Enzymatic control of anhydrobiosis-related accumulation of trehalose in the sleeping chironomid, Polypedilum vanderplanki

Larvae of an anhydrobiotic insect, Polypedilum vanderplanki, accumulate very large amounts of trehalose as a compatible solute on desiccation, but the molecular mechanisms underlying this accumulation are unclear. We therefore isolated the genes coding for trehalose metabolism enzymes, i.e. trehalose-6-phosphate synthase (TPS) and trehalose-6-phosphate phosphatase (TPP) for the synthesis step, and trehalase (TREH) for the degradation step. Although computational prediction indicated that the alternative splicing variants (PvTpsα/β) obtained encoded probable functional motifs consisting of a typical consensus domain of TPS and a conserved sequence of TPP, PvTpsα did not exert activity as TPP, but only as TPS. Instead, a distinct gene (PvTpp) obtained expressed TPP activity. Previous reports have suggested that insect TPS is, exceptionally, a bifunctional enzyme governing both TPS and TPP. In this article, we propose that TPS and TPP activities in insects can be attributed to discrete genes. The translated product of the TREH ortholog (PvTreh) certainly degraded trehalose to glucose. Trehalose was synthesized abundantly, consistent with increased activities of TPS and TPP and suppressed TREH activity. These results show that trehalose accumulation observed during anhydrobiosis induction in desiccating larvae can be attributed to the activation of the trehalose synthetic pathway and to the depression of trehalose hydrolysis.

