Role of trehalose phosphate synthase in anoxia tolerance and development in Drosophila melanogaster.

Recent studies have shown that trehalose plays a protective role in yeast in a variety of stresses, including heat, freezing and thawing, dehydration, hyperosmotic shock, and oxidant injury. Because (a) heat shock and anoxia share mechanisms that allow organisms to survive, (b) Drosophila melanogaster is tolerant to anoxia, and (c) trehalose is present in flies and is metabolically active, we asked whether trehalose can protect against anoxic stress. Here we report on a new role of trehalose in anoxia resistance in Drosophila. We first cloned the gene trehalose-6-phosphate synthase (tps1), which synthesizes trehalose, and examined the effect of tps1 overexpression as well as mutation on the resistance of Drosophila to anoxia. Upon induction of tps1, trehalose increased, and this was associated with increased tolerance to anoxia. Furthermore, in vitro experiments showed that trehalose reduced protein aggregation caused by anoxia. Homozygous tps1 mutant (P-element insertion into the third intron of the gene) leads to lethality at an early larval stage, and excision of the P-element rescues totally the phenotype. We conclude that trehalose contributes to anoxia tolerance in flies; this protection is likely to be due to a reduction of protein aggregation.

The ability of organisms to sustain O 2 deprivation is limited. Irreversible injury may occur to mammalian tissues within 5-10 min of severe hypoxia or ischemia. Nervous tissue is particularly vulnerable to hypoxia (1). The range of tolerance of animals to hypoxia is wide, and in contrast to mammalian sensitive tissues, there are certain vertebrate or invertebrate animals that have an amazing resistance to anoxic injury (2). For example, we and others (3)(4)(5) have shown that the brain tissue of the turtle Pseudemys scripta elegans is exceedingly resistant to anoxia, surviving hours of experimental O 2 deprivation with neurons continuing to fire with very little change in their electrophysiologic properties. Another example is the invertebrate Drosophila melanogaster: although its response to O 2 deprivation is different from that of the turtle, the Drosophila also shows an extreme resistance to anoxia as it recovers from hours of anoxia without evidence of cell damage (6,7).
The mechanisms underlying the susceptibility or tolerance to O 2 deprivation are not well understood. Clinical trials using a variety of agents to treat cardiac or brain ischemic injury have not been successful so far (8). Indeed, several questions still need to be addressed: 1) What mechanisms promote anoxia resistance in the turtle or fruit fly? 2) If metabolic energy (ATP) is depleted during anoxia, what mechanisms prevent protein degradation in the flies after hours of anoxia? and 3) Do similar mechanisms operate in other stress conditions, e.g. desiccation, heat, freezing, or oxidant injury? Mechanisms that protect protein integrity must exist and are likely to be operative during anoxia since the post-stress behavior of this organism appears to be normal.
It is well known that there are at least two major types of substances that play an important role in preserving proteins during heat shock: heat shock proteins and organic compounds such as disaccharides. For example, trehalose is a non-reducing disaccharide and is found in diverse organisms such as bacteria, fungus, algae, and insects. It has been shown to play a protective role in yeast in a variety of stress conditions such as heat, freezing and thawing, dehydration, hyperosmotic shock, and oxidant injury (9 -11). Singer and Lindquist (12) have demonstrated that trehalose reduces protein aggregation and maintains proteins in a partially folded state, facilitating their refolding by cellular chaperones at a later stage, possibly after the stress condition is terminated. Though trehalose is absent in mammals, it has been shown to improve the tolerance of human cells to desiccation (13) and tissue tolerance to cryopreservation (14).
The possible involvement of trehalose in cellular protection against anoxic damage has not been studied. Four reasons have led us to hypothesize that trehalose may do so: 1) Anoxia, like heat stress, induces an up-regulation of heat shock proteins; hence, anoxia may share similar types of mechanisms that allow organisms to survive. 2) We and others have shown that flies tolerate long term anoxia. 3) Flies synthesize and catabolize trehalose. and 4) Trehalose has been shown (as indicated above) to protect against heat shock.
The present studies were undertaken to investigate whether trehalose plays an important role in protecting flies against anoxic stress. We have tested whether overexpression or mutation of trehalose-6-phosphate synthase or tps1 1 gene, a gene that synthesizes trehalose in flies, increases or decreases tolerance to anoxia respectively, and if so, whether trehalose can reduce protein aggregation caused by anoxia in D. melanogaster.

