Functional and Conformational Plasticity of an Animal Group 1 LEA Protein

Group 1 (Dur-19, PF00477, LEA_5) Late Embryogenesis Abundant (LEA) proteins are present in organisms from all three domains of life, Archaea, Bacteria, and Eukarya. Surprisingly, Artemia is the only genus known to include animals that express group 1 LEA proteins in their desiccation-tolerant life-history stages. Bioinformatics analysis of circular dichroism data indicates that the group 1 LEA protein AfLEA1 is surprisingly ordered in the hydrated state and undergoes during desiccation one of the most pronounced disorder-to-order transitions described for LEA proteins from A. franciscana. The secondary structure in the hydrated state is dominated by random coils (42%) and β-sheets (35%) but converts to predominately α-helices (85%) when desiccated. Interestingly, AfLEA1 interacts with other proteins and nucleic acids, and RNA promotes liquid–liquid phase separation (LLPS) of the protein from the solvent during dehydration in vitro. Furthermore, AfLEA1 protects the enzyme lactate dehydrogenase (LDH) during desiccation but does not aid in restoring LDH activity after desiccation-induced inactivation. Ectopically expressed in D. melanogaster Kc167 cells, AfLEA1 localizes predominantly to the cytosol and increases the cytosolic viscosity during desiccation compared to untransfected control cells. Furthermore, the protein formed small biomolecular condensates in the cytoplasm of about 38% of Kc167 cells. These findings provide additional evidence for the hypothesis that the formation of biomolecular condensates to promote water stress tolerance during anhydrobiosis may be a shared feature across several groups of LEA proteins that display LLPS behaviors.


Introduction
Since their discovery in cotton seeds by Dure et al. [1], the importance of late embryogenesis abundant (LEA) proteins in the abiotic stress tolerance of plants has been firmly established. However, their actual function(s) are still largely unresolved [2][3][4]. Research over the past two decades has extended the role(s) of these intrinsically disordered proteins (IDPs) in promoting desiccation tolerance to animals [5,6]. Based on conserved sequence motifs, several LEA classifications schemes have been proposed, resulting in 6 to 11 distinct protein groups or families [7]. Using the classification proposed by Tunnacliffe and Wise, most investigated anhydrobiotic animals express LEA proteins from Group 3 (PF02987) [3]. The only exception is the brine shrimp, Artemia franciscana, which expresses LEA proteins from two additional families. These LEA proteins are from Group 1 (PF00477) and Group 6 (PF04927) and are also only expressed in the anhydrobiotic encysted embryos [5,[8][9][10]. Compared to other anhydrobiotic animals, the reasons for the more extensive LEA repertoire in Artemia are still unresolved. However, despite a uniquely broad range of expressed LEA proteins, the knockdown of Group 1 LEA proteins by RNAi reduced the cyst desiccation tolerance by over 90% [11]. Therefore, expressed proteins from other groups do not readily substitute the anhydrobiotic functions performed by Group 1 LEA proteins

Protein Cloning, Expression, and Purification
DNA encoding for the 180 amino acids-long Af LEA1 (highly similar to ABR67402; described in: [12]) polypeptide was cloned into the Ptxb1 (NEB Biolabs, Ipswich, MA, USA) vector and expressed using isopropyl-β-1-thiogalactopyranoside (IPTG) induction as previously described [24]. Briefly, BL21 Star strain Escherichia coli cells transformed with this vector were cultured on agarose plates containing 100 µg/mL ampicillin in Luria Bertani (LB) medium. For protein induction, single colonies were grown to an optical density of~0.6 at λ = 595 nm overnight at 37 • C and cells were harvested 2 h after adding 0.4 mM IPTG. The bacteria were harvested at 5000× g for 30 min at 4 • C and the pellets were resuspended in buffer A (500 mM NaCl, 50 mM Tris-HCl, pH 8.5) containing 1 mM phenylmethylsulphonyl fluoride (PMSF) and stored at −80 • C until protein purification.
For protein purification, the bacterial pellets were thawed and sonicated (Q500, Qsonica, Newtown, CT, USA), then centrifuged at 20,000× g for 30 min at 4 • C to remove cellular debris. The resulting supernatant was applied to a column containing 15 mL chitin resin (NEB Biolabs, Ipswich, MA, USA) and the column was washed with 3 column volumes of buffer A. Self-cleavage of the intein protein was induced by adding 50 mM dithiothreitol at 4 • C for 48 h and the eluted protein was dialyzed overnight against 20 mM TRIS-HCl at a pH of 8.0. Further purification was performed using a fast-performance liquid chromatography (FPLC) system (AKTA, Cytiva Life Sciences, Marlborough, MA, USA) and a 1 mL Resource Q anion exchange column (Cytiva Life Sciences, Marlborough, MA, USA). Bound Af LEA1 was eluted from the column using a gradient of 20 mM TRIS-HCl at pH 8.0 and NaCl concentrations from 0 to 125 mM over 10 column volumes. Af LEA1 eluted from the column was 99% pure based on band intensity measured using SDS PAGE and the protein was dialyzed against 50 mM phosphate buffer, pH 7.0. Aliquots of purified Af LEA1 were snap-frozen in liquid N 2 and stored at −80 • C until used for experimentation.

