFGP-ald. In the 2D HSQC experiment, which is sensitive for 1 J coupling shows clearly a correlation between the aldehyde proton at 9.2 ppm and the carbonyl carbon atom at 190.2 ppm.   The NMR signal at ~5.4 ppm (Figure 2b) is an interplay of all exchangeable hydroxyl protons in the system and depends on all the exchange values between these protons in agreement with prior study [12][13]          Movie S1 (separate file). Growth habit of native AFGP1-5 at 10 mg/mL in the hysteresis gap.
With the native AFGP it is difficult to melt the ice back to a round or oblong single crystal. In this video the single crystal has a protrusion at one end which morphs into the apex of the HBP as it is cooled. In other trials the seed is usually oblong in shape and as the temperature is decreased it grows into a blunt hexagonal bipyramid (BHBP), then morphs into a hexagonal bipyramid with a c-to-a ratio of 1.5-2 which is stable until the burst point. At the freezing point or burst point a fine spicule usually propagates from both ends to the water/oil interface followed by bundles of spicules parallel to the initial spicules. The spicules appear to thicken following their extension to the interface.
Movie S2 (separate file). Growth habit of AFGP-ald at 10 mg/mL in the hysteresis gap.
The growth morphology is the same or at least very similar to the native AFGP. In this video the cooling of the seed crystal was started from the HBP stage.
Movie S3 (separate file). Growth of AFGP-car at 10 mg/mL H2O in the hysteresis gap. As the temperature is slowly lowered, the round seed crystal (10 µm diam) rapidly grows into a BHPB and then a sharp tipped HBP. It continues to grow in "fits and starts" until both tips reach the water/oil interface at which point aaxis growth becomes predominant.
Movie S4 (separate file). Growth habit of AFGP-ipp at 10 mg/mL in the hysteresis gap.
The round seed rapidly grows through the BHPB stage forming a sharp tipped HBP. It continues to grow in "fits and starts" until both tips reach the water/oil interface and a-axes growth becomes predominant. As growth continues without lowering the temperature the crystal grows laterally eventually forming one large crystal that replaces the liquid phase. A similar growth pattern is seen with the native AFGP in 0.3M sodium borate buffer at pH 9.