EXPERIMENTAL PROCEDURES
D. melanogaster Stocks-Flies were maintained on standard Drosophila medium at temperatures of 24 -25°C or as otherwise specified. The following GAL4 fly lines were used: 32B-GAL4 (3rd) for constitutive general expression, pGMR-GAL4 (2nd) for specific eye tissue expression, and the heat shock inducible GAL4 line, Hs-GAL4 (3rd). Fly stocks with balancer chromosomes used were Adv/CyO and Adv/CyO; Sb/TM6B.
Cloning of Drosophila tps1 cDNA-When yeast tps1 cDNA was used to blast the Drosophila data base, a gene with a 2427-bp open reading frame was found. The Drosophila gene was found to be 30% similar to yeast tps1. Two primers in a conserved domain (sense: 5Ј-cacgactaccatctcatgct-3Ј and antisense: 5Ј-ctccaggttcatgccatcgcg-3Ј) were designed and used. PCR was performed, and its product, a 710-bp segment, was cloned into pCR-TOPO vector (Invitrogen) and used as a probe. A Drosophila cDNA library, which was constructed in our laboratory, was screened using the probe labeled with 32 P (a random primer labeling kit, Stratagene). Library screening yielded three positive clones of about 3000 bp, which were digested with EcoRI, subcloned into pCR2.1, and sequenced.
Confirmation of P-element Location-Genomic DNA was extracted from P838/CyO flies, digested with MspI, self-ligated, and PCR-amplified with primer Pry4: 5Ј-caatcatatcgctgtctcactca-3Ј and Plw3-1: 5Јtgtcggcgtcatcaactcc-3Ј (15). The PCR product was recovered from an agarose gel and cloned into the pCR-TOPO vector and sequenced with primer SpepI: 5Ј-gacactcagaatactattc-3Ј. The resulting sequence was used to blast the Drosophila genomic DNA data base.
Anoxia Test-The procedure for inducing anoxia and measuring recovery time has been described previously in detail (16). Briefly, groups of 10 -12 adult flies, age 3-5 days, were placed in a specially designed chamber and exposed to anoxia (O 2 concentration ϭ 0% with administration of 100% N 2 ) for 5 min prior to recovery in room air (O 2 ϭ 20.8%). After the introduction of N 2 in the chamber, flies became uncoordinated within ϳ30 s (when O 2 concentration was about 1%), resulting in their fall to the bottom of the chamber, where they remained motionless for the rest of the anoxic period. Recovery time was measured as the latency between the end of a 5-min anoxia treatment and the time point when flies recovered, which was usually a very discrete event.
Measurement of Trehalose Concentration in Flies-30 mg of fly tissues or fly heads were homogenized with a plastic pestle in 0.25 ml of 0.25 M Na 2 ϩ CO 3 Ϫ and incubated at 95°C for 2 h to denature proteins. Solutions of 1 M acetic acid (0.15 ml) and 0.2 M sodium acetate (0.6 ml) (pH 5.2) were mixed with the homogenates. The mixture was centrifuged at room temperature for 10 min at 12,000 rpm. Aliquots (200 l) of the supernatant were placed into two tubes; one was used as background, and the other was digested by trehalase (0.05 units/ml, 37°C overnight). Glucose concentration was measured using Infinity TM glucose reagent (Sigma) (17).