Larvae of an anhydrobiotic insect, Polypedilum vanderplanki, accumulate very large amounts of trehalose as a compatible solute on desiccation, but the molecular mechanisms underlying this accumulation are unclear. We therefore isolated the genes coding for trehalose metabolism enzymes, i.e. trehalose-6-phosphate synthase (TPS) and trehalose-6-phosphate phosphatase (TPP) for the synthesis step, and trehalase (TREH) for the degradation step. Although computational prediction indicated that the alternative splicing variants (PvTpsa ⁄ b) obtained encoded probable functional motifs consisting of a typical consensus domain of TPS and a conserved sequence of TPP, PvTpsa did not exert activity as TPP, but only as TPS. Instead, a distinct gene (PvTpp) obtained expressed TPP activity. Previous reports have suggested that insect TPS is, exceptionally, a bifunctional enzyme governing both TPS and TPP. In this article, we propose that TPS and TPP activities in insects can be attributed to discrete genes. The translated product of the TREH ortholog (PvTreh) certainly degraded trehalose to glucose. Trehalose was synthesized abundantly, consistent with increased activities of TPS and TPP and suppressed TREH activity. These results show that trehalose accumulation observed during anhydrobiosis induction in desiccating larvae can be attributed to the activation of the trehalose synthetic pathway and to the depression of trehalose hydrolysis. bacteria, fungi, plants and invertebrates, are known to accumulate a nonreducing sugar, such as trehalose or sucrose, at high concentrations prior to or on desiccation [3,4], although several tardigrades, including Milnesium tardigradum, and bdelloid rotifers, including Philodina roseola and Adineta vaga, can enter anhydrobiosis without trehalose or trehalose accumulation [5,6]. Trehalose, the focus of this paper, is thought to effectively protect organisms from severe desiccation stress owing to its ability for water replacement and vitrification [3,4,7]. In P. vanderplanki, as larvae are undergoing desiccation, a large amount of trehalose is produced in the fat body cells [8] and redistributed to other cells and tissues through a facilitated trehalose transporter, TRET1 [9]. The transported trehalose has been shown to vitrify in the completely desiccated insects [7]. Thus, the mechanisms underlying the diffusion of accumulated trehalose over the entire insect body, and the protective effect of trehalose on cell components, have been established. Nevertheless, the molecular mechanisms involved in trehalose accumulation in P. vanderplanki remain obscure.
In addition to its role as an anhydroprotectant, trehalose is generally known as a carbon and energy source for bacteria and yeast [10]. In bacteria and yeast, trehalose is synthesized from glucose-6-phosphate and UDPglucose, catalyzed by trehalose-6-phosphate synthase (TPS; EC 2.4.1.15) and trehalose-6-phosphate phosphatase (TPP; EC 3.1.3.12), and the relevant genes have been cloned and well characterized (Fig. 1A). This synthetic pathway is considered to be conserved in a wide range of taxa, including unicellular and multicellular organisms, because these genes have been found in algae, fungi, plants and invertebrates [11].
In numerous insect species, trehalose is the main hemolymph sugar, although many exceptions, including dipteran, hymenopteran and lepidopteran species, have been reported to contain both trehalose and glucose and even to completely lack trehalose, depending on the physiological conditions [12,13]. Trehalose is synthesized predominantly in the fat body, and then released into the hemolymph. After uptake by trehalose-utilizing cells and tissues, trehalose is hydrolyzed to glucose by trehalase (TREH; EC 3.2.1.28). To date, TREH has been studied extensively in many insect species because of its role as the enzyme responsible for the rate-limiting step in trehalose catabolism in eukaryotes [12]. In Bombyx mori, Tenebrio molitor, Pimpla hypochondriaca, Apis mellifera, Spodoptera exsigua and Omphisa fuscidentalis, TREH genes have been cloned and demonstrated to be implicated in certain physiological events [12,[14][15][16][17][18]. Several biochemical studies on insect TPS and TPP have been reported [12], but these are markedly less complete relative to those on TREH. Tps genes have been reported in many invertebrate species, including a model nematode, Caenorhabditis elegans, an anhydrobiotic nematode, Aphelenchus avenae, a crustacean, Callinectes sapidus, and insects, Drosophila melanogaster, Helicoverpa armigera and Spodoptera exigua [19][20][21][22][23]. Furthermore, insect genome projects have shown that Tps gene sequences are found in Apis mellifera, Tribolium castaneum, Locusta migratoria, Anopheles gambiae and Culex pipiens. Among the insect genes, Drosophila tps1 (dtps1) and Helicoverpa Tps (Har-Tps) are expressed heterologously, and TPS activity has been confirmed in the resultant proteins [21,22]. Furthermore, the effects of overexpression of dtps1 on trehalose levels in relation to anoxia tolerance [21], and the involvement of Har-Tps in diapause induction [22], have been reported. No information on the insect Tpp gene has been obtained, but, instead, it has been suggested that Filled circles and open circles represent glycogen and trehalose content, respectively; the broken line represents the amount of total carbohydrate. G-1-P, glucose-1-phosphate; G-6-P, glucose-6-phosphate; Glc, glucose; PGM, phosphoglucomutase; UDPGP, UDP-glucose pyrophosphorylase; Pi, inorganic phosphate; PPi, pyrophosphate; T-6-P, trehalose-6-phosphate.
DTPS1 and Har-TPS may act not only as TPS, but also as TPP [21][22][23]. The basis for this suggestion is that TPSs comprise both the Glyco_transf_20 (GT-20) motif responsible for trehalose-6-phosphate synthesis, and the trehalose_PPase (TrePP) motif, according to motif analysis on the Pfam (protein family) database (http://pfam.sanger.ac.uk/). However, on balance, the regulation of trehalose metabolism in insects has not been studied comprehensively. Thus, the elucidation of how enzymes control the rapid accumulation of trehalose in response to desiccation stress should provide important information for understanding the molecular mechanism of anhydrobiosis induction in P. vanderplanki as well as fundamental insect physiology. In this study, we identified the genes involved in trehalose metabolism and analyzed their expression and the functions of the gene products.

