A phase transition modulates the protective function of a tardigrade disordered protein during desiccation

Abstract

Water is essential for active life, yet some organisms, such as tardigrades, can survive prolonged periods of drying‐induced dormancy. Cytoplasmic abundant heat‐soluble (CAHS) proteins are disordered proteins that undergo a phase transition from the solution to gel state. CAHS proteins help tardigrades survive extreme drying, increase hyperosmotic stress tolerance in heterologous systems, and preserve the function of labile enzymes during drying in vitro. It has been speculated that the ability of CAHS proteins to form gels might be mechanistically linked to their protective capacity. However, recent evidence suggests that while gelation enhances hyperosmotic stress tolerance, it is not required for this phenomenon. Still, the extent to which gelation is necessary for other CAHS‐based protective functions, such as enzyme protection during drying, is unknown. Here, we show that rather than the solution or gel state of CAHS proteins being the sole protective phase, each phase is optimized to protect different enzymes during drying. Using in vitro assays that provide clear functional readouts and allow for precise control over CAHS and client enzyme ratios, we show that the gelled state of CAHS D, a model CAHS protein, promotes the protection of the enzyme lactate dehydrogenase during drying. We find that the opposite is true for the enzyme citrate synthase, with variants of CAHS D that do not gel providing optimal protection to this enzyme. Correlative analysis between protective capacity and sequence/ensemble features of CAHS D variants supports the notion that phase is a major driver of differential enzyme protection. Finally, we show that enhanced water binding is an emergent property of gelation that positively correlates with the protein's ability to protect LDH. These results demonstrate a link between the phase of CAHS proteins and their protective function, providing insights into how CAHS proteins help tardigrades counteract the spectrum of stresses encountered during different stages of drying. Broadly, this study advances our understanding of desiccation tolerance, while providing insights into engineering strategies to tune protein‐based excipients to protect specific clients. This study contributes to a broader discussion in the protein field about the functionality of phase behavior and states.

1. INTRODUCTION

Water is required for all metabolism and plays an important role in maintaining the structure and function of biological macromolecules (Ben‐Naim, 2009; Kauzmann, ¹⁹⁵⁹; Privalov & Crane‐Robinson, ²⁰¹⁷). Despite this, some organisms, spread across the kingdoms of life, can survive near complete desiccation for prolonged periods by entering into an ametabolic state known as anhydrobiosis (Greek for “life without water”) (Boothby, 2019; Crowe et al., ¹⁹⁹²; Kalemba et al., ²⁰²³; Rebecchi et al., ²⁰⁰⁷). How these organisms preserve their cells and cellular components in such a state for years, decades, or in some cases, millennia, is one of the enduring mysteries of organismal physiology (Boothby, 2019; Crowe et al., ¹⁹⁹²; Kalemba et al., ²⁰²³; Rebecchi et al., ²⁰⁰⁷; Waterworth et al., ²⁰²⁴).

Desiccation tolerance in animals is relatively rare, being found sporadically within only four animal phyla (Benoit et al., 2023; Ricci, ¹⁹⁹⁸; Ricci & Caprioli, ²⁰⁰⁵; Thorat & Nath, ²⁰¹⁸; Wharton, ²⁰⁰³; Wright, ²⁰⁰¹; Wright et al., ¹⁹⁹²). One phylum, the phylum Tardigrada, has recently been the subject of intense study with regard to anhydrobiosis, as many tardigrades—colloquially known as water bears—robustly survive many environmental extremes, including desiccation (Wright, 2001; Wright et al., ¹⁹⁹²). When faced with drying conditions, these microscopic animals retract their eight legs and head inside their cuticle, forming a ball‐like structure known as a “tun,” in which they become ametabolic and enter anhydrobiosis until rehydration—sometimes decades later—revives them (Crowe, 1972; Møbjerg & Halberg, ²⁰¹¹; Richaud & le Goff, ²⁰²⁰).

On a molecular level, relatively little is still known about how tardigrades perform this feat (Schill, 2018). One recent insight comes from the discovery of cytoplasmic abundant heat soluble (CAHS) proteins (Yamaguchi & Tanaka, 2012). These proteins, whose sequences are conserved in many tardigrade species but are not found in other taxa, are essential for robust tolerance to desiccation in tardigrades (Boothby & Tapia, 2017) and confer hyperosmotic (Sanchez‐Martinez & Nguyen, 2024; Tanaka & Nakano, ²⁰²²; Yamaguchi & Tanaka, ²⁰¹²) or increased desiccation tolerance (Boothby & Tapia, 2017) when heterologously expressed in various cells and organisms. Furthermore, purified CAHS proteins have been demonstrated to protect labile enzymes from desiccation in vitro (Biswas & Gollub, 2024; Boothby & Tapia, ²⁰¹⁷; Kc et al., ²⁰²⁴; Nguyen et al., ²⁰²²; Packebush & Sanchez‐Martinez, ²⁰²³; Piszkiewicz & Gunn, ²⁰¹⁹; Sanchez‐Martinez & Nguyen, ²⁰²⁴), though these studies have focused on the preservation of only a few client proteins.

Several groups have proposed mechanisms by which CAHS proteins might provide protection to cells and cellular constituents during drying (Bino et al., 2024; Boothby & Tapia, ²⁰¹⁷; Eicher et al., ²⁰²³; Eicher & Brom, ²⁰²²; Kc et al., ²⁰²⁴; Malki & Teulon, ²⁰²²; Nguyen et al., ²⁰²²; Sanchez‐Martinez et al., ²⁰²³; Sanchez‐Martinez & Nguyen, ²⁰²⁴; Yagi‐Utsumi & Aoki, ²⁰²¹; Yamaguchi & Tanaka, ²⁰¹²). Still, a unified understanding of the mechanism, or more likely mechanisms, that govern the protective function of these proteins is not well understood.

An emergent property of CAHS proteins that has garnered recent attention as a potentially important driver of desiccation protection is their ability to form gels (Eicher et al., 2023; Eicher & Brom, ²⁰²²; Kc et al., ²⁰²⁴; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴; Tanaka & Nakano, ²⁰²²; Wang et al., ²⁰²³), both in vitro and in vivo (Eicher et al., 2023; Eicher & Brom, ²⁰²²; Kc et al., ²⁰²⁴; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴; Tanaka & Nakano, ²⁰²²; Wang et al., ²⁰²³). CAHS gels form in a concentration‐dependent fashion and are reversible by heat or dilution (Eicher et al., 2023; Eicher & Brom, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴). Several groups have proposed mechanisms for the formation of CAHS gels (Eicher et al., 2023; Eicher & Brom, ²⁰²²; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴; Tanaka & Nakano, ²⁰²²). While differences in gelation mechanism may exist between CAHS orthologs, an emerging consensus is that the gelation of CAHS proteins is mediated through an assembly process reminiscent of many intermediate filaments (Sanchez‐Martinez & Nguyen, 2024; Tanaka & Nakano, ²⁰²²).

In their monomeric state, CAHS proteins are predicted to exist in a disordered ensemble that resembles a dumbbell, with two collapsed terminal regions held apart by an extended helical linker (Eicher et al., 2023; Eicher & Brom, ²⁰²²; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴). During self‐assembly, CAHS monomers first undergo an asymmetric dimerization event, mediated by coiled‐coil interactions between linker regions (Eicher et al., 2023; Eicher & Brom, ²⁰²²; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴). These dimers then further associate into supercoiled unit‐length filaments (Sanchez‐Martinez & Nguyen, 2024), which polymerize end‐to‐end into the fibers that make up the gel via intermolecular beta‐sheet interactions between their terminal regions (Eicher et al., 2023; Eicher & Brom, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴).