NMR Measurement of Perchloric Acid-extracted Flies-To determine whether trehalose is in a metabolically active pool in the fly, we resorted to NMR measurement of trehalose after [1-13 C]glucose feeding. Tissues from frozen fly heads (ϳ400 mg) were extracted with 3 M perchloric acid (3:1 v/w) and centrifuged, and the supernatant was neutralized with 3 M KOH. Following centrifugation to remove perchlorate salts, the supernatant was lyophilized, and the powder was dissolved in a solution of deuterium oxide containing 50 mM potassium phosphate (pH 7) and 3,3-trimethyl-sily phosphate (TSP) as a chemical shift reference. NMR spectra were acquired at 8.4 tesla using an AM360 NMR Spectrometer (Bruker Instruments, Bellerica, MA). Isotope enrichments were measured in the 1 H NMR spectrum using the proton-observed, carbon 13 -edited NMR pulse sequence (18). Briefly, two subspectra are acquired in an interleaved fashion: one consists of all proton resonances bound to 12 C or 13 C (i.e. 12 C ϩ 13 C), whereas the other is acquired with inversion of protons bonded to 13 C (i.e. 12 C Ϫ 13 C). Digital subtraction of these two subspectra yield only those proton resonances bonded to 13 C (i.e. 2 ϫ 13 C). Each subspectrum was acquired with 13 C decoupling to simplify the spectrum. 1 H spectra were acquired as the summation of 256 transients, 16,000 data points, and fully relaxed (repetition time/ transient of 21 s). A low power presaturation pulse was used to suppress the residual HDO resonance. Free-induction decays were processed using an exponential filter (0.5 Hz), zero-filled to 32,000 data points, and Fourier-transformed. The fractional 13 C enrichment of the trehalose H1␣ at 5.2 ppm was determined by integration of the resonances in the proton-observed, carbon 13 -edited subspectra and expressed as the ratio of 13 C-H1␣/[( 13 C ϩ 12 C)]-H1␣).
Isolation of Aggregated Protein-Aggregate proteins were isolated by differential centrifugation. The Triton-insoluble fraction was twice resuspended in extraction buffer, sonicated, and pelleted at 17,000 ϫ g for 30 min at 4°C. The resultant pellet was again resuspended in extraction buffer, sonicated, and pelleted at 5,000 ϫ g for 30 min at 4°C. The purified pellet, consisting of aggregated proteins, was resuspended in extraction buffer and stored at Ϫ70°C (19).
Western Blot-Protein determinations were carried out in duplicate in each sample with the bicinchoninic acid protein assay kit (Sigma) using bovine serum albumin as a standard. Protein samples (5 g) were electrophoresed through 4 -10% Novex bis-Tris denaturing gel (Invitrogen) and electrophoretically transferred to nitrocellulose membrane. Nonspecific binding sites were blocked, and the membranes were incubated for 1 h with primary antibodies (anti-␤ tubulin and anti-Na ϩ /K ϩ ATPase, Jackson ImmunoResearch Laboratories, Inc.) and detected using horseradish peroxidase-conjugated secondary antibody with enhanced chemiluminescence (ECL; Amersham Biosciences, Inc.).

Trehalose Is Metabolically Active in Wild Type Flies-To
assess whether trehalose is metabolically active in wild type flies, trehalose 13 C isotopic labeling was assessed in 13 C-edited, 1 H NMR spectra of acid extracts of overnight fasted flies following a 2-h exposure to a solution of [1-13 C]glucose (Fig. 1). The H1␣ doublet resonance of trehalose was visible at 5.2 ppm, whereas no resonance was detected for H1␤, consistent with the assignment of trehalose and indicating that free glucose ( 13 C-labeled or unlabeled) was not observed. Substantial 13 C labeling of trehalose C1␣ was observed (40.4%). Thus, the results indicate that wild type flies have significant levels of trehalose and that the trehalose pool is metabolically active.
Cloning of Drosophila tps1 cDNA-A total of 200,000 plaqueforming units from a primary fly head cDNA library were screened with the probe described above. Four positive clones were isolated, subcloned, and sequenced. Of these clones, three gave full-length cDNA of the tps1 gene. This gene encodes a FIG. 1. 13 C isotopic labeling of trehalose from [1-13 C]glucose in wild type fly heads. Upper spectrum, total intensity ( 12 C ϩ 13 C) trehalose-H1␣ at 5.2 ppm. Lower spectrum, the intensity of trehalose-H1␣ bonded to 13 C only. Spectra were acquired using 1 H-observed, 13 C-edited NMR as described in "Experimental Procedures." Only the region from 5.34 to 5.06 ppm is shown.
protein of 809 amino acids and contains both conserved domains of yeast tps1 and tps2 (Fig. 2). Using Flybase and Gadfly, we found that this gene is located on chromosome 2L at 24F01.