Results
Changes in trehalose and glycogen contents in P. vanderplanki during desiccation In insects, glycogen is the major substrate for trehalose synthesis [12,13,24]. During desiccation in P. vanderplanki, changes in trehalose and glycogen contents were correlated, i.e. the conversion of glycogen into treha-lose (Fig. 1B). As the sum of trehalose and glycogen was fairly constant, the fluctuations in trehalose and glycogen contents during desiccation indicate that trehalose is likely to be synthesized from glucose-6phosphate and UDP-glucose originating from the glycogen stored in fat body cells.
Changes in the activities of trehalose metabolism enzymes in P. vanderplanki during desiccation The activities of the enzymes involved in trehalose metabolism were investigated during the desiccation of P. vanderplanki. As desiccation progressed, the activities of TPS and TPP were enhanced prior to and parallel with trehalose accumulation, respectively, whereas TREH activity decreased ( Fig. 2B-D). Glycogen phosphorylase (GP) activity is generally controlled not only by gene expression, but also by reversible phosphorylation. Thus, GPb (inactive form) is reversibly converted into GPa (active form) by phosphorylation. In the results of GP assays, the GPa activity and total activity originating from both forms of GP protein were constant throughout the desiccation process ( Fig. 2A).
These results indicate that changes in the activity of TPS, TPP and TREH, rather than GP, are responsible for the accumulation of trehalose originating from glycogen. Cloning of PvTpsa ⁄ b, PvTpp and PvTreh cDNA To elucidate the molecular mechanisms of the enhancement of the trehalose biosynthetic activity during desiccation in P. vanderplanki, we cloned the genes for TPS, TPP and TREH. Full-length cDNAs of PvTps and PvTreh were isolated by RT-PCR and ⁄ or 5¢-and 3¢-RACE. For the isolation of cDNAs, degenerated primer sets were designed on the basis of the nucleotide sequences of Tps and Treh cDNAs that have been reported previously in many organisms [12,[25][26][27][28][29][30][31][32]. After cDNA fragments corresponding to each gene had been obtained, 5¢-and 3¢-RACE were performed. Information on the nucleotide sequence of PvTpp was obtained by screening in an expressed sequence tag (EST) database constructed with sequences of cDNAs prepared from desiccating larvae [33], and the full-length cDNA was determined by 5¢-RACE.
As a result of 3¢-RACE on PvTps, we isolated two distinct mRNAs, named PvTpsa and PvTpsb, that were different at each 3¢-end of the nucleotide sequence. PvTpsa cDNA consisted of 3026 bp (Fig. 3A). Because nucleotides (nt) 69-71 represent a stop codon (TAA), the downstream nt 90-92 were regarded as the initiation codon (ATG). nt 2628-2630 also represented a stop codon (TGA), thus suggesting a 2538-bp ORF (846 amino acids with a molecular mass of 95 300). PvTpsb cDNA consisted of 3094 bp; 68 nucleotides were inserted between nt 2291 and 2292 of PvTpsa. Because a frame shift occurred by insertion, the ORF in PvTpsb was shortened to 2373 bp, encoding 791 amino acids with a calculated molecular mass of 89 500 (Fig. 3A). The genomic DNA sequence of the PvTps gene confirmed that PvTpsa and PvTpsb were generated by alternative splicing (Fig. 3A). In the same manner, cDNAs of PvTpp and PvTreh were defined to consist of 1044 bp, including an 882-bp ORF (294 amino acids with a molecular mass of 33 400), and 2177 bp, including a 1734-bp ORF (578 amino acids with a molecular mass of 66 400), respectively ( Fig. 3B, C).
The deduced amino acid sequences of PvTPSa ⁄ b, PvTPP and PvTREH were subjected to Pfam search. PvTPSa and PvTPSb have both the GT-20 and TrePP motifs, whereas PvTPP has the TrePP motif only (Fig. 3A, B). The GT-20 motif, belonging to the glycosyl transferase family 20, is found in every TPS and several TPP proteins, and the TrePP motif is found in several TPSs and every TPP protein [32]. In PvTREH, we found TREH signature 1, TREH signature 2 and a glycine-rich region, which are the consensus sequences of the TREH protein (Fig. 3C). Thus, PvTpsa ⁄ b, PvTpp and PvTreh seemed to encode TPS, TPP and TREH, respectively, of P. vanderplanki.