Circular Dichroism
Af LEA1 protein (100 µg/mL in 20 mM phosphate buffer at a pH of 7.0) was measured in a sealed quartz cuvette with a path length 0.1 cm (Starna Scientific, Atascadero, CA, USA) using a wavelength range from 280 nm to 185 nm using a J-1500 circular dichroism spectrophotometer (Jasco, Easton, MD, USA). To reduce the light-scattering associated with a dried sample, Af LEA1 was dialyzed into ultrapure water three times at a ratio of 1:1000 protein solution to water. Next, Af LEA1 at 1 mg/mL in water was repeatedly plated on an open 0.01 cm path-length quartz cuvette (Starna Scientific, Atascadero, CA, USA). Each layer was rapidly dried by incubating them for 1 h at 0% relative humidity over anhydrous calcium sulfate. The rapid drying process produced an amorphous protein glass with minimal light scattering. Once the protein layer reached the thickness of the 0.01 cm path length, 1 µL of ultrapure water was added, allowing the cuvette to be assembled. A final incubation for 24 h at 25 • C and 0% relative humidity produced a protein glass of Af LEA1 without observable light scattering. Data were averaged over 5 measurements taken in 1 nm intervals. Secondary structure predictions were performed using the CONTIN and

Sample Preparation for Light Microscopy and Electron Scanning Microscopy (SEM)
A droplet of Af LEA1 in H 2 O was plated onto microscope slides directly adjacent to a droplet of ultrapure water. The droplets were then connected to allow diffusion of the protein and produce a protein gradient over 10 min at ambient relative humidity and 25 • C. The samples were dried at 0% relative humidity over anhydrous calcium sulfate at 25 • C to remove any residual water. Glass coverslips were placed above the samples and sealed with nail polish after desiccation-preserving the samples. The samples were stored in a sealed container at 0% RH until viewed using a specimen microscope. Sample preparation for SEM was performed as described above, with the difference that an aluminum SEM stage was used to deposit the protein solution. Once completely dried, the samples were sputter-coated with a 10 nm layer of gold to prevent sample rehydration reduce charging artifacts. The samples were stored at 0% RH until viewed using an electron microscope (Jeol, Peabody, MA, USA).

Lactate Dehydrogenase Activity Assays
Cell lysates were obtained by sonication of 20 × 10 6 Kc167 cells (Drosophila melanogaster) in 100 mM phosphate buffer at a pH of 6.4. Cell debris was removed by centrifugation at 15,000× g for 10 min. The supernatant was diluted to a total protein concentration of 2 mg/mL with phosphate buffer, and the total protein concentrations were confirmed via Bradford assays. Af LEA1 or BSA was then added to the lysate to yield a total concentration of 400 µg/mL of Af LEA1 or BSA, and the endogenous LDH activity was recorded. Aliquots of 50 µL were then desiccated for 7 days in a sealed container at 25 • C and 0% RH. Samples were rehydrated with 100 µL phosphate buffer, and activity was measured using UV-VIS spectrophotometry (UV-1800, Shimadzu, Tokyo, Japan). Absorbance was monitored at λ = 340 nm using the kinetics mode while the sample was stirred at 500 rpm, and the sample temperature was maintained at 25 • C.
To distinguish between protection of LDH and repair of LDH by Af LEA1, pure LDH was desiccated in the presence or absence of 400 µg/mL Af LEA1 or BSA in 100 mM sodium phosphate buffer at a pH of 6.4. Commercially obtained LDH (Millipore-Sigma, St. Louis, MO, USA) was diluted to 0.2 mg/mL with H 2 O and dialyzed against 100 mM phosphate buffer, pH 6.5. Initial LDH activity was determined, and 25 µL aliquots of samples containing LDH, LDH plus BSA, or LDH plus Af LEA1 were placed in microtubes and desiccated for 7 days at 0% RH. Samples were rehydrated with 50 µL of phosphate buffer. Additional samples of purified LDH were desiccated in the absence of Af LEA1 and BSA for 7 days at 0% RH and rehydrated on ice with either 50 µL of phosphate buffer or phosphate buffer containing Af LEA1 or BSA at 400 µg/mL to investigate the potential repair mechanisms. LDH activity after rehydration was measured as described above.

Screening AfLEA1 for RNA-Induced Liquid-Liquid Phase Separation
A solution resembling the crowded cytoplasm of the diapause cysts of A. franciscana (32 mM NaCl, 98 mM KCl, 11 mM K 2 PO 4 , 5 mM CaCl 2 , 340 mM trehalose, 2.9% w/v glycerol, and 25% Ficoll 400, pH of 6.5) was prepared, and 150 µg/mL of Af LEA1 was added to determine whether Af LEA1 undergoes LLPS during desiccation. The protein solution was deposited onto microscope slides, and the solutions were allowed to desiccate at ambient humidity (83% RH) and temperature (22-26 • C) while observed using an inverted microscope. RNA (20 ng/mL final concentration) isolated from A. franciscana using a total RNA isolation kit (Qiagen, Germantown, MD, USA) was added to the protein droplet to determine the impact on the LLPS behavior of Af LEA1.