Putative CAHS proteins were previously estimated to accumulate to ~0.05 ng (0.0019 pmol) in dried tardigrades (Nguyen et al., 2022). Because CAHS gels are concentration‐dependent (Eicher & Brom, 2022; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴) and the concentration of CAHS proteins is dramatically increased during drying due to loss of water as well as, in some species, the upregulation of CAHS encoding genes (Boothby & Tapia, 2017), the ability of these proteins to form gels has been speculated to be important for their protective capacity during desiccation (Boothby & Tapia, 2017; Eicher & Brom, ²⁰²²; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴; Tanaka & Nakano, ²⁰²²; Yagi‐Utsumi & Aoki, ²⁰²¹). However, direct testing of this hypothesized function of gelation has only been performed in a single prior study (Sanchez‐Martinez & Nguyen, 2024). In that study, variants of CAHS D (UniProt: P0CU50) that can and cannot form gels were assessed for their ability to confer hyperosmotic stress tolerance to human cells (Sanchez‐Martinez & Nguyen, 2024). Surprisingly, while gelling variants performed best, gelation was not essential for increased hyperosmotic stress tolerance, as a non‐gelling variant significantly outperformed controls in this regard (Sanchez‐Martinez & Nguyen, 2024). Furthermore, this study demonstrated a significant correlation between survival under hyperosmotic conditions and the ability of CAHS variants to reduce metabolism (Sanchez‐Martinez & Nguyen, 2024). Taken together, these limited prior results suggest that gelation is an emergent property of CAHS proteins that contributes to stress tolerance, likely in part through the slowing of metabolism, but that gelation is not the only mechanism governing CAHS‐mediated protection of cells faced with water deficit (Sanchez‐Martinez & Nguyen, 2024).

One point that remains unclear is whether gelation is necessary for other CAHS‐mediated protective functions, such as the protection of labile enzymes during drying. Here we address this question using in vitro enzyme protection assays and a range of CAHS D variants with different gelling capacities. We opt for in vitro assays to reduce the complexity of functional readouts associated with in vivo/ex vivo systems as well as to provide us with precise control over the concentrations and ratios of protectants and client enzymes.

Consistent with findings on cellular hyperosmotic stress tolerance (Sanchez‐Martinez & Nguyen, 2024), we show that while gelling variants of CAHS D confer optimal protection to the enzyme lactate dehydrogenase (LDH) during drying, non‐gelling variants also protect LDH, albeit to a lesser extent. Surprisingly, we observe the opposite trend for a second enzyme, citrate synthase (CS), where gelling variants protect this enzyme during drying worse than non‐gelling variants. To explain the difference in protective capacity observed for different client enzymes, we perform numerous analyses to identify sequence and ensemble features of CAHS variants that correlate with protection. From these, we find a single property, the propensity to form helices (a requirement for gelation), that correlates significantly with LDH protection. Thus, we conclude that gelation is a major driver of CAHS‐based protection of LDH, but not CS.

Exploring the mechanism underlying CAHS gel‐based protection of LDH, we identify enhanced water binding, the ability to associate tightly with water, as an emergent property of gelation. Furthermore, the ability of CAHS variants to bind water significantly correlates with increased LDH protection but negatively correlates with CS protection. Together, we conclude that the phase transition of CAHS D from a solution to a gel state tunes the protein's protective capacity for different client enzymes during drying. We propose a model for how this dynamic biophysical and functional behavior allows tardigrades to tolerate the continuum of protein‐based stresses encountered during different stages of the drying process.

These findings advance our fundamental understanding of anhydrobiosis. From an applied perspective, our study provides insights into how protein‐based mediators of desiccation tolerance can be engineered to tune their function toward optimal protection of different client molecules. Beyond desiccation tolerance, our findings contribute to a broader discussion on the functional significance of protein‐phase behavior and states.

2. RESULTS

2.1. Gelation of CAHS D improves the protection of the labile enzyme LDH during drying

To assess whether the gelation of CAHS proteins is important for the protection of client enzymes, we used CAHS D (UniProt: P0CU50), a model CAHS protein from H. exemplaris (Biswas & Gollub, 2024; Piszkiewicz & Gunn, ²⁰¹⁹; Sanchez‐Martinez & Nguyen, ²⁰²⁴), and eight variants of CAHS D with different propensities to form gels (Figure 1a). Previous characterization of CAHS D identified that this protein exists in a dumbbell‐like ensemble consisting of an extended linker region flanked by two terminal domains (N‐terminus and C‐terminus) (Biswas & Gollub, 2024; Sanchez‐Martinez & Nguyen, ²⁰²⁴) all of which are essential for CAHS D to form gels (Sanchez‐Martinez & Nguyen, 2024). The variants used here all perturb or augment CAHS D's ensemble in some way, resulting in proteins with different gelling propensities (Biswas & Gollub, 2024; Sanchez‐Martinez & Nguyen, ²⁰²⁴) (Figure 1a). For example, the variant termed 2X Linker Region (2X_LR) was generated via the tandem duplication of the endogenous linker region of CAHS D and robustly forms gels at lower concentrations than wild‐type CAHS D (Biswas & Gollub, 2024; Sanchez‐Martinez et al., ²⁰²³; Sanchez‐Martinez & Nguyen, ²⁰²⁴). Conversely, many variants do not form gels. For example, the variant Full Length Proline (FL_Pro) was generated by the insertion of several prolines in the C‐terminus of CAHS D (Sanchez‐Martinez et al., 2023; Sanchez‐Martinez & Nguyen, ²⁰²⁴). The addition of prolines disrupts three regions of beta‐structure, which are required for gelation (Sanchez‐Martinez et al., 2023; Sanchez‐Martinez & Nguyen, ²⁰²⁴). The other six variants were generated through the truncation of wild‐type CAHS D (N‐terminus, Linker Region, NL1, CL1) or the swapping of the terminal domains (NLN and CLC) (Sanchez‐Martinez & Nguyen, 2024) (Figure 1a).

FIGURE 1.

Engineered gelling and non‐gelling variants of CAHS D modulate its protective capacity for different labile enzymes during desiccation. (a) Schematic representing gelling and non‐gelling variants of CAHS D used in this study. (b) Concentration‐dependent protection of lactate dehydrogenase (LDH) (c) Bar graph represents PD50_LDH, n = 3. (d) Concentration‐dependent protection of citrate synthase (CS). (e) Bar graph represents PD50_CS, n = 3. ANOVA was used for statistical comparison between CAHS D and each variant. ns = not significant, * = p <0.05, ** = p <0.0), *** = p <0.001. Error bars represent the standard deviation. (f) Plot comparing protection of LDH and CS normalized to concentrations of protectant to client enzymes.

To test the protective capacity of CAHS D variants with different gelling propensities, we used the enzyme LDH. LDH is commonly used in stress‐tolerant studies to measure the capacity of molecules to preserve enzyme function during drying (Biswas & Gollub, 2024; Kc et al., ²⁰²⁴; Nguyen et al., ²⁰²²; Piszkiewicz & Gunn, ²⁰¹⁹). LDH is an ideal molecule for such studies, as it is extremely desiccation‐sensitive, losing ~98% of its enzymatic function upon drying and rehydration (Boothby & Tapia, 2017; Goyal et al., ²⁰⁰⁵; Nguyen et al., ²⁰²²; Piszkiewicz & Gunn, ²⁰¹⁹). This function can be preserved during drying through co‐incubation with protective molecules, including CAHS D (Biswas & Gollub, 2024; Kc et al., ²⁰²⁴; Nguyen et al., ²⁰²²; Piszkiewicz & Gunn, ²⁰¹⁹). Testing the ability of each variant in an LDH desiccation assay revealed that all variants protect LDH with varying degrees of efficiency in a concentration‐dependent manner (Figure 1b). It should be noted that the concentrations displayed in the figure represent the initial concentration of the protectant prior to desiccation. As the drying process progresses and water content is diminished, the concentration of these protective proteins will increase from this starting value. By fitting these data with a sigmoidal curve, we derived a protective dose 50 (PD50) value—the concentration (mg/ml) of protectant protein needed to preserve 50% of enzyme activity—for each of our variants (Figure 1c). Hence, a lower PD50_LDH indicates more efficient protection. In comparing the PD50_LDH of each variant, we observe that the gelling variants 2X_LR and CAHS D outcompeted the non‐gelling variants. N‐term and NL1 were the worst at protecting LDH. This trend is also observed when converting PD50_LDH into molar concentrations (Figure S1). These data suggest that the ability of CAHS D and its variants to form gels improves the protection of LDH during drying.