Overexpression of tps1 in Flies-To study the function of tps1, we overexpressed the tps1 gene using the UAS-GAL4 system (15). We first cloned tps1 cDNA into pUAST vector and generated a number of transgenic lines carrying the UAS-tps1 transgene on each of the three main chromosomes of the fly. When each of these specific transgenic lines were crossed with 32B-GAL4 for constitutive generalized overexpression (UAS-tps1/UAS-tps1 ϫ 32B-GAL4/32B-GAL), all larvae died at the stage of the first instar. Specific tissue overexpression of tps1, using pGMR-GAL4, resulted in a rough eye phenotype (Fig.  3C). The ommatidia were smaller and disorganized. However, the number of receptor cells in each ommatidia was normal with larger spaces present between ommatidia (Fig. 3D). When we overexpressed tps1 with the heat shock-inducible promoter Hs-GAL4 at either 24°or 18°C, flies died at a late pupa stage (Fig. 3E); however, animals survived and developed into the adult stage when raised at 15°C.
Overexpression of tps1 in Adult Flies: Trehalose Levels and Tolerance to Anoxia-Adult flies, grown at 15°C, were used. After hatching, these flies were first transferred to and kept at 18°C for 2 consecutive days before they were tested for their tolerance to anoxia. We used the same phenotypic assay we had validated previously and used extensively (16). We tested flies with tps1 overexpression and controls. Three kinds of controls were used: w 1118 , w 1118 ; UAS-tps1/ϩ and w 1118 ; Hs-GAL4/ϩ, which were also grown at 15°C. As in our previous studies, an FIG . 3. Overexpression of tps1 in flies. A, wild-type eye; B, wildtype eye section (ϫ160); C, tps1-pGMR eye; D, tps1-pGMR eye section (ϫ160). Compared with wild-type eye, the number of receptor cells in tps1-pGMR ommatidia is the same, but the size of the ommatidia is smaller, and larger spaces are present between ommatidia. E, overexpression of tps1 with Hs-GAL4 at 24°and 18°C causes lethality at late pupa stage. anoxia period of 5 min was used to determine the recovery latency and tolerance to O 2 deprivation. Overexpression of tps1 in adult flies reduced the recovery time and rendered flies even more tolerant to anoxia than the wild type flies. The recovery times from anoxia were all significantly longer in the controls than in the transgenic flies expressing tps1 and Hs-GAL4 (Fig.  4A). Parallel to the shortened recovery times in the transgenic flies were the trehalose levels, which were much higher in those flies containing Hs-GAL4 and tps1 than in the control flies (Fig. 4B).

FIG. 2. Alignment of
Mutation of tps1 in Flies: Lethality in the First Instar Larval Stage-To further our understanding of tps1 function, we needed to analyze the tps1 mutant. We used a homozygous lethal line, P838/CyO, with a P-element in the tps1 gene from the fly stock center. The location of the P-element was in the 3rd intron of the tps1 gene, and this was confirmed as detailed under "Experimental Procedures." In order to ascertain that the lethality is caused by the P-element insertion, we first crossed the P838/CyO to two other separate lines with deficiencies. The first deficiency was a deletion between cytological locations 24F04 and 25A01-04, and the second was a deletion between 24C03-05 and 25A02-03. We found that there was complementation with the first deletion but not with the second. In addition, excision of the P-element by jumping the P838 P-element restores the viability of the fly, indicating that lethality is due to P-element insertion into the tps1 gene. The lethality occurred in the first instar larval stage because about 1 ⁄4 of the larvae died in this stage when virgin females P838/ CyO were crossed to a male P838/Adv.

Long Term Anoxia and Protein Aggregation in Flies: Can Trehalose Reduce Aggregation or Prevent Na ϩ -K ϩ ATPase from
Aggregating?-The data presented above showed that overexpression of tps1 in flies almost doubled the trehalose level and increased the tolerance to anoxia. To further explore the mechanisms underlying the effect of trehalose, we investigated 1) whether anoxia causes protein aggregation in flies and 2) whether trehalose helps to reduce protein aggregation caused by anoxia. Wild type flies (w 1118 ) were treated with N 2 for 0, 1, 3, 4 h, put on dry ice immediately, and stored in Ϫ80°C. Of 18 mg of total protein extracted from flies, we found increased amounts of aggregated proteins in flies after 3 h but especially after 4 h of anoxia (mean Ϯ S.E.: 1931 Ϯ 96 g) as compared with flies that were not exposed to anoxia (415 Ϯ 21 g, Fig.  5A). Protein aggregation in samples from flies subjected to 4 h of anoxia was reduced with increasing trehalose concentration (Fig. 5B). Interestingly, protein aggregation was also decreased with other disaccharides, such as 0.5 M sucrose (765 Ϯ 41 g), although not to the same degree as with trehalose (470 Ϯ 36 g). However, glycerol at a similar concentration did not seem to have any effect on the solubility of proteins.