SDS-PAGE
Samples of Af LEA1 obtained after affinity or ion-exchange chromatography were dialyzed against 50 mM phosphate buffer at a pH of 7.0 and frozen in 10 mL aliquots. Before electrophoresis, the protein was boiled in Laemmli buffer (Bio-Rad, Hercules, CA, USA), and 2 µg of Af LEA1 was run per lane of a 10% SDS-PAGE (37.5:1, Bio-Rad, Hercules, CA, USA). A kaleidoscope protein ladder was used as a molecular weight standard (Bio-Rad, Hercules, CA, USA). The gel was run at 90 V for 1 h using a running buffer containing 25 mM Tris, 192 mM glycine, and 0.1% SDS at a pH of 8.3, and gels were destained overnight with 30% methanol and 10% glacial acetic acid in H 2 O.

Cell Culture, Transgenic Cell Line, and Confocal Microscopy
Cell culture, transgenic cell line development, and confocal microscopy were performed as previously described for AfrLEA6 with the following differences [21]. In place of Schneider's media, IPL-41 insect culture media (Gibco, Waltham, MA, USA) was supplemented with 2 g/L tryptone phosphate broth (VWR, Atlanta, GA, USA) and 10% heatinactivated FBS (Atlanta Biologicals, Flowery Branch, GA, USA) and used to culture the Kc167 cells. In addition, transfected and control cell lines were stained for only 2 min with Nile Red to prevent fluorescence saturation in Af LEA1-expressing cells. Cells co-expressing Af LEA1-mCherry and GFP were stained with 200 µL of 100 nM MitoView Blue in DPBS (Biotium, Fremont, CA, USA) for 10 min. The following primers were used to clone and amplify untagged AfLEA1: 5 -CGGAATTCCAAACATGGAACTGTCGTCGAGTAAGCTG-3 and 5 -TAATTGCGGCCGCATTTCTGTCTTGCGAGACCTCCTTTTTG-3 . The following primers were used to clone and amplify a chimeric protein composed of AfLEA1 and mCherry at the c-terminus: 5 -CGGCGAATTCATGGAACTGTCGTCGAGTAAGCTG-3 and 5 -AGGGCGGCCGCATTTCTGTCTTGCGAGACCTCCTTTTTG-3 .

Protein Cloning, Expression, and Purification
The protein Af LEA1 was readily expressed in E. coli, and an approximately 90% purity could be reached using chitin-affinity chromatography, as shown by SDS-PAGE ( Figure 1A). Surprisingly, when the crude protein sample was further purified using a strong anion exchanger, Af LEA1 elutes in two distinct fractions ( Figure 1B). Still, the protein in both fractions migrates identically during SDS-PAGE and has a purity of over 99% ( Figure 1A). The presence of two elution peaks containing Af LEA1 protein that behaves identically during electrophoresis may indicate higher-order assemblies that are lost during the denaturation step in electrophoresis or distinct conformational states undetectable by SDS-PAGE. denaturation step in electrophoresis or distinct conformational states undetectable by SDS-PAGE. Bioinformatics suggests that AfLEA1 has high conformational plasticity [17], which supports the hypothesis that the protein may exist in more than one tertiary state when bound to the column's polystyrene/divinylbenzene polymer matrix, causing two distinct elution fractions. Furthermore, a trend was observed indicating the concentration of salt required for elution of AfLEA1 correlates negatively with the quantity of AfLEA1 bound to the column for both elution fractions ( Figure S1, and Table S1, Supplementary Materials). This behavior may indicate interactions among AfLEA1 molecules on the column can change the affinity of the protein towards the strong anion exchanger for both conformational states. However, a more thorough investigation of this behavior would be needed to strengthen this interpretation.
The apparent molecular mass of AfLEA1 in both elution fractions was measured using size-exclusion chromatography to exclude SDS-sensitive higher-order oligomers (Figure 2). Independent of the fraction AfLEA1 eluted in during anion-exchange chromatography, size-exclusion chromatography indicates that AfLEA1 has an apparent molecular weight of 33 kDa, which is ~75% larger than the amino acid composition-based calculated mass (~19.7 kDa). However, intrinsically disordered proteins can have apparent molecular sizes up to 12 times as large as a similarly sized globular protein [30]. Overall, these results suggest AfLEA1 is likely in a molten globule state showing some secondary structure but not a tightly packed tertiary structure. In the molten globule state, when evaluated using size-exclusion chromatography, proteins can have an apparent molecular weight of up to twice that of a tightly packed globular protein of a similar length. Bioinformatics suggests that Af LEA1 has high conformational plasticity [17], which supports the hypothesis that the protein may exist in more than one tertiary state when bound to the column's polystyrene/divinylbenzene polymer matrix, causing two distinct elution fractions. Furthermore, a trend was observed indicating the concentration of salt required for elution of Af LEA1 correlates negatively with the quantity of Af LEA1 bound to the column for both elution fractions ( Figure S1 and Table S1, Supplementary Materials). This behavior may indicate interactions among Af LEA1 molecules on the column can change the affinity of the protein towards the strong anion exchanger for both conformational states. However, a more thorough investigation of this behavior would be needed to strengthen this interpretation.
The apparent molecular mass of Af LEA1 in both elution fractions was measured using size-exclusion chromatography to exclude SDS-sensitive higher-order oligomers ( Figure 2). Independent of the fraction Af LEA1 eluted in during anion-exchange chromatography, size-exclusion chromatography indicates that Af LEA1 has an apparent molecular weight of 33 kDa, which is~75% larger than the amino acid composition-based calculated mass (~19.7 kDa). However, intrinsically disordered proteins can have apparent molecular sizes up to 12 times as large as a similarly sized globular protein [30]. Overall, these results suggest Af LEA1 is likely in a molten globule state showing some secondary structure but not a tightly packed tertiary structure. In the molten globule state, when evaluated using size-exclusion chromatography, proteins can have an apparent molecular weight of up to twice that of a tightly packed globular protein of a similar length. Biomolecules 2022, 12, x 7 of 22