2.2. Gelation of CAHS D reduces its protective capacity for the labile enzyme CS during drying

To assess whether the observed link between gelation of CAHS D and CAHS D's ability to protect LDH during drying extends to other desiccation‐sensitive enzymes, we performed a similar protection assay on CS using CAHS D and the eight variants detailed above (Figure 1d,e). CS is an aggregation‐prone enzyme that is widely used in many stress‐related studies to assess the anti‐aggregation capacity of protective molecules (Ahrman et al., 2007; Goyal et al., ²⁰⁰⁵; Hibshman et al., ²⁰²³; Mishra et al., ²⁰⁰⁵; Mymrikov et al., ²⁰¹⁷; Nicholson et al., ²⁰²⁵). Similarly to LDH, we observed varying degrees of concentration‐dependent protection of CS from each CAHS D variant (Figure 1d). Using these data, we derived a PD50_CS for each CAHS D variant (Figure 1e). When comparing the resulting data, we observed that the gelling variants (CAHS D and 2X_LR) were some of the least efficient at protecting CS (Figure 1e). 2X_LR was the least effective at protecting CS and did so significantly worse than wild‐type CAHS D. Wild‐type CAHS D performed equally as poorly as the variant NLN, which was previously reported to not form gels, but rather highly viscous solutions at high concentrations (>10 mg/mL) (Hesgrove et al., 2021). Wild‐type CAHS D, NLN, and 2X LR all were significantly outperformed by variants that do not form gels or highly viscous solutions. Furthermore, the variants that protected CS most efficiently (NL1, FL_Pro, CL1) performed middlingly (or worse in the case of NL1) in the LDH assay. A direct comparison of each variant's PD50_LDH and PD50_CS reveals a negative trend, where the better a variant is at protecting one enzyme, the worse it is at protecting the other enzyme (Figure 1f).

These results indicate that CS is best protected by non‐gelling variants of CAHS D. Combined with the data from our LDH assay, these results demonstrate that rather than a single phase of CAHS D (gel vs. solution) being the functionally protective phase, changes in its phase modulate the protective capacity of CAHS D for different client enzymes during drying.

2.3. Viscosity and differential interactions are not linked to the protection of CS and LDH

The changes made to generate CAHS D variants might result in different material properties beyond gelation, such as changes to viscosity, which might in turn correlate with differential protection. To test for this possibility, we measured the viscosity (loss modulus) of CAHS D variants at 10 mg/mL using rheometry and correlated these values with LDH and CS PD50 values (Figure S2A–C). We saw no relationship between the viscosity and protective values for our variants.

Additionally, we wondered if CAHS D might provide differential protection to LDH and CS based on differential interactions. To test this, we performed crosslinking using mixtures of CAHS D and LDH, as well as CAHS D and CS. We did not see any emergence of new crosslinked species (Figure S2D). Similarly, we performed gel shift assays using non‐denaturing native polyacrylamide gels. Here we observed no shifts to either LDH or CS when coincubated with CAHS D (Figure S2E, F). Together, these data suggest that despite providing protection to both enzymes during drying, CAHS D does not form strong interactions with either LDH or CS, and thus differential interactions likely do not drive the differential protection observed above.

2.4. Sequence features that commonly dictate disordered protein function do not correlate with CAHS D protection

The results presented above indicate that the phase behavior of CAHS D might play a crucial role in modulating its ability to protect distinct client enzymes during drying. However, it is important to acknowledge that the generation of each variant involved modifications to the sequence parameters of wild‐type CAHS D. This fact highlights the possibility that the sequence of CAHS D variants, and not their phase, might be promoting the disparate degrees of protection they confer to LDH or CS during drying. This possibility motivated us to investigate whether IDP sequence features correlate with protection.

Much like the structure–function paradigm for well‐folded proteins, the conformational ensemble of IDPs can influence their function (Das et al., 2015, 2016; Marsh & Forman‐Kay, ²⁰¹⁰; Riback et al., ²⁰¹⁷). CAHS proteins are largely disordered, with their terminal and linker regions having the propensity to form transient secondary structure. Similarly, just as the primary structure of a well‐folded protein in part dictates its structure, the sequence of IDPs in part dictates their conformational ensembles and thus their function. Sequence features known generally to influence IDP form and function include charge patterning (κ), fraction of charged residue (FCR), net charge per residue (NCPR), and hydropathy (Das et al., 2015; Das & Pappu, ²⁰¹³; Tedeschi et al., ²⁰¹⁷). For each of our CAHS D variants, we investigated the correlation between these sequence features and the ability to provide protection to LDH or CS during drying (Figures 2, S3).

FIGURE 2.

Sequence features that dictate IDP ensemble do not correlate with CAHS D variants protection of LDH or CS. Correlation plots comparing PD50 against LDH and CS denaturation with (a), (b) Charge distribution, (c), (d) fraction of charged residue (FCR), (e), (f) Net charge per residue (NCPR), and (g), (h) hydropathy. Statistics were calculated using Pearson's correlation coefficient. (i) Das–Pappu diagram of states of the CAHS D variants. (j), (k) Correlation plots comparing PD50 against f₊:f₋ ratio.

The kappa (κ) parameter represents how well‐mixed oppositely charged residues in a sequence variant are (Das & Pappu, 2013). The charge patterning of the CAHS D variants showed differences, where variants composed of largely the N‐terminus (N‐Term, NL1, and NLN) showed more segregation of oppositely charged residues (κ: 0.093–0.172) as compared to the other variants (κ: 0.059–0.083) (Figure S3A). Bioinformatic analysis using CIDER (Holehouse et al., 2017) indicates that most of the variants are strong polyampholytes (FCR ≥0.3), or have high proportions of charged residues, with the exception of N‐term (FCR = 0.189) and NL1 (FCR = 0.237) (Figure S3B). Additionally, all variants were identified to have a net‐negative charge (NCPR <0) (Figure S3C) and a hydropathy score between 2 and 4 (Figure S3D). Although the above parameters are known to dictate an IDP's conformational ensemble (Das & Pappu, 2013), and hence presumably its function, we did not find any significant correlation with the ability of the variants to protect LDH or CS from desiccation‐induced inactivation (Figure 2a–h).

It is possible that these sequence features combine to elicit protective ensembles and thus, in addition to assessing the individual correlation of each sequence feature to protection, we predicted the ensemble state for each of our variants (Figure 2i). The CAHS D variants occupied different regions in the Das–Pappu state diagram (Das et al., 2015; Das & Pappu, ²⁰¹³). N‐term and NL1 fall into the weak polyampholyte region, suggesting that these variants exist as collapsed globules (Das et al., 2015). NLN and FL_Pro occupy the Janus region, indicating that their conformations are dynamic and highly dependent on their local environment. CLC, CL1, and Linker each fall under the strong polyampholyte region where they are predicted to predominantly exist as extended, coil‐like conformations. CAHS D exists on the boundary between the Janus and strong polyampholyte regions. Qualitatively, there seems to be an emerging trend where weak polyampholytes were the worst at protecting LDH, albeit this only represents two of our nine variants. There were no such trends for CS protection. Correlating PD50 to the f+:f‐ratio yielded no significant correlation for either client (Figure 2j,k).

Together, this analysis indicates that the differences in basic sequence features typically thought to be important for IDP function are insufficient to explain the differences in protective capacity observed between CAHS D variants. This supports the idea that the capacity of CAHS D to form gels is a property governing its differential protection of labile enzymes during drying.

2.5. Helicity correlates with LDH protection, but not CS protection

While correlative analysis suggests that sequence differences and their combined effect on the state of CAHS D cannot explain the variation in observed enzyme protection, it is possible that small changes in sequence could result in changes in a variant's ensemble and hence function, unrelated to the phase of the variant. To assess this possibility, we considered both the global (expansion/compaction) and local (secondary structure) ensemble features of each variant.

