The three species of sucrose were commercially obtained: D-sucrose (342 g/mol) from Sigma-Aldrich (S0389-500G), Ficoll 70 (polysucrose 70; 70 kDa) and Ficoll 400 (polysucrose 400; 400 kDa) from Sigma-Aldrich (SC-257529A and 26873-85-8 respectively). Glycerol was sourced from Biobasic (GB0232).

To prepare mixtures a 50 mL stock solution of 0.25 mM polysucrose 400 was made. Polysucrose 70 and monomeric sucrose solutions were made in reference to the polysucrose 400 solution such that each solution contained the same number of sucrose monomers.

From these stock solutions, a series of solutions were prepared via the addition of glycerol. Sucrose content was varied from 100.0% to 87.5% by adding glycerol in weight by weight ratio (sugar 100.0%, sugar 97.5%+glycerol 2.5%, sugar 95.0%+glycerol 5.0%, sugar 92.5%+glycerol 7.5% sugar 90.0%+glycerol 10.0%, and sugar 87.5%+glycerol 12.5%) (Supplementary Table S1).

2 mL from each aqueous sample was dispensed into individual plastic weigh boats and desiccated using a speedvac (Savant SpeedVac SC110 with a Thermo OFP400 vacuum pump) for 16 h. Each experiment was performed with triplicate. After the 16 h of desiccation, known weights of dried samples were loaded to pre-massed differential scanning calorimetry (DSC) aluminum hermetic pan (TA 900793.901) and aluminum hermetic lids (901684.901) while dried samples were loaded to pre-tared thermogravimetric analysis (TGA) platinum crucibles (TA 957207.904). Glass transition temperature (T _g ) and glass former fragility (m index) were derived from DSC thermograms (Supplementary Figure S1) and percentage of remaining water content was calculated using TGA (Supplementary Figure S2).

Dried samples (3–4 mg) were placed in pre-weighed DSC pans and loaded in the DSC machine. The glassy state of sugars was obtained experimentally via quench-cooling. The sample was heated up from 30°C to 200°C at a rate of 10°C/min and then cooled to 0°C immediately. Subsequently, the sample was heated again by raising the temperature to 250°C at a rate of 10°C/min. T _g for each of the three sugar species was obtained as the midpoint of the onset and offset (endset) temperatures using Trios software. Average midpoint values were taken from the triplicates from each type of sucrose-glycerol mixtures.

Samples were run on a TGA (TGA5500 instrument) in 100 μL platinum crucibles (TA 952018.906). Crucibles were tared prior to each run and prior to loading the samples. Then the analysis was performed to determine a material’s thermal stability and its fraction of volatile components by monitoring the weight changes along with the temperature. The dried samples (10–11 mg) were placed in TGA plates and loaded in the TGA machine. The program was started at 30°C and then heated to 250°C at 10°C/min rate. Water loss from each sample of three species of sugars was calculated in order to determine the remaining percentage of water content in the dried samples. Thermograms were used to calculate mass differences that occur after ∼100°C but before the thermal denaturation at ∼180°C. The Trios software “Smart Analysis” tool was used to identify the inflection point between mass loss events occurring from starting masses of samples and the mass of samples at the plateau. The average values were taken from the triplicates from each species of sucrose-glycerol mixtures.

LDH assays were performed according to the methodology described in Boothby et al., 2017. The stock solutions of 25 mM Tris HCl (pH 7.0), 100 mM sodium phosphate (pH 6.0), 2 mM pyruvate, and 10 mM NADH were prepared in bulk and stored at 4°C prior to the experiment. L-Lactate Dehydrogenase (LDH) was received from Sigma (SKU #10127230001). Prior to assay, LDH was diluted to a working concentration of 1 g/L using a pre-prepared stock solution of 25 mM Tris HCl (pH 7.0). Experimental sugar-glycerol mixtures were formulated with LDH at a 1:10 (LDH:sugar-glycerol) ratio. For all samples, 90 μL from each sample was added with 10 μL of LDH into a 1.5 mL microcentrifuge tube and a total volume of 100 μL was split in half as a control and an experimental set. The experimental set was kept in the vacuum desiccator and the control set was kept in the refrigerator (4°C) for 16 h. Experiment was performed with triplicates from each sample for each of the three different species of sucrose. After vacuum desiccation, control and experimental excipient mixtures were rehydrated using molecular grade water and brought to 250 μL total volume. A nanodrop instrument using quartz cuvettes was first blanked against the 980 μL of 100 mM sodium phosphate and 2 mM pyruvate solution. The nanodrop instrument was used to take absorbance readings every 2 seconds at the 340 nm wavelength (A340) for 60 s. Absorbance reading was taken by 10 μL of rehydrated control mixture combined with 10 μL of NADH in 980 μL of 100 mM sodium phosphate and 2 mM pyruvate solution. Then the absorbance reading was taken for the experimental mixture using the same procedure as the controls. A ratio of experimental over control, multiplied by one hundred, will produce the percent protection of the experimental mixture. Likewise percent protection was calculated for each mixture of three species of sucrose. The average protection percentages were taken from the triplicates from each type of sucrose-glycerol mixtures.