AfLEA1 Secondary Structural Analysis
The secondary structure of AfLEA1 was investigated by circular dichroism under three conditions to explore the conformational transition of AfLEA1 during the reduced availability of water ( Figure 3). In the hydrated state, about 40% AfLEA1 contains defined secondary structure motifs, and the protein is composed of 5% α-helices, 35% β-sheets, 18% turns, and 42% random coils. While AfLEA1 can be classified, like other LEA proteins, as a mostly intrinsically disordered protein, it is surprisingly ordered in the hydrated state. For example, the Group 3 LEA proteins AfrLEA2 and AfrLEA3m (A. franciscana) are only ordered to 21% and 25%, respectively [31]. However, AfLEA1 undergoes a dramatic conformational transition in the desiccated state. The protein is nearly 100% ordered and composed of 85% α-helices, 5% β-sheets, and 10% turns after removing the water ( Figure  3). Despite being more structured than most LEA proteins in the hydrated state, this conformational transition is the most dramatic shift from disorder to order reported for an LEA protein from A. franciscana. The transition from the hydrated to the desiccated conformation is partly mimicked by the molecular crowding effect observed in a solution containing 2% SDS (Figure 3). In the presence of SDS, AfLEA1 is slightly less ordered (38%) than in the hydrated state, and its β-sheets appear to transition into α-helices before the random coils. Generally, the secondary structure measurements and predictions produced by circular dichroism represent the average conformation of an ensemble of the different conformational states in the beam's path [32]. It appears that during desiccation, AfLEA1 is transitioning into a more uniform conformation compared to the hydrated