An IDP's global ensemble, and especially its global dimensions, can influence its overall function (Panova et al., 2019; Schrag et al., ²⁰²¹). To assess whether the global dimensions of CAHS D variants influence their protective capacity, we began using high‐throughput small‐angle x‐ray scattering (HT‐SAXS) to measure the radius of gyration (R _g) of each of our nine variants (Figures 3a and S4). In order to keep our samples monodispersed, we tested each protein at a concentration of 4 mg/mL, which is below the gelling concentration of CAHS D (Kc et al., 2024; Sanchez‐Martinez & Nguyen, ²⁰²⁴). The R _g of wildtype CAHS D was found to be ~55 Å, which is in line with previous reports (Malki & Teulon, 2022; Sanchez‐Martinez & Nguyen, ²⁰²⁴). The R _g of CAHS D variants ranged in size from ~30 to 55 Å (Figure 3a). To assess whether there is a relationship between the global ensemble of CAHS D variants and their protective function, we correlated R_g values with PD50 values for LDH and CS (Figure 3b,c). This analysis did not detect a significant relationship between R _g and PD50_LDH (p‐value = 0.541) or PD50_CS (p‐value = 0.242). This result suggests that the global dimensions of CAHS D do not contribute heavily to the protection of either LDH or CS and cannot be used to understand mechanisms underlying their differential protection during drying.

FIGURE 3.

Global ensemble changes in CAHS D do not correspond to protection against protein denaturation. (a) Empirical radius of gyration (R _g) measurements for CAHS D variants were conducted using HT‐SAXS. Correlation plots show the relationship between PD50 of (b) LDH and (c) CS against the R _g of all variants. (d) Empirical helicity measurements for CAHS D variants were conducted using circular dichroism (CD) spectroscopy at 10 μM. Correlation plots show the relationship between PD50 of (e) LDH and (f) CS against the percent helicity. Statistics were calculated using Pearson's correlation coefficient.

Next, we assessed the local ensemble (secondary structure) of CAHS D and its variants (Figures 3d and S5). Previous studies have demonstrated wildtype CAHS D can form transient secondary structures while maintaining overall disorder (Eicher et al., 2023; Malki & Teulon, ²⁰²²; Nguyen et al., ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴). Additionally, a prior study found that the helical content of CAHS D variants is a driver of LDH protection (Biswas & Gollub, 2024). Importantly, helicity is essential for the dimerization of CAHS proteins—an important step is the formation of CAHS gels, and helicity increases upon gelation and drying (Eicher & Brom, 2022; Kc et al., ²⁰²⁴; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴; Yamaguchi & Tanaka, ²⁰¹²). To assess the role of helicity in the protection of different client enzymes during drying, we used previous results from the literature (Biswas & Gollub, 2024) as well as novel circular dichroism (CD) spectroscopy experiments to assess the helical content of previously untested CAHS D variants (Figures 3d and S5). Consistent with previous work (Biswas & Gollub, 2024), we found a significantly negative correlation between helicity and PD50_LDH (R = −0.747, p‐value = 0.021) (Figure 3e), suggesting that increased helicity reduces the amount of protectant required to preserve 50% of LDH's activity. However, based on the strength of this correlation (R ² = 0.558), this data suggests that although helicity is important for predicting trends in LDH protection, it is not the sole determinant. In contrast, when correlating helicity with CS protection, we observed a nonsignificant correlation (R = 0.377, p‐value = 0.317) (Figure 3f), suggesting that helicity may not play a major role in CS protection. It should be noted that due to technical limitations, CD measurements were conducted at concentrations below the gelation threshold of CAHS D. Taken together, these differing correlations highlight the distinct role of secondary structure in client‐specific enzyme protection.

2.6. The dimer: Monomer ratio of CAHS variants negatively correlates with CS, but not LDH protection

While our initial observations suggested that gelation might drive LDH protection, it is possible that the formation of lower order oligomers rather than continuous gelation might account for this differential protection. A recent study demonstrated that the oligomerization of COR15‐A, a LEA protein involved in desiccation and freeze tolerance, modulates its specificity for the protection of distinct clients (Hernández‐Sánchez et al., 2024). To this end, we profiled the oligomeric states of our CAHS D variants at their PD50 concentration using 1 mM BS3 crosslinking. By doing this, we can relate the ratio of a variant's dimer to monomer to the ability of that variant to protect LDH and CS. Crosslinking provided a range of different dimerizations for the CAHS D variants at their PD50 concentrations (Figure 4a,b). Correlating the dimer: monomer ratio to LDH protection resulted in a nonsignificant correlation (R = 0.255, p‐value = 0.560) (Figure 4c). However, correlating the dimer: monomer ratio for CS yielded a negative correlation (R = 0.704, p‐value = 0.034) between protective capacity and dimerization (Figure 4d). These results demonstrated that dimerization of CAHS D variants is not a driver of LDH protection. Furthermore, for the CAHS D variants, oligomerization decreases their protective capacity for CS. This result helps us distinguish the potential protective effects of dimerization from gelation and supports the notion that gelation specifically is important for the protection of LDH.

FIGURE 4.

Degree of dimerization negatively correlates with protection of CS, while it does not correlate with LDH protection. Oligomeric profile of CAHS D variants using 1 mM BS3 crosslinking at (a) PD50_LDH and (b) PD50_CS concentrations. Correlation plots show the relationship between PD50 of (c) LDH and (d) CS against the band intensity ratio of dimers to monomers of each variant. Statistics were calculated using Pearson's correlation coefficient.

2.7. CAHS D gelation increases the ability to bind water

The results above suggest that gelation and properties with known roles in CAHS gelation, such as helicity (Eicher & Brom, 2022; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴; Yamaguchi & Tanaka, ²⁰¹²), drive the differential protection of enzymes during drying. We were curious about what makes gelation protective for LDH but not CS. With this in mind, we set out to identify emergent properties of CAHS D gelation that might account for these differences in protection.

One hypothesis that addresses how a protectant functions during desiccation stress is the water entrapment hypothesis (Arsiccio & Pisano, 2018; Belton & Gil, ¹⁹⁹⁴; Corradini et al., ²⁰¹³). This model proposes that protectants can coordinate layers of water between themselves and the client protein, allowing for the hydrogen bond network (HBN) that typically stabilizes labile proteins to be maintained instead of lost during drying (Arsiccio & Pisano, 2018; Belton & Gil, ¹⁹⁹⁴; Corradini et al., ²⁰¹³). Previous work has demonstrated that while gelation of CAHS D does not result in increased water retention in the dry state, CAHS D gelation can affect how tightly the protein binds residual water (Sanchez‐Martinez & Nguyen, 2024).

To investigate how the bound state of water induced by CAHS D gelation might relate to the protection of LDH and CS during drying, we used time‐domain nuclear magnetic resonance (TD‐NMR). TD‐NMR measures the transverse relaxation time (T₂ relaxation) of the hydrogen nuclei of water in a given environment, giving information on the molecular mobility of water molecules (Kuo et al., 2001). For pure water, T₂ relaxation was measured at approximately 2700 ms (Figure S6). Bound water molecules are more structurally organized and have less freedom of motion (Adhikari et al., 2020; Ahmed et al., ²⁰¹⁴; Laage et al., ²⁰¹⁷; Lerbret et al., ²⁰¹²; Raschke, ²⁰⁰⁶), allowing for faster progression back to equilibrium (Gossuin et al., 2000; Liu et al., ²⁰¹⁸). Thus, a faster T₂ relaxation, below 2700 ms, is indicative of water being more structured.

We quantified the T₂ relaxation of water in the presence of each CAHS D variant at increasing concentrations (Figures 5a, S6, S7). We reasoned that this analysis was appropriate due to the dynamic nature of the desiccation process. As desiccation occurs and water leaves the system, the relative volume decreases, and the concentration of CAHS proteins drastically increases, which causes the wild‐type proteins to gel (Romero‐Perez et al., 2023). By using increasing concentrations of our proteins as a proxy for desiccation, we can observe the relative dynamics of water during this process. We selected a concentration range (1, 2, 5, 10, 15, and 20 mg/mL) that allowed us to observe water dynamics at concentrations below (1–5 mg/mL), at (10 mg/mL), and above (15 and 20 mg/mL) the gel point of CAHS D (Sanchez‐Martinez & Nguyen, 2024).

FIGURE 5.