Trios software (TA Instruments TRIOS version #5.0.0.44608) was used to perform analysis of onset and endset points in order to calculate the glass former fragility (m index). Calculations of the glass m index were performed based on equations 10 and 14) proposed by Crowley and Zografi (Crowley and Zografi, 2001). On a thermogram with a completed heating ramp to 250°C, the degradation peak, melt peak, and glass transition were identified. The Trios software built-in onset and endset analysis was used to determine the glass transition onset and endset and m index was calculated using Crowley and Zografi’s equations 10 where m is the alternative fragility parameter, ΔE _Tg is the activation enthalpy of structural relaxation at T _g , R is the gas constant, T _g is the experimental glass transition temperature onset, and equation 14 where ΔE _𝜂 is the activation enthalpy for viscosity, R is the gas constant, T _g is the experimental glass transition onset temperature, T _g ^off is the experimental glass transition endset temperature, and constant is an empirical constant of 5 (Crowley and Zografi, 2001). A mean of each set of replicates was obtained.

One-Way ANOVA was utilized to compare the means in independent groups of sucrose- and polysucrose-glycerol mixtures in order to determine whether there is statistical difference between one or more mixtures. The Tukey post hoc test was used to assess the significance of differences between pairs of individual group means. Rstudio (R 4.2.2) and the package ‘ggpubr’ was used to calculate the One-Way ANOVA, Tukey post hoc test, p-value, and R ² results. A p-value of less than 0.05 being one level of statistical significance (*), less than 0.01 being two levels of statistical significance (**), and less than 0.001 being three levels of statistical significance (***). All error values in bar graphs represent one standard error (SE) and 95% confidence interval (CI) levels in correlation graphs. Regression analysis was performed using the simple linear model for monomeric sucrose and polysucrose 400 correlations. Since in each case examining polysucrose 70 produced a bi-phasic correlation, and for such relationships fitting data is less exacting and accurate, we opted not to fit these data.

To begin to address the question of how protectant size influences glassy properties and protection, we generated a series of mixtures, each utilizing one of three sucrose species; monomeric sucrose, polysucrose 70, and polysucrose 400, and increasing amounts of glycerol (0.0, 2.5, 5.0, 7.5, 10.0, and 12.5% glycerol in weight by weight ratio) (Supplementary Table S1). Samples were prepared such that the total number of sucrose monomers in each sample was 5 × 10⁻³ mol in 50 mL. We next assessed the ability of each of these 18 mixtures to preserve the function of the labile enzyme lactate dehydrogenase (LDH) during desiccation (Figures 1A, B). Previous reports have shown that LDH is sensitive to desiccation and that drying and rehydration of this enzyme results in loss in functionality. This functionality can be preserved by coincubation with protective vitrifying agents, including sucrose, prior to drying (Goyal et al., 2005).

Pure monomeric sucrose provided 54.243% protection to LDH (Figure 1C). A decrease in protection in monomeric sucrose mixtures was observed with the addition of 7.5% glycerol, though this was not significant, and with each step-wise addition beyond that, significant decreases in protection were observed relative to pure monomer sucrose (Figure 1C). Pure polysucrose 70, which provides the same number of sucrose monomers (5 × 10⁻³ mol/50 mL) as our pure sucrose monomer sample, provided 79.883% protection to LDH during drying. For mixtures containing polysucrose 70, protection did not decrease with additions of glycerol up to 7.5%, after which there was a dramatic drop-off in protection (Figure 1D). It was noted that the polysucrose 70 biphasic result was different from the sucrose and polysucrose 400 linear results. As a result, we have repeated these assays multiple times (3 technical replicates and 3 biological replicates) on different days. In all cases we obtained the similar results. Thus, given our experience with the assay and our repeated and replicable trials with polysucrose 70 we concluded that while the result is surprising but none-the-less real. For pure polysucrose 400, which provides the same number of sucrose molecules as our pure monomer sample and our polysucrose sample, provided 59.037% protection to LDH subjected to desiccation. Mixtures containing polysucrose 400 differed further from mixtures containing monomeric sucrose or polysucrose 70, in that protection was not affected by the addition of glycerol at any levels used in this study (Figure 1E). Apart from that, there was an increase of protection with addition of 10.0% and 12.5% glycerol to the mixture, but this was not statistically significant relative to pure polysucrose 400 (Figure 1E).