Functional analysis of PvTpsa/b, PvTpp and PvTreh
To corroborate whether these genes encode functional proteins, recombinant proteins were prepared using an in vitro transcription and translation system (TnT, Promega, Madison, WI). First, we checked that protein synthesis was successful via SDS ⁄ PAGE and western blot analysis (Fig. 4A). The expression of PvTPP protein was very faint. The coexistence of both PvTpsa and PvTpsb cDNAs with PvTpp cDNA in the TnT reaction mixture was successful for the expression of these proteins, although the expression levels were slightly lower. In the TPS assay, PvTPSa and PvTPSb showed no activity; trehalose-6-phosphate was not produced from glucose-6-phosphate and UDP-glucose (data not shown). TPS activity was also not detected when PvTPSb and PvTPP were present with PvTPSa. In the TPP assay with PvTPP only, or mixed with PvTPSa and PvTPSb, catalyzed dephosphorylation of trehalose-6-phosphate into trehalose occurred (Fig. 4B). As neither PvTPSa nor PvTPSb (or both) was able to dephosphorylate trehalose-6-phosphate, we conclude that PvTPP is responsible for dephosphoryla-    tion. The incubation of PvTREH with trehalose resulted in the production of glucose, indicating that PvTREH functions as TREH by hydrolysis of the a-1,1-glycosidic bond in trehalose (Fig. 4C, D). TPS activity was not detected in the recombinant PvTPSa or PvTPSb in vitro. Genetic techniques using yeast deletion mutants are also a powerful tool for the functional analysis of TPS [34][35][36]. In order to confirm the function of PvTPSa and PvTPSb, we employed yeast tps1 deletion mutants. The yeast deletion mutant of TPS1 (tps1D), lacking the TPS1 gene corresponding to TPS, was transformed with the PvTpsa or PvTpsb expression vector. These transformants were examined for their ability to synthesize trehalose. The tps1D + PvTpsa strain, but not the tps1D + PvTpsb strain, accumulated trehalose comparably to the wild-type (Fig. 4E). We checked the expression of the PvTPSa and PvTPSb proteins in each transformant, and found that PvTPSa was successfully expressed, but that PvTPSb was not (Fig. 4E). From these results, the catalytic activity of the PvTPSa protein was demonstrated, although the function of PvTPSb as an enzyme was not shown.
Complementation of the yeast tps1 or tps2 deletion mutant phenotype by the corresponding PvTpsa or PvTpp gene The yeast deletion mutant tps1D has been reported to be osmosensitive [34][35][36]. In the tps2D strain, the yeast deletion mutant lacking the TPS2 gene corresponding to TPP, thermosensitivity to high temperature was reported [37,38]. Thus, we examined whether PvTpsa ⁄ b in tps1D and PvTpp in tps2D rescued the deletion mutants from osmosensitivity and thermosensitivity, respectively (Fig. 5). The tps1D + PvTpsa strain grew at the same level as the wild-type on hypertonic medium containing 1 m NaCl, 50% sucrose or 1.5 m sorbitol (Fig. 5A). However, the tps1D + PvTpsb strain showed little improvement in growth rate compared with the tps1D strain on 1 m NaCl and 50% sucrose plates (Fig. 5A); these results are consistent with the absence of PvTPSb expression (Fig. 4E). Nevertheless, tps1D + PvTpsb on 1.5 m sorbitol plates showed slightly lower growth than the tps1D + PvTpsa strain (Fig. 5A). At present, we have no adequate explanation for this modest rescue; it may be caused by a kind of side-effect of transformation or the presence of trace amounts of the PvTPSb protein.
Thermosensitivity in the tps2D + PvTpp strain was rescued to almost the same level as the wild-type (Fig. 5B). These results clearly demonstrate that PvTpsa and PvTpp function genetically as Tps and Tpp, respectively.

Expression profiles of PvTpsa/b, PvTpp and PvTreh mRNAs and proteins
As shown in Fig. 1B, in P. vanderplanki, trehalose is likely to be synthesized from glycogen en route to anhydrobiosis. In eukaryotes, the metabolic pathway from glycogen to trehalose is highly conserved ( Fig. 1A). Hence, in order to elucidate the molecular mechanisms underlying the regulation of the enzymes involved in trehalose metabolism on desiccation, we first investigated the expression profiles of PvTpsa ⁄ b, PvTpp and PvTreh mRNAs (Fig. 6A). The accumulation of PvTpsa ⁄ b and PvTpp mRNAs was induced within 1 h and 3 h, respectively, during desiccation treatment. For PvTreh, the induction of mRNA accumulation was delayed by 48 h after the beginning of desiccation treatment compared with the other two genes. Real-time PCR analyses of these mRNAs confirmed the results (data not shown). However, the amount of PvGp mRNAs remained constant during treatment, which is consistent with the constancy of GP activity on desiccation ( Fig. 2A). Western blot analyses revealed that the proteins of PvTPSa ⁄ b, PvTPP and PvTREH were also accumulated, as were the corresponding mRNAs (Fig. 6B).