AfLEA1 Secondary Structural Analysis
The secondary structure of Af LEA1 was investigated by circular dichroism under three conditions to explore the conformational transition of Af LEA1 during the reduced availability of water ( Figure 3). In the hydrated state, about 40% Af LEA1 contains defined secondary structure motifs, and the protein is composed of 5% α-helices, 35% β-sheets, 18% turns, and 42% random coils. While Af LEA1 can be classified, like other LEA proteins, as a mostly intrinsically disordered protein, it is surprisingly ordered in the hydrated state. For example, the Group 3 LEA proteins AfrLEA2 and AfrLEA3m (A. franciscana) are only ordered to 21% and 25%, respectively [31]. However, Af LEA1 undergoes a dramatic conformational transition in the desiccated state. The protein is nearly 100% ordered and composed of 85% α-helices, 5% β-sheets, and 10% turns after removing the water ( Figure 3). Despite being more structured than most LEA proteins in the hydrated state, this conformational transition is the most dramatic shift from disorder to order reported for an LEA protein from A. franciscana. The transition from the hydrated to the desiccated conformation is partly mimicked by the molecular crowding effect observed in a solution containing 2% SDS (Figure 3). In the presence of SDS, Af LEA1 is slightly less ordered (38%) than in the hydrated state, and its β-sheets appear to transition into α-helices before the random coils. Generally, the secondary structure measurements and predictions produced by circular dichroism represent the average conformation of an ensemble of the different conformational states in the beam's path [32]. It appears that during desiccation, Af LEA1 is transitioning into a more uniform conformation compared to the hydrated state, which is a strong indication of the disorder-to-order regulation of its protective functions.
state, which is a strong indication of the disorder-to-order regulation of its protective functions. Figure 3. Circular dichroism analysis of hydrated and desiccated AfLEA1, wherein α-helices (red), β-sheets (green), turns (blue), and random coils (magenta) are represented as proportions of the total protein structure. In the hydrated state, the secondary structure of AfLEA1 was on average 5% α-helices, 35% β-sheets, 18% turns, and 42% random coils. In the desiccated state, the secondary structure of AfLEA1 was on average 85% α-helices, 5% β-sheets, and 10% turns. In the presence of 2% SDS, the secondary structure of AfLEA1 was on average of 25% α-helices, 13% β-sheets, 16% turns, and 46% random coils (for CD spectra see Figure S2, Supplementary Materials).
The sequence of AfLEA1 is highly repetitive and is mainly composed of the eight repeats of the Group 1 consensus sequence 'GGOTRREQLGEEGYSQMGRK' [33,34] (Figure 4A). AfLEA1's propensity for α-helices is predicted by the regular appearance of alanine-arginine-alanine "helix cap" motifs found at the ends of the repeating Group 1 LEA motif [35][36][37]. Given the amount of predicted α-helical structure, an α-helical protein projection is informative. Unlike AfrLEA2 and AfrLEA3m [38], AfLEA1 does not have sufficient secondary structure orientation of its charged residues to offer alpha-helical stabilization ( Figure 4B) [39]. However, the protein does have a thin hydrophobic face, which can stabilize α-helical formation in the absence of water [40]. Nevertheless, the hydrophobic moment is too low to offer sufficient stabilization to explain AfLEA1's α-helix content measured in the desiccated state [17,41]. . Circular dichroism analysis of hydrated and desiccated Af LEA1, wherein α-helices (red), β-sheets (green), turns (blue), and random coils (magenta) are represented as proportions of the total protein structure. In the hydrated state, the secondary structure of Af LEA1 was on average 5% α-helices, 35% β-sheets, 18% turns, and 42% random coils. In the desiccated state, the secondary structure of Af LEA1 was on average 85% α-helices, 5% β-sheets, and 10% turns. In the presence of 2% SDS, the secondary structure of Af LEA1 was on average of 25% α-helices, 13% β-sheets, 16% turns, and 46% random coils (for CD spectra see Figure S2, Supplementary Materials).
The sequence of Af LEA1 is highly repetitive and is mainly composed of the eight repeats of the Group 1 consensus sequence 'GGOTRREQLGEEGYSQMGRK' [33,34] (Figure 4A). Af LEA1's propensity for α-helices is predicted by the regular appearance of alanine-arginine-alanine "helix cap" motifs found at the ends of the repeating Group 1 LEA motif [35][36][37]. Given the amount of predicted α-helical structure, an α-helical protein projection is informative. Unlike AfrLEA2 and AfrLEA3m [38], Af LEA1 does not have sufficient secondary structure orientation of its charged residues to offer alpha-helical stabilization ( Figure 4B) [39]. However, the protein does have a thin hydrophobic face, which can stabilize α-helical formation in the absence of water [40]. Nevertheless, the hydrophobic moment is too low to offer sufficient stabilization to explain Af LEA1's α-helix content measured in the desiccated state [17,41]. Biomolecules 2022, 12, x 9 of 22

Computational Interpretation of the Structure of AfLEA1 in the Desiccated State
The I-Tasser program uses X-ray crystallography data to predict protein structure. The program is surprisingly effective at predicting the ordered arrangements of LEA proteins that undergo conformational transitions during desiccation [25], and the secondary structure predictions of I-Tasser fall within 2% of the secondary structure content measured by circular dichroism for desiccated AfLEA1. The α-helices predicted by I-Tasser appear to be organized into a sizeable helical bundle composed of four helix-turn-helix structures, with each helix formed by a single Group 1 LEA consensus sequence and a short spacer containing a histidine residue ( Figure 5A). The α-helices propagate from the alanine-arginine-alanine caps, as previously hypothesized [17]. Still, the internal space of the helical bundle is highly enriched in positively charged residues and aromatic residues ( Figure 5B).

Computational Interpretation of the Structure of AfLEA1 in the Desiccated State
The I-Tasser program uses X-ray crystallography data to predict protein structure. The program is surprisingly effective at predicting the ordered arrangements of LEA proteins that undergo conformational transitions during desiccation [25], and the secondary structure predictions of I-Tasser fall within 2% of the secondary structure content measured by circular dichroism for desiccated Af LEA1. The α-helices predicted by I-Tasser appear to be organized into a sizeable helical bundle composed of four helix-turn-helix structures, with each helix formed by a single Group 1 LEA consensus sequence and a short spacer containing a histidine residue ( Figure 5A). The α-helices propagate from the alaninearginine-alanine caps, as previously hypothesized [17]. Still, the internal space of the helical bundle is highly enriched in positively charged residues and aromatic residues ( Figure 5B).
Although aromatic residues can interact with positively charged residues due to their delocalized π orbitals [42], the small number of aromatic residues does not appear to be sufficient to prevent the electrostatic repulsion of the α-helices in Af LEA1. Despite this source of instability, this structure is similar to a de novo-designed protein built to emulate armadillo repeat proteins (RCSB PDB ID: 5CWH), which has remarkably similar primary, secondary, and tertiary structures [43] to Af LEA1. Armadillo repeat proteins are highly stable and are commonly involved in cell signaling and misfolded protein degradation due to their promiscuous binding behaviors [44,45]. The structure of Af LEA1 appears to be stabilized by similarly charged stripes found on helices proposed for AfrLEA2 and AfrLEA3m [38]. The charge patterns are generated in these two LEA proteins from Group 3 through a specific arrangement of basic and acidic amino acids in the primary sequence ( Figure 6A,B). If I-Tasser has accurately predicted the tertiary structure of Af LEA1, then charged stripes arise in this LEA protein from the tertiary association of alpha-helices, thereby stabilizing these helical clusters ( Figure 6C) [17]. Although aromatic residues can interact with positively charged residues due to their delocalized π orbitals [42], the small number of aromatic residues does not appear to be sufficient to prevent the electrostatic repulsion of the α-helices in AfLEA1. Despite this source of instability, this structure is similar to a de novo-designed protein built to emulate armadillo repeat proteins (RCSB PDB ID: 5CWH), which has remarkably similar primary, secondary, and tertiary structures [43] to AfLEA1. Armadillo repeat proteins are highly stable and are commonly involved in cell signaling and misfolded protein degradation due to their promiscuous binding behaviors [44,45]. The structure of AfLEA1 appears to be stabilized by similarly charged stripes found on helices proposed for AfrLEA2 and Afr-LEA3m [38]. The charge patterns are generated in these two LEA proteins from Group 3 through a specific arrangement of basic and acidic amino acids in the primary sequence ( Figure 6A,B). If I-Tasser has accurately predicted the tertiary structure of AfLEA1, then charged stripes arise in this LEA protein from the tertiary association of alpha-helices, thereby stabilizing these helical clusters ( Figure 6C) [17].