These data demonstrate that the polymer size of sucrose affects its protective capacity, with pure polysucrose 70 providing more protection to LDH than mixtures of pure monomeric sucrose or polysucrose 400 that are composed of the same number of sucrose monomers. Furthermore, these results indicate additions of glycerol affect the protective capacity of different sized sucrose polymers in distinct ways.

After observing that the protective capacity of sucrose in different polymerized states differs during drying, we were curious if there are any glassy properties of these sugars that correlate with their ability to provide protection from desiccation. To begin to assess the correlation between properties of vitrified systems composed of different sized sucrose species and the protection, we first assessed whether the water content could account for these differences or not.

In the field of desiccation tolerance, it has been suggested that desiccation tolerant organisms may be able to retain more water than their desiccation sensitive organisms (Holzinger and Karsten, 2013; Hibshman et al., 2020). A higher amount of water could help increase the proportion of proteome, membranes, and other vital cellular components that remain solvated within the system. Conversely, retaining too much water in a dry/drying state might also be deleterious, by allowing for reduced viscosity and increased alpha- and/or beta-relaxations (Fan and Roos, 2016). Thus, desiccation tolerant organisms likely need to precisely tune the amount of water retained within their cells to maintain optimal viability. Therefore, water content could have an important impact on in vitro stabilization of biological material in a dry state.

To assess whether there are differences in water retention in vitrified systems composed of different sucrose species and whether or not these differences correlate with differences in protection, we tested each of our mixtures using thermogravimetric analysis (TGA) (Figure 1A). TGA produced thermograms with two distinct areas of sample mass loss, the first centered around 100°C and the second around 200°C (Supplementary Figure S2). These mass losses correspond to the evaporation of water from our samples and the degradation of the sample at high temperature, respectively (Hurtta et al., 2004). The water contents of our 18 dry mixtures ranged from 3.141% to 12.172% (Figures 2A–C). Pure dry monomeric sucrose contained 5.636% retained water, while pure dry polysucrose 70 and 400 showed significantly higher water content (12.172% and 7.585%, respectively) (Figures 2A–D; Supplementary Table S2). The water content of the dried sucrose monomer mixtures did not change significantly with the addition of glycerol into the mixtures (Figure 2A). For polysucrose 70, water content significantly decreases upon addition of 12.5% glycerol (Figure 2B). Polysucrose 400 displayed a non-monotonic response to glycerol addition, with an initial decrease, followed by an increase, a decrease, and then a further increase (Figure 2C).

After observing distinct water retentive behaviors in glasses composed of different sucrose species, we assessed the relationship between water content and LDH enzyme protection. For monomeric sucrose mixtures we observed a negative, albeit statistically insignificant, relationship between protection and water content (Figure 2E). This differed from our series of mixtures composed with polysucrose 70, which showed a statistically significant bi-phasic relationship between water retention and protection, with samples at low average water contents (7.940%) being the least protective and samples with the most retained water (10.979%) conferring the most protection (Figure 2F). Finally, mixtures containing polysucrose 400 differed further, displaying a slightly positive, but statistically insignificant, trend between retained water and protection (Figure 2E).

Taken together these data demonstrate that water is retained in vitrified sucrose samples differently depending on the polymer state of sucrose. However, these data indicate that retained water is not a strong predictor of protection for vitrified sucrose at any size of the species tested.