Discussion
In this study, we have isolated and characterized three desiccation-inducible genes, PvTpsa ⁄ b, PvTpp and PvTreh, encoding the enzymes involved in trehalose metabolism in P. vanderplanki (Fig. 3). In addition to P. vanderplanki, many anhydrobiotes, such as A. avenae, and Artemia cysts accumulate trehalose as they undergo desiccation. In these organisms, trehalose accumulation correlates significantly with anhydrobiosis induction [3,4,39]. In contrast, several rotifers and tardigrades enter anhydrobiosis without trehalose accumulation, but possess other anhydroprotectants, such as late embryogenesis abundant proteins [4,6].
The induction of trehalose synthesis is necessary for P. vanderplanki to achieve anhydrobiosis. The larvae, if rapidly dehydrated, cannot enter anhydrobiosis because of an insufficient amount of trehalose [40,41]. Furthermore, it has been hypothesized that trehalose is replaced with water or can vitrify to exert its protective function against dehydration [3,4,7]. Indeed, trehalose is produced in fat body cells in desiccating P. vanderplanki larvae [8], redistributed to other cells and tissues through a facilitated trehalose transporter, TRET1 [9], and vitrified in completely desiccated insects [7]. Thus, the successful induction of anhydrobiosis in P. vanderplanki must occur via a sequence of events: expression of trehalose metabolism-related genes, de novo synthesis and accumulation of trehalose, redistribution and vitrification.
PvTpsa rescued the growth of the yeast tps1D mutant, and PvTpp rescued the growth of the tps2D mutant, providing evidence that PvTpsa and PvTpp encode genetically functional trehalose synthases (Fig. 5). Furthermore, we confirmed the enzymatic activities for PvTPSa in vivo (Fig. 4E) and PvTPP in vitro (Fig. 4B), but not for PvTPSb. Thus far, all cloned insect Tps genes encode both GT-20 and TrePP motifs, and insect TPP has been proposed to be identical to TPS [21][22][23]. Although PvTpsa ⁄ b also has both of these motifs, we cloned a PvTpp gene distinguishable from PvTpsa ⁄ b and demonstrated the TPP activity of PvTPP. This is the first report of an insect Tpp gene. BlastP and Pfam searches have shown that TPP orthologs possessing only the TrePP motif are likely to occur in several insects, including four dipteran species, such as Culex quinquefasciatus, Anopheles gambiae, Aedes aegypti, Drosophila melanogaster and Drosophila pseudoobscura, and a hemipteran species, Maconellicoccus hirsutus (CPIJ009402 in C. quinquefasciatus; AGAP008225 in Anopheles gambiae; AAEL010684 in Aedes aegypti; CG5171 and CG5177 in D. melanogaster; GA18712 and GA18709 in D. pseudoobscura; and ABN12077 in M. hirsutus). We therefore propose that insect Tps and Tpp genes exist independently, as reported in other organisms, e.g. bacteria, yeast and plants [32].
In Saccharomyces cerevisiae, trehalose synthase forms a heterotetramer with TPS1, TPS2, TPS3 and TSL1 subunits [42,43]. In the complex, the TPS3 and TSL1 subunits, both of which possess GT-20 and TrePP motifs without TPS or TPP activity, act as regulators [27,28,[42][43][44]. In addition, the activity of TPS is enhanced by its aggregation, indicating that heteromeric and ⁄ or homomeric multimerization of the TPS-TPP complex should be important for the production of TPS activity [45]. Similar to S. cerevisiae, other regulatory subunits might constitute the trehalose synthase complex in P. vanderplanki. No cDNAs homologous to TPS3 and TSL1 have been found thus far in the EST database of P. vanderplanki. Although we could not detect TPS activity in PvTPSb (Fig. 5A), acceleration of its expression by desiccation (Fig. 7) suggests that the protein also plays a role in anhydrobiosis induction. PvTPSb might act as a regulatory subunit, in a similar manner to TPS3 and TSL1, interacting with PvTPSa and PvTPP. The absence of enzymatic activity in PvTPSa ⁄ b proteins prepared by an in vitro transcription and translation system might be caused by the inappropriate interaction of components. If PvTPSa also possesses the same property as TPS in yeast, aggregation of PvTPSa caused by dehydration could lead to an enhancement of its activity en route to anhydrobiosis. Further investigation is required to answer these questions. During the induction of dehydration in an anhydrobiotic nematode, A. avenae, lipid is used as the most likely carbon source to synthesize trehalose via the glyoxylate cycle, and glycogen degradation also contributes to trehalose synthesis [39,46]. In addition, in the trehalose synthesis mechanism of A. avenae during anhydrobiosis induction, it has been reported that the excess substrate influx into TPS is caused by the saturation of glycogen synthase as a result of the increase in UDP-glucose and glucose-6-phosphate as dehydration progresses [47]. However, as shown in Fig. 1B, glycogen degradation and trehalose accumulation during the induction of anhydrobiosis in P. vanderplanki occur as a mirror image. This result indicates that, in drying P. vanderplanki larvae, glycogen is the largest