FIG. 5. Anoxia increases the aggregate protein subfraction and trehalose reduces this protein aggregation in vitro.
Flies were subjected to 0, 1, 3, or 4 h of anoxia (A). Aggregated proteins were extracted from 18 mg of total proteins. The aggregate subfraction increases with the time of anoxia, especially after 4 h (means Ϯ S.E.). **, p Ͻ 0.002; ***, p Ͻ 0.0001. In 18 mg of total proteins from flies subjected to 4 h of anoxia, the addition of trehalose to protein extraction buffer reduces aggregate subfraction (means Ϯ S.E.) in a dose-dependent manner (B). **, p Ͻ 0.002; ***, p Ͻ 0.0001 subjected to 4 h of anoxia. Western blots of the total aggregated protein fraction revealed much less aggregated Na ϩ /K ϩ ATPase in 0.5 M trehalose than in buffer containing no trehalose (Fig. 6).

DISCUSSION
In this study, using Drosophila as a genetic model, we have made four novel observations regarding the importance of trehalose and the tps1 gene in the tolerance to O 2 deprivation. First, the fly has only one tps gene (tps1), and this gene has domains that are conserved when compared with both yeast genes, namely tps1 and tps2. Second, overexpression of fly tps1 almost doubled trehalose levels and reduced recovery time from anoxia, enhancing their tolerance to anoxia. Third, eyespecific overexpression of tps1 led to gross eye abnormalities, and mutation in tps1 (P-element insertion) resulted in lethality at the first instar larval stage. Fourth, although there was little effect on protein aggregation in Drosophila when anoxia exposure was less than 3 h, there was substantial aggregation at 4 h; trehalose decreased this aggregation in vitro considerably in a dose-response fashion.
Unicellular and multicellular organisms are endowed with mechanisms that allow them to cope with environmental stresses, especially when the stress is not too severe. The best example, which has been studied for more than a decade, is the response to heat shock, which induces the production of specialized heat shock proteins (e.g. heat shock proteins and heat shock transcription factors) that are important for cellular survival. However, the heat shock response involves more than the induction of heat shock proteins. For example, there is evidence that organic compounds and solutes may take part in this response. Sucrose and maltose have been known to protect tissues and cells against temperature swings (20). Trehalose, a glucose dimer, is known to protect yeast during heat, freezing, dehydration, and oxidant injury (9,10). In addition, trehalose is used as a cryoprotectant of human lymphocytes, red blood cells, and lung slices (14). Recently, it was shown that trehalose increases the tolerance in human fibroblast cells in culture against desiccation (13). Because trehalose seems to protect yeast and human cells, it is well known now that there is considerable conservation of genetic pathways from plant to yeast to man, and hypoxia and heat are two environmental stresses to which organisms respond often in similar ways, we elected to ask whether trehalose, which is present in a metabolically active pool in Drosophila as we show in our NMR work, also protects flies from anoxic stress. Our results demonstrate that this is the case, and the evidence is based on several observations. First, when trehalose was increased in adult flies, the recovery from anoxia was enhanced. These are the first data showing that in a whole organism, an increase in trehalose is beneficial and can protect against anoxia. Second, trehalose markedly reduced protein aggregation seen after 4 h of anoxia. In this regard, we tested whether this reduction in protein aggregation could be seen with some important proteins that have been shown in mammalian tissue to be protected during anoxic stress (19). Indeed, we found that the aggregation of the Na ϩ /K ϩ ATPase protein decreases in a major way with increasing amounts of trehalose when trehalose was added to tissues taken from flies exposed to 4 h of anoxia. Third, although not necessarily related to anoxic stress directly, when the gene that encodes for trehalose synthesis was mutated with a P-element insertion, flies did not develop, and they died early in the larval stage. This indicates that tps1/trehalose are important as cell protectants in early life such as during the process of burrowing into food when larvae are constantly or intermittently exposed to anoxia. Alternatively, it is possible that the lack of tps1/trehalose is deleterious in early fly development because of their involvement in functions that are not related to protection from anoxia.
