Synergism between a simple sugar and a small intrinsically disordered protein mitigate the lethal stresses of severe water loss

Anhydrobiotes are rare microbes, plants and animals that tolerate severe water loss. Understanding the molecular basis for their desiccation tolerance may provide novel insights into stress biology and critical tools for engineering drought-tolerant crops. Using the anhydrobiote, budding yeast, we show that trehalose and Hsp12, a small intrinsically disordered protein (sIDP) of the hydrophilin family, synergize to mitigate completely the inviability caused by the lethal stresses of desiccation. We show that these two molecules help to stabilize the activity and prevent aggregation of model proteins both in vivo and in vitro. We also identify a novel role for Hsp12 as a membrane remodeler, a protective feature not shared by another yeast hydrophilin, suggesting that sIDPs have distinct biological functions.


INTRODUCTION 22
In the near future, global food security will be challenged by the effect of climate change on 23 crop yields (Schwalm et al., 2017;Tirado et al., 2013). Potential solutions to more frequent 24 drought may come from the study of drought-tolerant organisms. Indeed, diverse organisms, 25 collectively named anhydrobiotes, can survive extreme water loss. By identifying and 26 understanding the molecular mechanisms underlying desiccation tolerance, we hope to 27 provide key new insights into stress biology and suggest fruitful avenues for the engineering of 28 drought-tolerant crops. 29 For almost a half a century, scientists have known that almost all anhydrobiotes have high 30 levels of the disaccharide trehalose and hydrophilins, which are small, hydrophilic, and 31 intrinsically disordered proteins (sIDPs) (Crowe et al., 1992;Crowe, 1971;Potts, 2001). 32 Demonstrating a functional significance for this striking correlation eluded researches until very 33 desiccation tolerance, this screen did not inform on potential stress effectors because changes 67 in respiration alters the expression of many hundreds of genes, including many potential stress 68 effectors. Strikingly, this genetic screen failed to identify deletions in non-regulatory genes. The 69 absence of these deletions supported the notion that stress effectors with overlapping 70 functions in desiccation tolerance must exist, precluding their identification by a single gene 71 deletion. . These results showed that trehalose and sIDPs could act as potential 77 stress effectors in the same organism. However, the level of desiccation tolerance in cells with 78 elevated trehalose or the tardigrade sIDPs was still significantly less than the tolerance in 79 stationary-phase cells, further reinforcing the notion that tolerance likely requires multiple 80 effectors working in concert. 81 We sought to identify such effectors and to characterize their molecular functions. Here, 82 using our genetic assays with stationary-phase and exponentially growing yeast cells, we 83 demonstrate that trehalose and Hsp12, a yeast hydrophilin, synergize to completely alleviate 84 all of the lethal damage that occurs due to severe water loss. We provide evidence that these 85 two small molecules counteract the misfolding and aggregation of protein reporters that occurs 86 upon desiccation both in vitro and in vivo. We identify a novel role for Hsp12 as a membrane 87 remodeler, an activity likely required to provide desiccation tolerance. Finally, our data 88 suggests that despite sharing many of the same properties (high glycine content, small size, 89 high hydrophilicity, disordered secondary structure), yeast hydrophilins clearly have distinct 90 functions. We discuss the implications of these findings for both stress biology and potential 91 translational applications. 92 93 RESULTS 94

Trehalose and Hsp12 are necessary and sufficient for desiccation tolerance 95
We reasoned that the previous screen of the non-essential yeast deletion collection for 96 sensitivity to short-term desiccation sensitivity was unable to identify individual stress effectors 97 because likely more than one stress effector needed to be inactivated to compromise tolerance. 98 At least one of these factors was likely trehalose, given its role in long-term desiccation 99 unstressed dividing cells. Therefore, we expressed Hsp12 under a strong constitutive promoter 134 to drive expression in dividing cells to a level similar to stationary cells. These dividing cells 135 exhibited a 1000-fold increase in the desiccation, comparable to the effect of increased 136 intracellular trehalose ( Figure 1B). Thus, Hsp12, like trehalose, had a significant ability to 137 mitigate one or more of the lethal stresses imposed by desiccation. 138 Next, we looked for potential synergy between trehalose and Hsp12 by examining the 139 desiccation tolerance of exponentially dividing yeast cells containing high levels of both 140 protectants. The desiccation tolerance of these cells was ~60-fold greater than cells that 141 expressed only Hsp12 or trehalose. Furthermore, the absolute amount of tolerance (65-80% 142 survival) was even greater than the tolerance of wild type cells in stationary phase (20-40% 143 survival), the highest tolerance previously reported for yeast cells (Figure 1B