Water binding at gelling concentrations of CAHS D correlates positively with lactate dehydrogenase protection but negatively with citrate synthase protection. CONTIN analysis of Carr–Purcell–Meiboom–Gill T₂ relaxation measurements for (a) CAHS D, (b) FL_Pro, and (c) 2X_LR. The dashed line represents T₂ relaxation at 10 mg/mL CAHS D. (d)–(f) Correlation plots comparing PD50 of CAHS D variants for LDH to T₂ relaxation at gelling concentrations of CAHS D (10, 15, and 20 mg/mL). (g)–(i) Correlation plots comparing PD50 of CAHS D variants for CS to T₂ relaxation at gelling concentrations of CAHS D. Statistics were calculated using Pearson's correlation coefficient.

For CAHS D, we observe a gradual decrease in T₂ relaxation time at concentrations below the gel point of the protein (1–5 mg/mL) (Figure 5a). This modest decrease in T₂ relaxation time is expected because higher protein concentrations provide a greater surface area for water molecules to interact with (Wierzuchowska & Blicharska, 2014). At 10 mg/mL, the concentration at which CAHS D begins to form a gel (Sanchez‐Martinez & Nguyen, 2024), we observe a large decrease in T₂ relaxation from water (~2700 ms) to ~1000 ms. Increasing gelation at concentrations of 15 and 20 mg/mL resulted in further restriction of water (~500 ms) (Figure 5a). This indicates that water binding increased concurrently with the formation of the hydrogel. This was confirmed by using rheometry and correlating the protein's viscoelastic properties to the T₂ relaxation time (Figure S8A).

The gelling variant 2X_LR exhibits a similar trend in water binding, where at gelling concentrations (10 mg/mL), we see an increase in water binding (~500 ms) (Figure 5b) and in the viscoelastic property (Figure S8B). Additionally, we see a greater increase in water binding measurements at 15–20 mg/mL, consistent with the fact that 2X_LR gels at a lower concentration than CAHS D (Sanchez‐Martinez & Nguyen, 2024) (Figures 5b and S8B). In contrast, when comparing the T₂ relaxation patterns to non‐gelling variants, for example, FL_Pro, the jump in T₂ relaxation at 10 mg/mL is not present (Figures 5c and S7). At increasing concentrations, these non‐gelling variants did not exhibit the large decreases in T₂ observed for CAHS D and 2X_LR (Figure S6). Together, these data suggest that the observed increase in water binding corresponds to gel formation.

2.8. Water binding correlates with LDH protection

Next, we assessed whether water binding correlates with the protection of LDH and CS. We correlated the PD50_LDH and PD50_CS of each variant with its T₂ relaxation at the different concentrations tested (Figure 5d–i, Fig. Sis 9). At the gelling concentrations of CAHS D, we observe a significant positive correlation (p‐value $\le$0.05, R >0) with T₂ relaxation and PD50_LDH, suggesting that the binding of water promotes the protection of LDH during drying stress (Figure 5d–f). When correlating T₂ relaxation with PD50_CS, we observe a significant negative correlation (p‐value ≤0.05, R <0) for CS protection (Figure 5g–i).

Since helicity and water binding both correlated significantly with LDH protection, we investigated whether there was a relationship between these two properties. However, a comparison of helicity and T₂ relaxation among CAHS D variants did not yield a significant correlation (R = −0.252, p‐value = 0.512), suggesting that these two properties are independent of each other. The lack of multicollinearity between helicity and T₂ relaxation allowed us to perform a multiple linear regression analysis to assess how these two independent variables together correlated with PD50_LDH. This analysis resulted in an adjusted‐R ² = 0.872, meaning that helicity and T₂ relaxation together strengthened the correlation with LDH protection. These results suggest that both helicity and water binding are drivers of CAHS D‐mediated LDH protection but work, at least in part, independently to provide this protection.

Taken together, these results suggest that the CAHS D variants' ability to bind water is a determinant in their efficiency at protecting LDH and CS, which is directly linked to the phase of the CAHS D protein.

3. DISCUSSION

CAHS protein gelation has been the focus of several recent studies, as this phase transition to a condensed state has been speculated to play an important role in conferring protection during tardigrade desiccation (Boothby & Tapia, 2017; Eicher & Brom, ²⁰²²; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴; Tanaka & Nakano, ²⁰²²; Yagi‐Utsumi & Aoki, ²⁰²¹). Both within and beyond the desiccation tolerance field, how protein phase behavior impacts biological function is a widely discussed topic (Belott et al., 2020; Hatters, ²⁰²³). While several studies have demonstrated that CAHS D effectively protects labile enzymes from desiccation‐induced inactivation (Biswas & Gollub, 2024; Nguyen et al., ²⁰²²; Piszkiewicz & Gunn, ²⁰¹⁹; Sanchez‐Martinez & Nguyen, ²⁰²⁴), it remains unclear whether CAHS protein phase transitions, specifically gelation, contribute to this function (Sanchez‐Martinez & Nguyen, 2024). Here, we show that CAHS D gelation modulates protective effects on two different desiccation‐sensitive enzymes, LDH and CS. Compared to non‐gelling variants, gelling variants of CAHS D provided greater protection for LDH but reduced protection for CS. Likewise, non‐gelling variants protect CS best, but have diminished protective capacity for LDH. We further demonstrated that helicity, an ensemble feature critical for CAHS D gelation, drives LDH protection but does not influence CS protection. Finally, we show that enhanced water binding is an emergent property of gelation and correlates positively with LDH protection. Collectively, our results suggest that rather than one phase being the protective phase of CAHS proteins, both the condensed (gelled) and non‐condensed phases of CAHS D are protective during drying, but that these phases serve to tune protection for specific clients. We speculate that this allows CAHS D to provide protection to drying systems across a spectrum of different desiccation‐related stresses (see below).

3.1. CAHS D sequence variants modulate the protection of labile enzymes

In this study, we demonstrated that sequence variants of wild‐type CAHS D have altered protective capacities in our in vitro protection assays as measured by PD50. While PD50 provides us with a uniform quantified value for comparison, it should be noted that other features from our concentration‐dependent sigmoidal plots are also informative. For example, the steepness of the linear region of these plots denotes how quickly protection increases once a critical protective threshold is met. It is of note that some variants that have rather middling PD50s in either the LDH or CS protection assays have relatively steep linear regions. This indicates that while these protectants do not provide protection at low concentrations, once they do start protecting, they become even more protective quickly. It will be interesting to follow up on why this might be the case in future studies. One possible reason could be that these variants form protective glasses when dry, and that after some critical threshold is met, the protective properties—such as glass transition temperature and/or glass former fragility—may increase or decrease exponentially.

3.2. Why does gelation of CAHS D increase protection for LDH but decrease protection for CS during drying?

One possible explanation for the differential protection between LDH and CS is the distinct ways these enzymes lose function during desiccation. During drying, CS has been observed to form irreversible aggregates (Goyal et al., 2005). In contrast, LDH does not form aggregates upon desiccation (Popova et al., 2015) but is instead suggested to misfold or dissociate from its tetrameric conformation during drying, leading to a loss of function (Simongini et al., 2023). These differences in the mechanism of proteolytic dysfunction during drying may reflect CAHS D's differential protection of enzymes in a gelled and non‐gelled state.

The CAHS D gel network has previously been shown to conformationally immobilize large protein complexes (Sanchez‐Martinez & Nguyen, 2024), potentially stabilizing them through excluded volume or confinement, which compounds the effects of molecular crowding on client proteins (Eggers & Valentine, 2001; Lucent et al., ²⁰⁰⁷; Ping et al., ²⁰⁰³; Sanchez‐Martinez & Nguyen, ²⁰²⁴). This increased confinement, in turn, could destabilize the unfolded state of proteins and promote their functional folded state (Richards, 1977; Simpson et al., ²⁰²⁰).

In addition to simply increasing the excluded volume effect, our work demonstrates that the gelation of CAHS D increases its water‐binding affinity. Since water contributes greatly to protein folding via the formation of a HBN between water and client proteins, CAHS D‐mediated water binding could further increase the stability of client proteins via the maintenance of HBNs. These gelation‐driven mechanisms (excluded volume and HBN maintenance via water binding) align well with the enhanced protection of LDH by gelling variants, which are known to be destabilized during drying, but why is gelation not optimal for the protection of CS?