After observing that water retention is a poor indicator of the protective capacity of sucrose species-glycerol glasses, we were curious if glass transition temperature (T _g ) might be a good predictor. T _g is the temperature at which a glassy material will begin to transition into a more rubbery phase. Previous work has established that even small amounts of additives, such as glycerol, can lead to changes in the T _g of a vitrifying material, which can correlate with changes to protective capacity (Weng and Elliott, 2015; Ramirez et al., 2024). With this in mind, we were curious if the T _g differs between sucrose-glycerol mixtures composed of different sized sucrose species. Furthermore, we were curious if T _g , which has previously been observed to be a good predictor of protection conferred by sucrose, is also a good predictor of protection conferred by sucrose species (Ramirez et al., 2024). To this end, we performed differential scanning calorimetry (DSC) on dried sucrose species-glycerol mixtures to obtain T _g onset, midpoint, and endset values (Supplementary Figure S1).

The T _g of pure sucrose was observed to be significantly lower than that of pure polysucrose 70 and polysucrose 400. For example, the midpoint T _g for monomeric sucrose was measured to be 73.310°C, while the T _g of polysucrose 70 and 400 were 126.603°C and 130.205°C, respectively (Figures 3A–C; Supplementary Table S2). The addition of glycerol to all three sucrose species tested had a similar plasticizing effect, however the range of observed plasticization varied (Figures 3A–C). For example, addition of glycerol decreased the T _g of monomeric sucrose from 73.310°C (pure) to 40.907°C (87.5%), a range of 32.403°C (Figure 3A; Supplementary Table S2). This same range for polysucrose 70 and 400 was 64.183°C and 69.282°C, respectively (Figures 3B, C; Supplementary Table S2).

Next, we evaluated the relationship between protective capacity and the T _g onset, midpoint, and endset for each mixture (Figures 4A–C). Regardless of whether onset, midpoint, or endset was considered, the same trends were observed (Figures 4A–C). For monomeric sucrose there was a statistically significant, strong positive correlation between protection and T _g . This is in stark contrast to polysucrose 400, which showed a statistically insignificant negative correlation between protection and T _g . For polysucrose 70, we observed a statistically significant bi-phasic relationship between protection and T _g , where mixtures with a T _g average onset lower than 60.010°C, T _g average midpoint lower than 69.771°C, and Tg average endset lower than 79.532°C provided essentially no protection, but for mixtures with a T _g average onset higher than 97.052°C, T _g average midpoint higher than 104.942°C and T _g average endset higher than 112.834°C we observed the highest levels of protection (Figures 4D–F).

Taken together, these results indicate that the size of the sucrose species influences the T _g of mixtures, with monomeric sucrose transitioning at much lower temperature than either polysucrose 70 or 400. Furthermore, while all size species of sucrose were observed to plasticize with the addition of glycerol, monomeric sucrose plasticized over a much smaller range than either polymer. Finally, and most strikingly, there were correlations between T _g and protection. However, protection provided by sucrose in different polymerized states correlated with T _g in dramatically distinct ways (e.g., positive versus negative correlations).

Since changes to T _g can influence the glass former fragility (m index) of a material, we were interested in whether glass former fragility also differed between our samples as well as if there is a correlation between glass former fragility and protection during drying for mixtures composed of different sized sucrose species. For pure samples lacking glycerol we observed an increase in glass former fragility as the size of sucrose species increased (Figures 5A–C). For each of our three species of sucrose, we observed a decrease in glass former fragility with the addition of 2.5% glycerol, which was statistically insignificant for monomeric sucrose, but statistically significant for both polymers (Figures 5A–C). With the rest of the additions of glycerol (up to 12.5%) there were mild decreases in glass former fragility observed in polysucrose 400, but these were slight and almost always statistically insignificant. Monomeric sucrose and polysucrose 70 however, showed non-monotonic behavior regardless of size of the sucrose species (Figures 5A–C).

Next, we assessed the relationship between the glass former fragility of our mixtures and protection of LDH. This correlative analysis produced only statistically insignificant trends (Figure 5D), indicating that glass former fragility is not a strong predictor of protection. Polysucrose 70, which still showed a statistically significant bi-phasic relationship between glass former fragility and protection, with samples with low average m index (42.142) being the least protective and samples with higher average m index value (56.126) conferring the most protection (Figure 5E).

Taken together these results suggest that glass former fragility is not a mechanistic driver of desiccation protection conferred by sucrose, since the glass former fragility of pure monomeric sucrose is significantly lower than that of pure polysucrose 70 and 400, but monomeric sucrose does not offer more protection than polysucrose 70 or 400 (Figures 1C–E). Furthermore, the addition of glycerol decreased glass former fragility in all cases, but again, there was no strong trend between this decrease in fragility and augmented protection.