Trehalose and Hsp12 modulate proteostasis in vivo and in vitro 149
Having previously established that trehalose helped prevent desiccation-induced 150 proteotoxicity in vivo (Tapia and Koshland, 2014), we tested the in vivo role of Hsp12, with and 151 without trehalose to modulate proteostasis. We first utilized the inactivation of firefly luciferase 152 as a proxy for misfolding and aggregation in vivo. As a control, we expressed luciferase in 153 exponentially dividing wild-type cells that are expressing neither trehalose nor Hsp12. We also 154 expressed luciferase in exponentially dividing cells with high levels of Hsp12, trehalose or both 155 protectants together. After desiccation, luminescence in wild type cells was reduced one 156 hundred thousand-fold ( Figure 2A). Thus, luciferase activity was extremely sensitive to 157 desiccation. Cells with elevated trehalose retained a four-fold increase in luminescence; while 158 cells with Hsp12 alone failed to retain any increase in luminescence after drying. However, the 159 luminescence of cells that contained both elevated trehalose and elevated Hsp12 was 14-fold 160 greater than wild type and 4-fold greater than cells with just elevated trehalose (Figure 2A). 161 These results demonstrate that Hsp12 can enhance the small, but reproducible, ability of 162 trehalose to prevent luciferase inactivation. 163 To complement the luminescence assay, we examined desiccation-induced luciferase 164 aggregation. Lysates were prepared from our panel of desiccated exponentially dividing cells 165 immediately after rehydration. The lysates were subjected to centrifugation to separate soluble 166 luciferase from insoluble luciferase that had aggregated with itself or other insoluble proteins. 167 For all strains tested, equivalent amounts of total luciferase were present in cell lysates (Figure 168 2B). In the wild type strain lacking trehalose and Hsp12, no luciferase was detected in the 169 soluble fraction ( Figure 2B). Increased intracellular trehalose or Hsp12 significantly increased 170 luciferase solubility. Despite increasing solubility, Hsp12 failed to rescue any luciferase activity 171 ( Figure 2A). Thus Hsp12 was more potent at blocking luciferase aggregation than it was at 172 preventing its denaturation. We conclude that desiccation, a naturally-occurring environmental 173 condition, potentiates denaturation and aggregation of luciferase, and these proteotoxic effects 174 can be mitigated in vivo by either trehalose, Hsp12, but more pronouncedly by a combination 175 of the two. propagation was dramatically reduced in the descendants of desiccated tps1∆ cells (Figure 192 2C) as expected from our previous study (Tapia and Koshland, 2014). In contrast, the lack of 193 Hsp12 (hsp12∆) alone did not affect the propagation of [PSI + ] after desiccation more than wild 194 type cells, nor did it exacerbate the loss of [PSI + ] propagation in cells unable to make trehalose 195 (tps1∆ hsp12∆) ( Figure 2C). These results suggest that trehalose is more effective than Hsp12 196 in preventing aggregation of this cytoplasmic prion. 197 [GAR + ], a membrane prion. In tps1∆ cells, the percentage of [GAR + ] cells was ~10-fold lower 199 than wild type after desiccation ( Figure 2D) as expected from our previous study (Tapia and 200 Koshland, 2014). The percentage [GAR + ] cells did not change significantly in hsp12∆ cultures 201 before or after transient desiccation ( Figure 2D). In tps1∆ hsp12∆ cultures, the percentage of 202 [GAR + ] cells was undetectable after desiccation, as expected given that tps1∆ cells were 203 unable to propagate [GAR + ] through desiccation. However, even in the absence of desiccation, 204 the percentage of [GAR + ] cells in tps1∆ hsp12∆ cultures reduced almost 500-fold compared to 205 wild type cells ( Figure 2D). Thus, these two small effectors act synergistically to modulate the 206 proteostasis of the [GAR + ] membrane prion even in the presence of water. This synergism also 207 may occur during desiccation, but was masked by the severe effect of the loss of trehalose 208