source of trehalose synthesis and is gradually converted into trehalose to act as an anhydroprotectant, although we have not yet verified the involvement of the glyoxylate cycle. Neither the expression of PvGp mRNA nor the activity of GP was elevated on desiccation (Figs 2A and 6A), indicating that PvGP is not involved in the degradation of glycogen. However, TPS and TPP activities increased prior to and parallel with trehalose accumulation, respectively, as a result of the upregulation of the expression of the corresponding mRNAs and proteins (Figs 2B, C and 6A, B). In contrast with the case of TPS and TPP, TREH activity was depressed during desiccation treatment, even though the mRNA and protein of PvTreh increased (Figs 2D and 6). These interesting results indicate that In vitro recombinant PvTREH without modification, such as phosphorylation, showed hydrolytic activity (Fig. 4C, D), implying that PvTREH activity in desiccating larvae might be negatively modified post-translationally. In insects, TREH activity is thought to depend on transcriptional regulation, as reported in the ovary and midgut of B. mori [48,49], or on the coexistence of a TREH inhibitor, as in the hemolymph of Periplaneta americana [50]. In S. cerevisiae, TREH is activated through phosphorylation by cdc28 and inactivated by an inhibitor of TREH (DCS1 ⁄ YLR270W) [51][52][53]. Post-translational modification of PvTREH activity could be occurring in a similar manner, such as by phosphorylation or the coexistence of an inhibitor for rapid accumulation and breakdown (see [54]) of trehalose, in dehydrated and rehydrated larvae, respectively.
In P. vanderplanki, the expression and activity of the enzymes of trehalose metabolism are regulated by desiccation stress (Figs 2 and 6). This is the first report concerning the comprehensive analyses of trehalose metabolism enzymes and the corresponding genes in a single insect species, and provides evidence that multiple pathways control trehalose concentration appropriately according to its physiological role. In insects, including P. vanderplanki, trehalose production and utilization as a hemolymph sugar are under hormonal control via the central nervous system under normal conditions [12]. However, in dehydrating P. vanderplanki larvae, trehalose accumulation as an anhydroprotectant is independent of the control of the central nervous system [40], and is instead triggered by an increase in internal ion concentration [41]. A requirement for rapid adaptation to a desiccating environment could have led to the evolution of the cell autonomous responsive systems in P. vanderplanki larvae.
Here, we summarize a probable molecular mechanism underlying trehalose metabolism that is involved in anhydrobiosis induction in P. vanderplanki (Fig. 7). Once larvae are exposed to drying conditions, fat body cells receive the desiccation signal through the elevation of internal ion concentration and rapidly activate certain desiccation-responsive transcription factors to enhance the transcription of PvTpsa ⁄ b and PvTpp genes participating in trehalose synthesis. Indeed, mRNAs of PvGp, PvTpsa ⁄ b and PvTpp are abundantly expressed in fat body tissue, but the PvTreh mRNA level is less than that in other tissues (Fig. S1, Table S2 and Doc. S1). Furthermore, the PvTPSa ⁄ b protein localizes only to fat body tissue (Fig. S2 and Doc. S1). Concomitant with the accumulation of PvTPSa ⁄ b and PvTPP proteins, the aggregation of PvTPSa ⁄ b-TPP complexes, facilitated by dehydration of the cells, might potentiate the activity of the complex, resulting in the very rapid production of trehalose. Synthesized trehalose then diffuses via the hemolymph through TRET1 to protect all cells and tissues from irreversible desiccation damage (see [7][8][9]). Just before the completion of anhydrobiosis, the expression of PvTreh is accelerated, and the activity of PvTREH is depressed, for subsequent activation during rehydration. Consequently, strict temporal regulation of the pathway of trehalose metabolism, in response to desiccation stress, seems to be the key for the completion of anhydrobiosis in P. vanderplanki. Interestingly, P. nubifer, a desiccation-sensitive and congeneric chironomid to P. vanderplanki, contains trehalose at a comparable level to that in P. vanderplanki under normal conditions, but it does not accumulate trehalose during desiccation (data not shown). Therefore, among the chironomid species, P. vanderplanki seems to be specifically adapted to dehydration by controlling the expression of trehalose metabolism-related genes and the activities of the proteins. In future studies, the determination of the cis-elements and trans-factors of PvTps and other desiccation-inducible genes will be essential in order to obtain a comprehensive understanding of the regulatory mechanisms underlying the induction of anhydrobiosis. Such an understanding could also lead to the exploitation of desiccation-responsive heterologous gene expression systems that are crucial for the reconstitution of the anhydrobiotic state.

