Ice-binding proteins that accumulate on different ice crystal planes produce distinct thermal hysteresis dynamics

Abstract

Ice-binding proteins that aid the survival of freeze-avoiding, cold-adapted organisms by inhibiting the growth of endogenous ice crystals are called antifreeze proteins (AFPs). The binding of AFPs to ice causes a separation between the melting point and the freezing point of the ice crystal (thermal hysteresis, TH). TH produced by hyperactive AFPs is an order of magnitude higher than that produced by a typical fish AFP. The basis for this difference in activity remains unclear. Here, we have compared the time dependence of TH activity for both hyperactive and moderately active AFPs using a custom-made nanolitre osmometer and a novel microfluidics system. We found that the TH activities of hyperactive AFPs were time-dependent, and that the TH activity of a moderate AFP was almost insensitive to time. Fluorescence microscopy measurement revealed that despite their higher TH activity, hyperactive AFPs from two insects (moth and beetle) took far longer to accumulate on the ice surface than did a moderately active fish AFP. An ice-binding protein from a bacterium that functions as an ice adhesin rather than as an antifreeze had intermediate TH properties. Nevertheless, the accumulation of this ice adhesion protein and the two hyperactive AFPs on the basal plane of ice is distinct and extensive, but not detectable for moderately active AFPs. Basal ice plane binding is the distinguishing feature of antifreeze hyperactivity, which is not strictly needed in fish that require only approximately 1°C of TH. Here, we found a correlation between the accumulation kinetics of the hyperactive AFP at the basal plane and the time sensitivity of the measured TH.

Keywords: antifreeze proteins; ice-binding; thermal hysteresis.

Figure 1.

Stages of the measurement are indicated by text. Sample temperature in the nanolitre osmometer is represented by the bold line. TH is indicated by the double-headed vertical arrow. (Online version in colour.)

Figure 2.

Microfluidic chip design. The control layer (yellow) was used as pneumatic valves and was fabricated on top of the flow layer (white), which was designed to fit the viewing hole of the cooled stage (5 mm in diameter). The crystals were formed in the small side compartments in the flow layer. During AFP accumulation, these compartments were isolated from any flow using the pneumatic valves.

Figure 3.

Effect on TH of the exposure time between ice and TmAFP. Solutions of different concentrations were tested for TH as a function of TmAFP exposure time: 1 µM (black squares), 4 µM (red circles), 10 µM (blue triangles) and 40 µM (green inverted triangles). (a) Plots of TH versus exposure time on a log scale for the different TmAFP concentrations. TH values were the average of 3–10 measurements for each data point, with the variability indicated by the vertical error bars. (b) The untagged TmAFP were tested for time dependence at a concentration of 10 µM. TH values were the average of 3–5 measurements for each data point, with the variability indicated by the vertical error bars. (c) Re-plotted data from (a) where the closest data points to the intersection of a 0.15°C min⁻¹ constant cooling rate (dashed line) are highlighted in yellow. (d) The highlighted data points from (b) are plotted to examine the dependence of TH on the concentration of TmAFP. The curved lines in (a,d) were fitted using Origin software.

Figure 4.

Sensitivity of TH to the AFPIII exposure time. The TH activity of a 40 µM AFPIII solution was examined as a function of ice crystal exposure time. TH values are the average of at least three measurements taken at 0.02°C below T_m (open circles) and 0.1°C below T_m (filled squares), with the variability indicated by the vertical error bars.

Figure 5.

Sensitivity of TH activity to MpAFP exposure time. The TH activity of a 2.4 µM MpAFP solution was plotted against the log of the exposure time. TH values (black squares) are the average of at least three measurements with the variability indicated by the vertical error bars. (Adapted with permission from [22].)

Figure 6.

Sensitivity of the TH activity to sbwAFP exposure time. TH activity of 4 µM (open circles) and 8 µM (filled squares) sbwAFP solutions plotted against the log of the exposure time. TH values are the average of at least three measurements with the variability indicated by the vertical error bars.

Figure 7.

Accumulation of TmAFP and AFPIII on different ice crystal planes. (a,b) TmAFP and (c,d) AFPIII. (a,c) Ice crystals with surface-adsorbed AFP-GFP. Crystal axes are indicated by white arrows, and their dimensions are indicated by the scale bar. In the upper image (a), the two concentric lemon-shaped crystals are the seed for the observed basal plane. See the electronic supplementary material, movie S5, for the development of the basal planes from these seeds. The seed crystals are slightly tilted thus the projection of the basal planes that emerge from them is seen in the optical observation direction that is perpendicular to the microfluidic channel. Note that the apparent lens shape of the basal plane (marked with dotted lines) is caused by the round shape of the microfluidic channels (see the electronic supplementary material, figure S3). The scale bar is the same for both images. (b,d) Measurement of the fluorescence intensity of surface-adsorbed AFP-GFPs over time on the basal planes (black squares, TmAFP) and prism plane (black squares, AFPIII). The red lines were plotted using the equations presented in the results section for each AFP.