Behavior of AfLEA1 during Desiccation
Although AfLEA1 appears to evenly distribute itself as an amorphous, glassy deposition during fast desiccation in pure water, as used for circular dichroism, it undergoes crystallization when dried more slowly for hours rather than minutes ( Figure 7A). The formation of crystals indicates that AfLEA1 assumes a uniform structure in the desiccated

Behavior of AfLEA1 during Desiccation
Although Af LEA1 appears to evenly distribute itself as an amorphous, glassy deposition during fast desiccation in pure water, as used for circular dichroism, it undergoes crystallization when dried more slowly for hours rather than minutes ( Figure 7A). The formation of crystals indicates that Af LEA1 assumes a uniform structure in the desiccated state, as indicated by the circular dichroism data and structural predictions by I-Tasser. Unfortunately, scanning electron microscopy (SEM) reveals that the crystals formed by Af LEA1 during slow drying are of insufficient quality to employ X-ray crystallography to verify the structure of Af LEA1 in the desiccated state ( Figure 7B). However, SEM also revealed spherical structures in the desiccated sample similar to those reported for the Group 6 LEA protein AfrLEA6, which is known to undergo LLPS in vitro during desiccation [21]. The program catGranule predicts that AfLEA1 has an exceptionally high propensity to undergo LLPS, particularly when interacting with RNA ( Figure 8A,B). Desiccation experiments were conducted in a buffer system that resembles the cellular environment of the encysted embryos of A. franciscana to determine whether AfLEA1 undergoes LLPS in a physiologically more relevant background. At high levels of desiccation, AfLEA1 does appear to undergo LLPS, but this occurs at nearly complete desiccation ( Figure 9A), and at these extreme levels of water loss, the observed phenomenon may not be physiologically relevant. Behaviors such as salt-induced precipitation, which may cause proteins to separate from solution, are likely commonplace in the drying cell when protein-protein interactions become unavoidable [46]. The occurrence of LLPS at extreme dehydration may simple be a consequence of water loss, not a protective mechanism during anhydrobiosis. After adding total RNA isolated from A. franciscana to AfLEA1 samples before drying, the protein undergoes an LLPS readily after minor water removal by evaporative drying ( Figure 9B). However, additional studies are required to verify that RNA is selectively incorporated into the AfLEA1 dense phase and if the AfLEA1 condensate offers protection against RNA degradation. The program catGranule predicts that Af LEA1 has an exceptionally high propensity to undergo LLPS, particularly when interacting with RNA ( Figure 8A,B). Desiccation experiments were conducted in a buffer system that resembles the cellular environment of the encysted embryos of A. franciscana to determine whether Af LEA1 undergoes LLPS in a physiologically more relevant background. At high levels of desiccation, Af LEA1 does appear to undergo LLPS, but this occurs at nearly complete desiccation ( Figure 9A), and at these extreme levels of water loss, the observed phenomenon may not be physiologically relevant. Behaviors such as salt-induced precipitation, which may cause proteins to separate from solution, are likely commonplace in the drying cell when protein-protein interactions become unavoidable [46]. The occurrence of LLPS at extreme dehydration may simple be a consequence of water loss, not a protective mechanism during anhydrobiosis. After adding total RNA isolated from A. franciscana to Af LEA1 samples before drying, the protein undergoes an LLPS readily after minor water removal by evaporative drying ( Figure 9B). However, additional studies are required to verify that RNA is selectively incorporated into the Af LEA1 dense phase and if the Af LEA1 condensate offers protection against RNA degradation. Biomolecules 2022, 12, x 13 of 22  The readily observed phase transition behavior of some LEA proteins may warrant revisiting the hydration buffering hypothesis. LLPS behavior involves the reordering of water as water molecules interact with the separating protein to produce a unique fluid with different surface tension, resulting in spherical protein-water droplets. The water contained in the protein-water phase can have properties that vary significantly from the bulk fraction of cytosolic water [47]. The hydration buffer hypothesis postulates that water  The readily observed phase transition behavior of some LEA proteins may warrant revisiting the hydration buffering hypothesis. LLPS behavior involves the reordering of water as water molecules interact with the separating protein to produce a unique fluid with different surface tension, resulting in spherical protein-water droplets. The water contained in the protein-water phase can have properties that vary significantly from the bulk fraction of cytosolic water [47]. The hydration buffer hypothesis postulates that water The readily observed phase transition behavior of some LEA proteins may warrant revisiting the hydration buffering hypothesis. LLPS behavior involves the reordering of water as water molecules interact with the separating protein to produce a unique fluid with different surface tension, resulting in spherical protein-water droplets. The water contained in the protein-water phase can have properties that vary significantly from the bulk fraction of cytosolic water [47]. The hydration buffer hypothesis postulates that water molecules that are retained and interact at the surface of LEA proteins during dehydration serve as a water reservoir during desiccation [23,48]. In addition to LEA proteins serving as a water reservoir by organizing water molecules at the protein surface, a new organizational feature of water may occur inside of the protein-water droplet that protects desiccation-sensitive biomolecules at different stages of dehydration. Furthermore, changes in water-loss kinetics caused by some LEA proteins could be caused by the different physiochemical behavior of the water fraction in proteinaceous droplets. While associating biomolecular condensates with a general mechanism of desiccation tolerance still appears somewhat premature, research providing evidence for a significant role of proteinaceous phase transitions as part of the anhydrobiotic toolbox has begun to accumulate [6,17,21,49,50].