Insects
Polypedilum vanderplanki larvae were reared on a milk agar diet under a controlled photoperiod (13 h light : 11 h dark) at 27°C [40,55]. Procedures for the desiccation treatment for the induction of anhydrobiosis-related genes have been described previously [41].

Determination of glycogen and trehalose content in P. vanderplanki
Larvae of P. vanderplanki desiccated for various periods were homogenized in 80% ethanol to obtain soluble and insoluble fractions. The soluble fractions were prepared for the determination of trehalose as described previously [40]. The insoluble fractions were boiled for 30 min in the presence of 30% KOH; glycogen was then precipitated in 80% ethanol and collected by centrifugation at 20 000 g for 15 min at room temperature. The resulting glycogen precipitates were dissolved in distilled water. The glycogen content was determined by the phenol-sulfuric acid method [56].

Determination of the PvTps gene structure
Genomic DNA was extracted from the larvae of P. vanderplanki using a DNeasy Tissue Kit (Qiagen, Hilden, Germany). The construction of the fosmid library and the screening of the clones containing the PvTps gene were entrusted to TaKaRa Bio Inc., Shiga, Japan. The positive clones were subjected to sequencing analysis, and the structure of the PvTps gene was determined. The primer sets used are shown in Table S1.

Northern blot analysis
Total RNA was isolated from dehydrating larvae using TRIzol (Invitrogen, Carlsbad, CA). Northern blot analysis was performed as described previously [9,33]. Briefly, 15 lg of RNA was electrophoresed on 1% agarose-20 mm guanidine isothiocyanate gels, blotted onto Hybond N-plus membrane (GE Healthcare Bioscience, Piscataway, NJ) and probed with the full length of the corresponding cDNA fragments labeled with [a-32 P]dATP using a Strip-EZ labeling kit (Ambion, Austin, TX). The hybridized blot was analyzed by BAS 2500 (Fuji Film, Tokyo, Japan).

Protein extraction
For western blot analyses, the larvae were homogenized in a 10-fold volume of SDS ⁄ PAGE sample buffer without dye reagent, and boiled for 10 min. The homogenates were centrifuged at 20 000 g for 10 min at room temperature, and the supernatants were collected. The concentration of protein was determined as described previously [14]. The preparation of yeast protein extract was carried out according to Clontech's Yeast Protocols Handbook (PT3024-1; http:// www.clontech.com). For the determination of enzyme activities, the larvae were homogenized in a 20-fold volume of protein extraction buffer (T-PER; Pierce Biotechnology, Rockford, IL) containing a protease inhibitor cocktail (Complete; Roche Diagnostics, Basel, Switzerland), and the supernatants containing the crude protein were obtained by centrifugation at 20 000 g for 5 min at 4°C. The concentration of protein was determined with a BCA Protein Assay Kit (Bio-Rad, Hercules, CA).

Western blot analysis
Using the protein extracts described above, western blot analysis was performed as described previously [9,33]. The blots were treated with anti-PvTPS, TPP or TREH polyclonal IgGs as the primary antibodies to detect the corresponding proteins, and subsequently with goat anti-rabbit IgG (H + L) conjugated with horseradish peroxidase (American Qualex, La Mirada, CA) as the secondary antibody, and reacted with Immobilon Western Chemiluminescent HRP substrate (Millipore, Billerica, CA) to analyze the chemiluminescent signals by LAS-3000 (Fuji Film). The recognition sites of antibodies for PvTPS, TPP and TREH are the following amino acid sequences: (592)GIEGITYAGNH-GLE(605) of PvTPSa ⁄ b, (108)GIDGIVYAGNHGLE(121) of PvTPP and (109)LDKISDKNFRD(119) of PvTREH.