alone. 209
To test whether the impact of trehalose and Hsp12 on proteostasis was direct, we 210 examined their effect in vitro on the desiccation-induced aggregation of citrate synthase (CS). 211 We desiccated solutions of CS alone, or with trehalose, Hsp12 or both of the stress effectors 212 together. These samples were rehydrated, and CS proteostasis was assessed by assaying its 213 aggregation and its enzymatic activity. desiccation tolerance and failed to provide any synergistic tolerance with trehalose ( Figure 4A). 265 Additionally, deletion of Stf2 alone (stf2∆), or in combination with a loss of trehalose synthesis 266 (tps1∆ stf2∆) had no effect on the short-term desiccation tolerance of stationary phase yeast, 267 unlike the pronounced increase in desiccation sensitivity displayed by tps1∆ hsp12∆ cells 268 ( Figure 1A and 4A). Unlike Hsp12, Stf2 also did not exhibit any detectable secondary structure 269 in the presence or absence of DMPG ( Figure 4B). Additionally, the vesiculation activity of Stf2 We exploited the conditional desiccation tolerance of yeast to provide important new 280 insights into the stress effectors of desiccation tolerance. First, we demonstrated a dramatic 281 synthetic loss in desiccation tolerance in stationary cells that were deleted for HSP12, a 282 hydrophilin, and TPS1, a biosynthetic enzyme of trehalose. This synthetic sensitivity reveals a 283 function for this hydrophilin in desiccation tolerance that was missed because it was masked 284 by an overlapping function with trehalose. The function of Hsp12 as a stress effector of 285 desiccation was supported further by its ability, when highly expressed, to confer partial 286 desiccation tolerance to exponentially dividing cells. While Hsp12 has previously been reported 287 to mitigate lethality due to heat stress and osmolarity, we were unable to repeat those findings 288 ( Figure S1) (Welker et al., 2010). Furthermore, the previous study used a temperature for heat 289 shock of 58°C, a temperature that budding yeast is unlikely to experience neither in the wild 290 nor fermenting vats (Welker et al., 2010). In contrast, desiccation occurs readily in nature, so 291 the desiccation tolerance conferred by Hsp12 is likely one of its true physiological functions. 292 These results, coupled with our previous demonstration for the importance of a subset of 293 tardigrade hydrophilins in desiccation tolerance, provide compelling evidence for the causal 294 roles of a hydrophilin in desiccation tolerance in most anhydrobiotes. 295 Hsp12's lipid-induced folding, its in vitro membrane remodeling activity, and its in vivo role 296 in desiccation tolerance are not shared with Stf2, another yeast hydrophilin. These features of 297 Hsp12 provide new insights into hydrophilins and stress biology. Like yeast, most organisms 298 have large families of hydrophilin-like genes. The distinct causal role of Hsp12 in desiccation 299 tolerance reveals that the generic properties of hydrophilins of charge, size and intrinsic 300 disorder, are not sufficient to mitigate the stresses of desiccation. Thus, the other hydrophilins 301 like Stf2 must have other unrelated functions. One simple idea is that these other functions 302 mitigate yet to be determined stresses. If so, hydrophilins may be a major new class of diverse 303 stress effectors. 304 A second major insight from this study is the demonstration that trehalose and Hsp12 can 305 synergize to counteract all the major stresses that yeast cells encounter upon severe water 306 loss. Exponentially dividing cells increased their desiccation tolerance 1000-fold when they 307 were engineered to have levels of Hsp12 or trehalose similar to levels in stationary cells. This 308 dramatic but partial desiccation tolerance was increased another 60-fold in exponentially 309 dividing cells expressing both trehalose and Hsp12, surpassing the tolerance seen in 310 stationary phase cells. The fact that this simple sugar and small protein together can confer 311 complete tolerance to sensitive cells speaks to their potency to mitigate major cellular stresses. we propose that, alongside protecting against proteotoxicity, Hsp12 and trehalose cooperate in 345 membrane homeostasis. We propose that desiccation induces membrane damage. This 346 damage is prevented by the intercalation of trehalose into membranes. If damage occurs, it is 347 removed by Hsp12's membrane remodeling activity. These distinct activities explain the 348 synergistic requirement for trehalose and Hsp12 in both short-and long-term desiccation 349 tolerance. Additional genetic and biochemical experiment will be needed to test this model. 350 Regardless of their mode of function, the fact that just trehalose and Hsp12 are sufficient to 351 mitigate the major stresses of desiccation has major implications for engineering 352 desiccation/drought tolerance in other organisms. Water performs so many important functions 353 in biology; it was easy to imagine that the removal of water would generate many stresses that 354 require many stress effectors to combat. The need for a plethora of stress effectors would 355 make transmitting this trait to another organism too complex. The ability to generate 356 desiccation tolerance with only two factors makes this engineering eminently more feasible. 357 Attempts to engineer plants to synthesize more trehalose have been met with technical 358 difficulties because of the sugar's additional roles in metabolism. Our results provide new 359 impetus to overcome those hurdles. Furthermore, Hsp12 has no known metabolism side 360 effects. Given the sufficiency of Hsp12 alone to confer partial desiccation tolerance, it will be 361 very interesting to test whether Hsp12 by itself might confer drought tolerance in plants. NaCl. The protein was then concentrated using a MW3000 filtration unit, aliquoted, frozen in 463 liquid nitrogen, and stored at -80° C until use (Supple Fig. 2). were then placed inside a vacuum desiccator and sealed. Vacuum was then applied to the 499 desiccator, and the samples were left at room temperature under vacuum overnight to remove 500 any remaining chloroform. Then, they were resuspended in 10mM sodium phosphate, pH 7.4 501 and moved to a 1.5 ml plastic microcentrifuge tube. Resuspended aliquots were then vortexed 502 until a homogenous, cloudy white solution was formed. Tubes were placed on ice in a 4° C 503 room, and sonicated using a probe tip sonicator equipped with a blunt tip and set to 20% 504 amplification, 7.5 minutes, with pulses of 2 seconds on and 2 seconds off. Lipids were then 505 spun down at 14,000 rpm at 4° C for 10 minutes to precipitate any metal shards from 506 sonication. Supernatant was then moved to a new 1.5 ml microcentrifuge tube and left on ice 507 until use in circular dichroism. 508