AfLEA1 Protection of Lactate Dehydrogenase (LDH) Activity during Desiccation and Rehydration
Although the above results may indicate that Af LEA1 is involved in RNA stabilization, one of the leading hypotheses for LEA protein function is the protection of other proteins and membranes during desiccation and rehydration via molecular shielding. Therefore, LDH was desiccated in the presence of Af LEA1 or BSA ( Figure 10). Endogenous LDH activity in Kc167 cell lysates that lack LEA proteins was measured to simulate cellular conditions more closely than experiments using purified enzymes can achieve. A significantly higher residual LDH activity after rehydration was observed in samples containing Af LEA1, whereas BSA did not show any improvement over the control ( Figure 10). This observation indicates that Af LEA1 offers protection against desiccation-induced protein denaturation, which could have been due to Af LEA1 having a chaperone-like refolding activity. Purified LDH was dried in the absence of Af LEA1 or BSA, and samples were rehydrated with a phosphate buffer or buffer containing either Af LEA1 or BSA to test if enzyme activity is higher if rehydrated with protein-containing solutions. Furthermore, the purified enzyme was dried in the presence or absence of these proteins and rehydrated with phosphate buffer. Af LEA1 preserved nearly 100% of LDH activity during desiccation if added before the onset of water removal but failed to affect LDH activity when only included in the rehydration solution ( Figure 11). Therefore, Af LEA1 improved the desiccation tolerance through interactions with LDH during drying or in the desiccated state. BSA offered lesser protection against desiccation-induced loss of LDH activity to the purified enzyme than Af LEA1 when desiccated together with LDH and did not improve the LDH activity if only present during rehydration ( Figure 11). It is noteworthy that the protection conferred by BSA to LDH activity was only observed for experiments using the purified enzyme but not in a more complex proteome sample. These results again demonstrate that the insights gained by experiments using purified targets are limited, and caution is warranted when extending these results to physiologically relevant contexts.