While the excluded volume effect imposed by CAHS protein gelation may be beneficial to proteins prone to destabilization, for aggregation‐prone proteins, this effect is likely detrimental (or at least not helpful). Gelation likely leads to an environment that accelerates CS:CS interactions, aggregation, and consequently loss of function (Arakawa et al., 2017; Nanev et al., ²⁰¹⁷). Rather than gelation of CAHS D promoting CS protection, we observed that non‐gelling variants preserve the function of the aggregation‐prone enzyme best. This is in line with the “molecular shielding” hypothesis (Chakrabortee et al., 2007, 2012), positing that aggregation of proteins during drying could be reduced by protectants that act as long‐entropic springs. These “molecular shields”' are speculated to prevent aggregation by essentially taking up space within the cell, acting as a sort of packing material, which slows or altogether prevents the association of aggregation‐prone proteins (Chakrabortee et al., 2007, 2012).

3.3. Helicity drives trends in enzyme protection during desiccation

Previous studies, primarily focused on late embryogenesis abundant (LEA) proteins, another broad class of IDP involved in desiccation, have speculated that helicity is important for the proteins' protective function during drying (Bremer et al., 2017; Cuevas‐Velazquez et al., ²⁰¹⁶; LeBlanc & Hand, ²⁰²¹; Shimizu et al., ²⁰¹⁰; Sowemimo et al., ²⁰¹⁹; Tunnacliffe et al., ²⁰¹⁰; Yamaguchi & Tanaka, ²⁰¹²). Similarities can be drawn between LEA and CAHS proteins, as both are generally protective and acquire helicity during desiccation (Biswas & Gollub, 2024; Cuevas‐Velazquez et al., ²⁰¹⁶; Hincha & Thalhammer, ²⁰¹²; Sanchez‐Martinez & Nguyen, ²⁰²⁴; Yamaguchi & Tanaka, ²⁰¹²). Previous work identified the helicity of CAHS D as a driver of LDH protection during drying (Biswas & Gollub, 2024). Here, we extend these insights, identifying helicity as an important driver for LDH protection during drying, but not for CS. However, similar to the previous study, while the correlation between helicity and LDH protection was significant, the weak correlative power (R ² _LDH = 0.573) suggests that additional factors likely contribute to the positive relationship between helicity and LDH protection.

How might helicity contribute to LDH protection? Helicity is known to be important for CAHS protein gelation, specifically for the dimerization of CAHS, which is a critical step in the gelation process (Eicher & Brom, 2022; Hesgrove & Boothby, ²⁰²⁰; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴). This oligomerization process is facilitated through helix–helix interaction between the linker regions of CAHS proteins (Eicher & Brom, 2022; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴; Tanaka & Nakano, ²⁰²²). Given that gelation depends on helicity and that gelling variants provide the strongest protection to LDH, the observed correlation between helicity and LDH protection further supports the notion that gelation is an important driver of LDH protection.

Our observation that helicity correlates with LDH protection brings up the question of whether this protection is being conferred in the dilute or gel phase, as CD measurements were conducted at concentrations below the gelation threshold of CAHS D. We previously demonstrated that increasing concentration of CAHS D promotes helical adoption (Kc et al., 2024; Sanchez‐Martinez & Nguyen, ²⁰²⁴). Thus, if helicity is linked to LDH protection as we show here and in Biswas and Gollub (2024), then while the dilute phase might provide some protection, the gelled state should increase this protection because of the increase in helical content observed to be concomitant with this phase transition. Perhaps stabilization of transient helices is why the gel phase enhances protection of LDH relative to the dilute phase. This would make sense with respect to our TD‐NMR results as well, where we see that gelling variants slow water more than non‐gelling variants. The helices of CAHS D are predicted to be strongly amphipathic, with a hydrophobic and hydrophilic face (Hesgrove & Boothby, 2020; Yamaguchi & Tanaka, ²⁰¹²). Perhaps stabilized helices in the gel phase with hydrophilic faces can better bind and restrict the motion of water molecules compared to the transient helices seen in monomers of the dilute phase.

Importantly, it should be noted that LEA proteins have also been reported to undergo phase changes, which are proposed to be connected to their protective capacity during desiccation (Belott et al., 2020). Interestingly, in these studies, similar to CAHS proteins, phase behavior was found to be governed by SMP domains—low complexity domains with propensity for helical formation (Belott et al., 2020).

Many studies investigating IDP helicity in desiccation protection focus on its role in preserving membrane organization and stability (Moore et al., 2016; Navarro‐Retamal et al., ²⁰¹⁸; Popova et al., ²⁰¹¹). However, apart from the fact that helicity is important for LDH protection (Biswas & Gollub, 2024), little is known about its mechanistic role in protein protection. Our observation that CAHS proteins do not tightly bind to LDH or CS implies that helicity likely is not promoting protectant‐client interactions. However, the adoption of helicity can also influence the surface chemistry of proteins. Helices in CAHS proteins are predicted to have a strong amphipathic nature, with hydrophobic and hydrophilic faces (Hesgrove & Boothby, 2020; Yamaguchi & Tanaka, ²⁰¹²). The formation of a strong hydrophilic face could in turn influence protectant‐water interactions, in line with observations that CAHS D gelation promotes water binding and slowdown in our TD‐NMR experiments. Protectant‐water interactions have been proposed to influence client protection during desiccation (Hesgrove & Boothby, 2020) which also fits with our observations that CAHS‐dependent water slowdown correlates with protection of LDH.

3.4. Gelation of CAHS D enhances the binding and slowdown of water

Here we report the slowdown on water to be an emergent property of CAHS gelation, which positively correlates with the protection of LDH, but not CS. A long‐proposed mechanism of protection during desiccation is the idea of water retention, where protectants can allow an organism to avoid losing water during drying (Boothby & Pielak, 2017; Vieira et al., ²⁰²²; Wright, ¹⁹⁸⁹). However, studies have demonstrated that desiccated CAHS D gels do not retain any more or less water than non‐gelling proteins or proteins unrelated to desiccation tolerance (Brom & Pielak, 2022; Crilly et al., ²⁰²²; Sanchez‐Martinez et al., ²⁰²³). Rather, previous work suggests that the gelation of CAHS D instead influences the strength of interaction between the CAHS D and water by binding water more tightly (Sanchez‐Martinez et al., 2023).

Here, we expanded on this study by performing the first reported experiments on CAHS proteins utilizing TD‐NMR to monitor the effects of gelation on water T₂ relaxation. In correlating water relaxation with the protection of LDH and CS, we find that the increase in CAHS D‐bound water correlated significantly with LDH protection but negatively with CS protection. The restriction of free water into the bound state and its correlation with LDH protection is parsimonious with the water entrapment hypothesis, wherein a protectant entraps a layer of water between itself and a client, helping to “coordinate”' water to the client, providing an HBN that reinforces stability during drying (Arsiccio & Pisano, 2018; Belton & Gil, ¹⁹⁹⁴; Corradini et al., ²⁰¹³). One possible explanation as to why this is not protective for CS is that by accumulating and concentrating water, CAHS D gels might allow for extra molecular mobility that could lead to protein aggregation.

3.5. CAHS phase transitions as a mechanism for tuning protection across a spectrum of drying‐induced perturbations

Surviving desiccation is not about withstanding a singular stress, but rather enduring a continuum of stresses. In the early stages of drying, when the cell begins to lose water, a major perturbation is the reduction in cellular volume and the resulting increase in macromolecular crowding (Rivas & Minton, 2016; Romero‐Perez et al., ²⁰²³). This increased crowding can accelerate stochastic self‐collisions between aggregation‐prone molecules (Chakrabortee et al., 2007; Chakrabortee et al., ²⁰¹²; Hatanaka et al., ²⁰¹³). We propose that at these early stages of drying, where CAHS proteins have not yet started to oligomerize into a gel, the monomeric form of CAHS D is optimal for preventing aggregation (Figure 6). CAHS proteins likely prevent aggregation during these early stages of desiccation via molecular shielding (Romero‐Perez et al., 2023). This adaptation strategy would help preserve functional proteins during a critical window when the cell must adjust to water loss and associated loss of cellular volume/increased crowding (Romero‐Perez et al., 2023).

FIGURE 6.