Circular dichroism spectra 509
Circular Dichroism (CD) spectra were obtained using an Aviv 2000 Circular Dichroism 510 Spectrometer Model 410 (Aviv Biomedical Inc.). CD signal was measured from 250 to 190 nm 511 at 25° C, averaging measurements at every nm for 10 seconds. Hydrophilins and DMPG were 512 kept on ice until added to samples. Once added, samples were mixed well by flicking and 513 incubated at room temperature for 10 minutes before spectra were measured. A 1M stock of 514 trehalose and 100mM stock of SDS were prepared in 10mM sodium phosphate, pH 7.4 and 515 added to achieve desired concentrations. Proteins were added to samples to achieve a 516 concentration of 0.32 mg/ml. For samples with DMPG, DMPG was added to achieve a 517 concentration of 1.32 mM (1.2 mg/ml. Samples were made at a volume of 400 µL and 300 µL 518 were pipetted into a 0.1 cm cuvette for CD measurements. The cuvette was washed with water 519 and ethanol between measurements, drying completely each time. Samples with protein were 520 blanked against those with the same buffer and concentration of DMPG without hydrophilins. 521 All spectra were converted to units of mean residue molar ellipticity before plotting on graph. 522

High Throughput Desiccation Tolerance Assay 538
Single colonies of newly constructed double mutants are placed into 200 µL of YEP + 2% 539 Galactose in wells of 96-well plates and allowed to grow to saturation at 30°C with agitation. 540 Strains were then pinned using a 48-density prong onto YEP + Galactose plates and allowed 541 to grow for 2 days as non-desiccated controls. 20 µL of each strain was also transferred into 542 new 96-well plates to let air dry for 6 and 30 days at 23°C. Drying was done as previously 543 mentioned. After desiccation, strains are rehydrated in 200 µL of YEP + Galactose and pinned 544 onto solid media as before. Desiccation tolerance is assayed by comparing growth of strains 545 after desiccation with non-desiccated controls. 546