AfLEA1 Expressed in Kc167 Cells from D. melanogaster Cells Localizes to the Cytoplasm and Improves Cellular Structural Integrity during Desiccation
Kc167 cells from D. melanogaster ectopically co-expressing Af LEA1-mCherry and GFP were stained with MitoView Blue and imaged using confocal microscopy. Af LEA1 was primarily found to localize to the cytoplasm and not the mitochondria, corroborating previously published results ( Figure 12). Structures resembling proteinaceous condensates were observed in approximately 38% of Af LEA1-expressing cells when observed by confocal microscopy ( Figure 12 and Figure S3, Supplementary Materials). These structures were not detected in the previous work, utilizing conventional fluorescence microscopy [12]. The exclusion of GFP from these condensates suggests that this compartment may be an LLPS of Af LEA1, an interpretation strengthened by the observation that Af LEA1 in vitro readily undergoes an LLPS in the presence of RNA ( Figure 9B). In addition, Af LEA1 appeared to accumulate in the nucleus in approximately 23% of cells ( Figure 12 and Figure  S3, Supplementary Materials). We speculate that the formation of an Af LEA1 condensate and the protein accumulation in the nucleus may be due to interactions with nucleic acids that differ among cell cycle stages. However, additional cell-based studies exploring alternative expression and visualization approaches are needed to confirm any physiological significance of these observations.  . Residual activity of purified LDH was significantly increased when desiccated in the presence of BSA or AfLEA1 and then rehydrated with 100 mM sodium phosphate buffer at pH 6.5 (light gray box plots). LDH activity was not restored after desiccation in the absence of BSA or AfLEA1, and then rehydrated with 100 mM sodium phosphate buffer at pH 6.5 containing BSA or AfLEA1 (dark gray box plots). Control samples (white box plot) were desiccated and rehydrated in Figure 10. Af LEA1 (0.4 mg/mL) addition protects endogenous LDH activity in Kc167 cell lysates after desiccation and rehydration, while addition of BSA (0.4 mg/mL) did not significantly increases recovered LDH activity after rehydration compared to the control (n = 9; ±SD, p < 0.05; * different from control).  . Residual activity of purified LDH was significantly increased when desiccated in the presence of BSA or AfLEA1 and then rehydrated with 100 mM sodium phosphate buffer at pH 6.5 (light gray box plots). LDH activity was not restored after desiccation in the absence of BSA or AfLEA1, and then rehydrated with 100 mM sodium phosphate buffer at pH 6.5 containing BSA or AfLEA1 (dark gray box plots). Control samples (white box plot) were desiccated and rehydrated in Figure 11. Residual activity of purified LDH was significantly increased when desiccated in the presence of BSA or Af LEA1 and then rehydrated with 100 mM sodium phosphate buffer at pH 6.5 (light gray box plots). LDH activity was not restored after desiccation in the absence of BSA or Af LEA1, and then rehydrated with 100 mM sodium phosphate buffer at pH 6.5 containing BSA or Af LEA1 (dark gray box plots). Control samples (white box plot) were desiccated and rehydrated in the absence of BSA or Af LEA1 (n = 9-18; ±SD, p < 0.05; different letters indicate significant differences between groups).
During desiccation, Kc167 cells ectopically expressing untagged Af LEA1 displayed a robust increase in intracellular viscosity over the vector control cells when stained with Nile Red (Figure 13). Nile Red is a solvatochromatic dye that can increase in fluoresce intensity with increasing viscosity of the cytoplasm [51][52][53][54][55]. Increasing intracellular viscosity during modest dehydration likely helps impede molecular movement and prevent undesirable chemical reactions until more efficient glasses of protein and sugar can form [5,56,57]. Furthermore, the fusion of the plasma membranes during desiccation was significantly reduced in Af LEA1-expressing cells compared to the vector control. These results are like those obtained using AfrLEA6, possibly suggesting that both proteins utilize similar molecular mechanisms to confer protection during desiccation. However, despite sharing similar promiscuous protection mechanisms, different groups of biomolecules may be targeted by both proteins (e.g., proteins, nucleic acids, or lipids).   1 and 2). In some cells, Af LEA1-mCherry was also observed to undergo an LLPS and exclude GFP from the biomolecular condensate (Example 1). Af LEA1-mCherry was observed to accumulate in the nucleus of other cells (Example 2). Cells were stained with 200 µL of MitoView Blue, and Af LEA1-mCherry does not localize to the mitochondria. Images are representative images, and fluorescence intensities are not relative between examples (for a quantitative assessment, see Figure S3, Supplementary Materials).

Conclusions
Af LEA1 and its protein variants in the animal extremophile A. franciscana are the only known Group 1 LEA proteins in this kingdom of life. Based on our findings, their highly repetitive structure aids in consistent protein folding behaviors during desiccation. The repetitive motifs in Af LEA1 may further act as multivalence sites for protein-protein and protein-RNA interactions. We speculate that this will be similar for LEA proteins from Group 1 found in other organisms. The readily observed phase separation of Af LEA1 in the presence of RNA in vitro strongly suggests a direct interaction between both biomolecules. However, Af LEA1 also improves the desiccation tolerance of proteins, such as LDH, making this protein promiscuous in its target selection. Based on these findings, we speculate Group 1 LEA proteins such as Af LEA1 may offer multiple modes of protection against desiccationinduced damage at different cellular hydration levels to various classes of biomolecules.
The reason(s) for multiple Group 1 LEA proteins that virtually only differ in their number of Group 1 repeats in A. franciscana is intriguing and requires further investigation. Similar Group 1 LEA proteins with different molecular mobilities may be advantageous during desiccation. Larger Group 1 LEA proteins may rapidly increase the viscosity of the cellular compartments at high water contents, thereby preventing protein aggregation and membrane collapse. In contrast, smaller Group 1 LEA proteins have higher mobility and can more readily interact with targets in this highly dense environment. We suspect the observed LLPS of Af LEA1 in the presence of RNA is a feature of other LEA proteins from Group 1, strengthening the notion that this recently observed phenomenon could be a common protective mechanism in the toolbox of anhydrobiotic organisms.
Supplementary Materials: The following supporting information can be downloaded at: https://www. mdpi.com/article/10.3390/biom12030425/s1. Table S1: Both Af LEA1 elution fractions display a tendency to elute a lower sodium concentration with increasing amounts of protein loaded. Figure S1: The elution maxima (retention time) of Af LEA1 tends to correlate with the amount of loaded protein (peak height). Figure S2: The CD spectra of Af LEA1 measured in solution (green), in the desiccated state (blue), and when crowded via 2% v/v SDS (magenta). Figure