Proposed model of how phase behavior optimizes CAHS D protection of diverse proteins across the spectrum of desiccation‐induced perturbations. (Upper panels) Before desiccation stress, proteins are fully hydrated. The initial reduction in cellular volume caused by water loss promotes aggregation‐induced damage mainly due to increased crowding. Further loss of water results in loss of the HBN, leading to protein misfolding. (Lower panels) Under hydrating concentrations, low‐oligomeric/non‐condensed forms of CAHS D are best suited to protect proteins from aggregation‐induced damage caused by the reduction in cellular volume. Further loss of water increases CAHS D concentrations and promotes gelation. This allows for stabilization due to confinement, molecular crowding, and increased hydration facilitated by increased water binding.

As drying progresses and water loss continues, the concentration of CAHS D increases, due in part to the loss of solvent but also through its increased expression (Boothby & Tapia, 2017) and this increase in concentration is known to trigger CAHS gelation (Eicher et al., 2023; Eicher & Brom, ²⁰²²; Malki & Teulon, ²⁰²²; Sanchez‐Martinez & Nguyen, ²⁰²⁴; Yagi‐Utsumi & Aoki, ²⁰²¹). As small oligomers of CAHS D combine to form a reticular network, clients could become ensnared between and within assembled CAHS D fibrils, which themselves are concentrating the small amounts of residual water left in the system (Figure 6). This ensnarement within the CAHS gel provides several different protective effects to client proteins (Sanchez‐Martinez & Nguyen, 2024). First, it is known that CAHS gelation provides increased excluded volume and stabilizing effects of proteins and protein complexes (Sanchez‐Martinez & Nguyen, 2024). In addition to excluded volume, the concentration of water within CAHS gels could help to provide an HBN to desiccation‐sensitive enzymes and help to maintain their stability (Arsiccio & Pisano, 2018; Belton & Gil, ¹⁹⁹⁴; Corradini et al., ²⁰¹³) (Figure 6).

Here, we have assessed protection under a single drying regime (vacuum desiccation). How CAHS protein protection varies under different drying regimes, for example, slower air drying or hyperosmotic shock, that are more relevant to real‐world conditions experienced by tardigrades, will be interesting to assess in future studies. Previous work has shown that wild‐type CAHS D and variants behave similarly (forming gel‐like fibers or not) under hyperosmotic stress conditions in living cells (Sanchez‐Martinez & Nguyen, 2024). Interestingly, in this study wild‐type CAHS D provided enhanced protection relative to both gelling (e.g., 2X_LR) and non‐gelling (e.g., LR) variants in cells exposed to hyperosmotic stress. These observations will motivate future studies not only on the effect of drying regimes on protection, but also studies looking at how insights from in vitro protection assays can be applied to more complex biological systems such as cells and whole organisms.

Additionally, wild‐type CAHS D forms concentration‐dependent gels, but ultimately, as water is lost, it will become a vitrified (glass‐like) solid (Boothby, 2021; Boothby & Tapia, ²⁰¹⁷). Owing to the disordered nature of CAHS D variants, we expect they would behave similarly when dried, regardless of whether they gel. While not identical, this behavior provides an interesting comparison to molecular condensates that are observed to undergo solid‐like aging transitions due to solvent depletion (Biswas & Potoyan, 2024). In our current work, we have only examined protection after 16 h of drying, but future longevity studies, which would assess how dry CAHS D and its variants “age” and how this ultimately impacts protection will be informative.

Overall, our study gives insight into the importance of CAHS protein gelation and its possible importance to desiccation tolerance as a whole. In addition, our study suggests that tuning properties of engineered IDP—such as water‐binding, helicity, and propensity to form higher‐order oligomers/gels—can guide the design of excipients tailored to protect specific clients. Beyond desiccation tolerance, our findings all contribute to a broader discussion on the functional significance of protein‐phase behavior.

4. MATERIALS AND METHODS

4.1. Cloning

CAHS D and variants g‐blocks (Integrated DNA Technologies) were codon optimized for expression in Escherichia coli and cloned into the pET28b expression vector using Gibson assembly (New England Biolabs). Clones were propagated in DH5α (Catalog C2987H, NEB). Sanger sequencing was used to confirm the full incorporation of CAHS D and variants into pET28b (Eton Bioscience).

4.2. Protein expression

The expression of CAHS variants follows the methods in Sanchez‐Martinez and Nguyen (2024). CAHS D and variant expression constructs were transformed in BL21 (DE3) competent E. coli strains (New England Biolabs) and were plated on LB agar plates containing 50 μg/mL Kanamycin. Large‐scale expression was performed in 1L LB/Kanamycin cultures, shaken at 37°C until an optical density of 0.6, at which point expression was induced using 1 mM IPTG for 4 h. Cells were harvested via centrifugation at 4000g at 4°C for 30 min. Cell pellets were resuspended in 5 mL of resuspension buffer (20 mM Tris, pH 7.5) and 30 μL of protease inhibitor (Sigma‐Aldrich, St. Louis, MO). Pellets were stored at −80°C.

4.3. Protein purification

Purification of CAHS D variants follows the methods in Sanchez‐Martinez et al. (2023). Pellets were allowed to thaw at room temperature and were heat‐lysed by boiling for 10 min. Insoluble components were removed through centrifugation at 5000g at 4°C for 30 min. The supernatant was collected and sterile‐filtered with 0.45 and 0.22 μm syringe filters (Foxx Life Sciences, Salem, NH) in succession. The filtered lysate was diluted 1:2 in buffer UA (8 M Urea, 50 mM sodium acetate, pH 4). The protein was purified using a cation‐exchange HiPrep SP HP 16/10 column (Cytivia, Marlborough, MA) on the AKTA Pure 25 L (Cytivia). Protein was eluted using a gradient of 0%–50% UB (8 M Urea, 50 mM sodium acetate, and 1 M NaCl, pH 4), over 20 column volumes.

Eluted fractions were confirmed using SDS‐PAGE and pooled for dialysis in 3.5 kDA MWCO dialysis tubing (SpectraPor 3 Dialysis Membrane, Sigma Aldrich). Pooled fractions were dialyzed at room temperature for 4 h against 20 mM sodium phosphate, pH 7.0. This was followed by an additional 6 rounds of dialysis in Milli‐Q water (18.2 MΩ cm). Proteins were then quantified fluorometrically (Qubit 4 Fluorometer, Invitrogen, Waltham, MA) and lyophilized (FreeZone 6, Labconco, Kansas City, MO) for 48 h. Purified, lyophilized protein was stored at −20°C.

4.4. Lactate dehydrogenase protection assay

LDH desiccation protection assays were performed in triplicate following the methods in Boothby et al., (2017). Protectants were reconstituted in a concentration range from 0.01 to 20 mg/mL in resuspension buffer (25 mM Tris, pH 7.0) and were co‐incubated with L‐Lactate Dehydrogenase (L‐LDH) from rabbit muscle (Catalog 10,127,230,001, Sigma‐Aldrich) at 0.1 mg/mL with a final volume of 100 μL. Each sample was split into 50 μL halves in which one set was stored at 4°C, while the other half was desiccated in a speed vacuum (OFP400, Thermo Fisher Scientific, Waltham, MA) for 16 h without heating. Following desiccation, all samples were reconstituted to a final volume of 250 μL with sterile water. Samples were added 1:100 to assay buffer (100 mM Sodium Phosphate, 2 mM Sodium Pyruvate, 1 mM NADH, pH 6). Conversion of NADH to NAD+ was measured by enzyme kinetics at 340 nm for 1 min by UV–VIS (NanodropOne, Thermo Fisher Scientific). The protective capacity was calculated as a ratio of NAD+ absorbance in desiccated samples normalized to non‐desiccated controls.