Although we show in this work that trehalose prevents anoxia-induced aggregation, this does not argue against a similar function of other disaccharides or small molecules such as glycerol. Indeed, we have shown that sucrose does help against aggregation but to a lesser degree. Interestingly, glycerol did not help in this activity.
Another important issue is the concentration of trehalose. The overall average concentration of trehalose that we have measured in flies is much less than 500 mM, but this does not imply that concentrations of that high magnitude cannot be reached in specific locations of the cytosol. Major intracytosolic gradients have been shown to exist, such as for ATP and Ca 2ϩ concentrations.
How does trehalose protect cells? Is it possible that trehalose is used as another carbon source that enhances the metabolic machinery in cells? We doubt that trehalose acts mainly by providing an alternative energy source for at least three reasons. (i) Studies from yeast have shown that trehalose is not produced when nutrients are abundant but that it is produced even in the absence of glucose in the medium and when glycogen is metabolized (21). Furthermore, trehalose is not used until all glycogen stores are depleted and yeast is close to cell death (22). (ii) Trehalose has been found to suppress the aggregation of denatured proteins and maintain them in a semifolded state through hydrogen bond stabilization promoting refolding and reattainment of bio-activity (23,24). This stabilizing effect of trehalose is accounted for by the change in the transfer of free energy on unfolding: the interaction between trehalose and proteins is less favorable with the denatured than with the native protein (25). Whether trehalose protects proteins by interacting with these proteins or whether there are intermediary components is not clear from this work. However, it is possible that trehalose "prepares" proteins for the chaperones to act upon and rescue them (26). (iii) Our results show that protein aggregation, including the Na ϩ /K ϩ ATPase, is reduced considerably and in a dose-response fashion when trehalose is added with these proteins. Thus, the results from the literature and our current study would strongly suggest that this disaccharide has a function that can help in the refolding of proteins when cells and tissues are exposed to stresses such as O 2 deprivation, although we cannot rule out totally the idea that trehalose may contribute metabolically as a carbon source in extraordinary circumstances of energy depletion.
If trehalose can assist in refolding proteins and maintenance of protein integrity during stress, why is it that overexpression of tps1 and trehalose at a relatively high temperature (at 18°a nd 24°C) leads to lethality, and why is it that trehalose leads to grossly abnormal eye development when overexpressed with Flies were subjected to 4 h of anoxia, and 5 g of aggregated proteins were prepared by adding 0, 0.3, and 0.5 M trehalose. Although the same amount of Na ϩ /K ϩ ATPase seemed to be present in total proteins, much less Na ϩ /K ϩ ATPase was present in the aggregated proteins with 0.5 M trehalose as compared with those lacking trehalose. Note that ␤-tubulin did not change in each of T, S, or A proteins as a function of trehalose. pGMR? The answers are not totally apparent at present but may lie in the molecular interactions between trehalose, heat shock proteins, and proteins that are being refolded. Normally, trehalose suppresses the aggregation of protein and prevents native protein from unfolding (12), thus promoting interactions of the target protein with heat shock proteins to enhance proper refolding to the native state. If trehalose is present in large quantities during stress (or in the absence of stress), it is believed that it would interfere with refolding of the targeted proteins by heat shock proteins. Whether this is the reason for larval lethality or the grossly abnormal eyes with higher expression of trehalose is not known, but this is a possible explanation. It is interesting to note that survival until first instar larva or late pupae stage depends on the level of overexpression. With the 32B-GAL4 promoter at 24°C, early larvae died; with Hs-GAL4 at 18°and 24°C, late pupae did hatch but died shortly after (Fig. 4E). It was only when the temperature was rather low during development (15°C) that we could obtain adult flies and test them.
In summary, we have shown that constitutive overexpression of tps1 enhances anoxia tolerance. Furthermore, tps1 and trehalose seem to play a critical role during development, and a mutation in the gene leads to early larval lethality. We hypothesize that enhanced anoxia tolerance shown following tps1 and trehalose overexpression is due to a reduction of protein aggregation and the likelihood of protein denaturation during O 2 deprivation.