4.5. Citrate synthase protection assay

The Citrate Synthase Kit (Catalog CS0720‐1KT, Sigma‐Aldrich) was adapted for use in this assay (Chakrabortee et al., 2012; Goyal et al., ²⁰⁰⁵). All samples were prepared in triplicate, except desiccated negative control samples, which were prepared in quadruplicate, so that the extra sample could be used for assessment of desiccation efficiency. Lyophilized variants were resuspended in either purified water (samples to be desiccated) or 1X assay buffer (control samples) to a concentration of 20 g/L and diluted as necessary for lower concentrations. CS was added at a ratio of 1:10 to resuspended protectants. Non‐desiccated control samples were measured as described in the assay kit immediately following resuspension. Desiccated samples were subjected to 5–6 rounds of desiccation and rehydration (1‐h speedvac desiccation [Thermo Fisher Scientific] followed by resuspension in water). Following the 5th round of desiccation, a negative control sample was resuspended and assayed to determine if activity remained. If the negative control sample retained more than 10% activity, a 6th round of desiccation/rehydration was performed. After the final round of desiccation, samples were resuspended in 10 μL of cold 1X assay buffer. Samples were diluted 1:100 in the assay reaction mixture supplied, and all subsequent steps followed the kit instructions. The colorimetric reaction was measured for 90 s at 412 nm using the Spark 10 M (Tecan).

4.6. Rheometry

For the rheological experiments, CAHS D variants were dissolved in water at concentrations ranging from 20 to 1 mg/mL. Rheological characterization of the protein solutions was performed using a Discovery HR20 rheometer (TA Instruments) equipped with an 8 mm parallel plate and a stainless steel humidity chamber. The gap between the plates was set to 500 μm, and the humidity chamber was maintained at 50%. All tests were conducted at a constant temperature of 20°C, with a 30‐min soak time allowed for gel formation before each run. Samples were loaded onto the bottom plate, and any excess material was trimmed to prevent overflow, if required. The rheometer was programmed to perform a logarithmic sweep with 1% strain at angular frequencies ranging from 0.1 to 100 rad/s, recording five data points per decade. Raw loss modulus and Tan δ values were collected at an angular frequency of 10 (Kauzmann, 1959).

4.7. BS3 crosslinking

BS3 (bis[sulfosuccinimidyl] suberate) (Catalog A39266, Thermo Fisher Scientific, USA) is an 11.4 Å amine ester crosslinker. Crosslinking experiments were done according to the manufacturer's instructions. Proteins were resuspended in 20 mM HEPES at pH 7.5. 1 mM of BS3 crosslinker was added to CAHS D variants at PD50 concentrations. The mixture was incubated for 30 min at room temperature. The reaction was quenched by adding 2X lamelli gel loading buffer (Bio‐Rad catalog #1610737), and 10 μL of each sample was loaded on a 4%–20% Criterion™ TGX™ Precast Midi Protein Gel (Catalog 5671094, Bio‐Rad). Gels were imaged using the Azure 200 (Catalog AZI200‐01) gel imager. ImageJ was used to quantify band intensities for dimer: monomer calculations.

BS3 crosslinking experiments with CAHS D and clients (LDH or CS) were done by mixing 1 to 2 mM BS3 crosslinker with protein mixtures, each with a final concentration of 1 mg/mL.

4.8. Gel‐shift assay

1 mg/mL of CAHS D was mixed with an equal volume of each client (LDH or CS) at 1 mg/mL. Each sample was mixed at a 2:1 ratio with Native Sample Buffer (Catalog 1610738, Bio‐Rad). Native gel electrophoresis was run at 120 V, constant voltage, for 2–3 h. NativeMark™ Unstained Protein Standards (Catalog LC0725, Invitrogen™) were used to visualize protein size.

4.9. HT‐SAXS sample preparation

Lyophilized proteins were resuspended in 20 mM Tris HCl, pH = 7.00. Because of their instability in a plain tris buffer, the proteins CLC and N‐Term were resuspended in a Tris buffer containing 250 and 100 mM NaCl, respectively. All proteins were subsequently quantified with the Qubit Protein Assay from ThermoFisher Scientific (catalog# Q33212). Protein samples were then diluted with tris HCl into stocks of 8 and 4 mg/mL. 100 μL of these samples were then passed through a 0.22 μm filter. After allowing the mixture to settle, 30 μL of each sample was loaded onto an Axygen 96‐well Polypropylene PCR Microplate (Catalog 3596, Corning). Each protein sample was tested along with a unique buffer blank consisting of the same aliquot of buffer that was used to resuspend the protein. The plate was then sealed with an AxyMat Sealing Mat (Catalog AM‐96‐PCR‐RD, Corning) and wrapped in parafilm. It was shipped at 4°C to Lawrence Berkeley National Labs where it was tested on the HT‐SAXS beamline (12.3.1) (Dyer et al., 2014; Trame et al., ²⁰⁰⁴).

4.10. HT‐SAXS—Guinier analysis

All analyses were performed in BioXTas RAW v. 2.1.4 (Hopkins et al., 2018; Nielsen et al., ²⁰⁰⁹). For all proteins, a qmaxRg of less than or equal to 1.1 was used to fit the Guinier region (Kachala et al., 2015). Samples were checked for characteristic signs of protein aggregation or repulsion. All Guinier region fits and data points can be seen in Figure S4, along with the raw data points and a Kratky plot of each sample.

4.11. CD spectroscopy

Lyophilized CAHS D and its variants were resuspended in tris buffer pH 7 to a concentration of 10 μM. Protein concentrations were quantified using Qubit™ Protein Assay (Thermo Fisher Scientific, USA). Each variant was measured in either 1 or 0.01 mm quartz cuvettes in a CD spectrometer (AVIV Associates, Model 420, Lakewood, NJ). Each measurement was performed in three replicates.

4.12. Secondary structure analysis

The Beta Structure Selection method was used to calculate the detailed structure information from the CD spectra (Micsonai et al., 2023).

4.13. Time‐domain NMR sample preparation

Quantitated and lyophilized protein samples were transferred as powder into 10 mm TD‐NMR tubes (Wilmad Lab Glass) and resuspended in 500 μL of water to a final concentration of 1L, 5, 10, 15, and 20 g/L. Samples were left at room temperature for 5 min to solubilize. If solubilization was not occurring (determined visually), samples were moved to 55°C for intervals of 5 min until solubilized. T₂ relaxation measurements were taken immediately after solubilization.

4.14. Measurement of T₂ relaxation

T₂ relaxation measurements were performed using a Bruker mq20 minispec low‐field nuclear magnetic resonance spectrophotometer, with a resonance frequency of 19.65 MHz. Samples were kept at 25°C during measurements through the use of a chiller (F12‐MA, Julabo USA Inc., Allentown, PA) circulating a constant‐temperature coolant. T₂ free induction decays were measured using a Carr–Purcell–Meiboom–Grill pulse sequence with 8000 echoes and an echo time of 1000 μs. Pulse separation of 1.5 ms, recycle delay of 3 ms, and 32 scans were used for all samples. The gain was determined for each sample individually and ranged from 53 to 56. Conversion of the free induction decay to T₂ relaxation distribution was processed using the CONTIN ILT software provided by Bruker. Each variant was measured for the full concentration range, along with a water control.

AUTHOR CONTRIBUTIONS

Kenny Nguyen: Conceptualization; data curation; investigation; formal analysis; visualization; writing – original draft; writing – review and editing. Sourav Biswas: Investigation; data curation; methodology; writing – review and editing. Shraddha KC: Methodology; data curation; investigation; writing – review and editing. Annie Walgren: Methodology; investigation; writing – review and editing. Vincent Nicholson: Investigation; data curation; methodology; writing – review and editing. Charles Childs: Methodology; investigation; resources; writing – review and editing. Bryan X. Medina‐Rodriguez: Methodology; investigation; writing – review and editing. Vladimir Alvarado: Methodology; resources; writing – review and editing; supervision. Shahar Sukenik: Methodology; resources; writing – review and editing; funding acquisition; supervision. Alex Holehouse: Funding acquisition; methodology; writing – review and editing; supervision; resources. Thomas C. Boothby: Conceptualization; supervision; funding acquisition; project administration; writing – original draft; writing – review and editing.

CONFLICT OF INTEREST STATEMENT

The authors declare no competing interests.

Supporting information

Data S1: Supporting Information.

Data S2: Supporting Information.

Nguyen K, Biswas S, KC S, Walgren A, Nicholson V, Childs C, et al. A phase transition modulates the protective function of a tardigrade disordered protein during desiccation. Protein Science. 2025;34(10):e70300. 10.1002/pro.70300

DATA AVAILABILITY STATEMENT

All data and code associated with this manuscript are provided in Files S1 and S2, respectively.