In vitro transcription and translation
In vitro transcription and translation of PvTPSa ⁄ b, PvTPP and PvTREH were performed using a TnT Ò T7 Quick for PCR DNA kit (Promega). Briefly, approximately 200 ng of each PCR product, flanked by a T7 promoter at the 5¢-end and a poly(A) at the 3¢-end of the ORF, were incubated for 90 min at 30°C in a 50-lL reaction mixture containing 1 lL of 1 mm methionine or [ 35 S]methionine (> 37 TBqAEmmol )1 , 400 MBqAEmL )1 ; Muromachi Chemical, Tokyo, Japan). The reaction products were separated by 15% SDS ⁄ PAGE, and the gel was applied to western blot analyses as described above, or for autoradiography to confirm protein synthesis.
Determination of enzyme activity GP (EC 2.4.1.1) assays were performed as follows: 100 lL of 45 mm potassium-phosphate buffer (pH 6.8), containing 0.1 mm EDTA, 15 mm MgCl 2 , 4 lm glucose-1,6-bisphosphate, 0.1 U phosphoglucomutase, 0.6 U glucose-6-phosphate dehydrogenase, 2 mgAEmL )1 glycogen, 0.4 mm NADP and 10 lL of protein extract, were incubated at 30°C for 30 min, monitoring the change in the absorbance at 340 nm (A 340 ). Because the inactive form of GP is activated by an allosteric effector, such as AMP, to determine total GP (active 'a' form and inactive 'b' form) activity, the reactions were performed in the presence of an additional 1 mm 5¢-AMP.
For TPS assays, 200 lL of reaction mixture, containing 2.5 mm glucose-6-phosphate, 2.5 mm UDP-glucose, 2.5 mm MgCl 2 , 100 mm KCl, 1.25 mm phosphoenolpyruvate, 20 lL pyruvate kinase ⁄ lactate dehydrogenase (34 lLAEmL )1 ), 0.3 mm NADH, 30 mm Tris ⁄ HCl (pH 7.4) and 5 lL of protein extract, were incubated at 30°C for 30 min, monitoring the change in A 340 that depends on NADH oxidation. In the case of samples from in vitro transcription and translation, 1.2 lL each of the products were incubated at 30°C for 2 h, and then at 95°C for 10 min to stop the reaction.
Assays for TPP activity were performed in 200 lL of reaction mixture containing 2.5 mm trehalose-6-phosphate, 2.5 mm MgCl 2 , 30 mm Tris ⁄ HCl (pH 7.4) and 20 lL of protein extract. In assays for the in vitro transcription and translation products, 1.2 lL of each of the preparations was used. The mixtures were incubated at 30°C for 1 h, and then at 95°C for 10 min to stop the reaction. The reaction product (trehalose) was measured by HPLC [40].
TREH activity was assayed in 250 lL of 15 mm phosphate buffer (pH 6.0) containing 20 mm trehalose and an appropriate amount of protein preparation. After incubation at 30°C for 0.5-1 h, the reaction mixture was boiled for 5 min. As a control, another reaction mixture was immediately boiled without incubation. The reaction products (trehalose and glucose) were measured by HPLC [40].
A desiccation treatment of 48 h is required to completely desiccate larvae under laboratory conditions [40,41]. Enzyme activities in the larvae were measured from 0 to 40 h after the beginning of desiccation, as it seems likely that no metabolic activity would be detectable in vivo in completely desiccated larvae [2].

Yeast complementation assay
The S. cerevisiae deletion mutants were purchased from Open Biosystems, Huntsville, AL. The deletion strains for TPS1 (MATa; his3D1; leu2D0; lys2D0; ura3D0; YBR126c::kanMX4) and TPS2 (MATa; his3D1; leu2D0; lys2D0; ura3D0; YDR074w::kanMX4) were transformed with pUG35 (http:// mips.gsf.de/proj/yeast/info/tools/hegemann/gfp.html; U. Gueldener and J. H. Hegemann, Heinrich-Heine-Universita¨t Du¨sseldorf, unpublished results), which contains the ORF of PvTpsa, PvTpsb and PvTpp under the MET25 promoter [57]. For the positive and negative controls, wild-type and deletion mutants were transformed with pUG35 containing the GFP ORF instead of the target genes. After selection on synthetic defined (SD) medium lacking uracil, transformants were confirmed by colony PCR. Three independent colonies were picked up for each strain. For the complementation test of the tps1 mutant, transformants of the tps1 deletion mutant with PvTpsa and PvTpsb were grown in SD medium containing 2% galactose and lacking uracil and methionine at 30°C to an exponential phase. After harvesting of the yeast cells, a dilution series of 10 4 -10 1 cells was prepared, and each solution was spotted onto yeast extract and peptone (YP) medium containing galactose conditioned in hyperosmolarity with 1 m NaCl, 50% sucrose or 1.5 m sorbitol. For complementation tests of the tps2 mutant, diluted series of transformants of the tps2 deletion mutant with PvTpp were prepared as for tps1. Each cell suspension was spotted onto SD medium containing 2% galactose and lacking methionine and uracil. To confirm the rescue of the temperature sensitivity of the tps2D mutant, the plates were incubated at 45°C for 5 h and then at 30°C for 3-4 days.

Quantification of trehalose by HPLC
The amount of trehalose was determined by HPLC according to Watanabe et al. [40]. For the determination of intracellular trehalose content, PvTpsa-or PvTpsb-introduced yeast strains were cultured in SD medium containing galactose and lacking uracil and methionine at 30°C for 48 h until the growth curve entered the stationary phase. Yeast cells were harvested and homogenized with glass beads in 80% ethanol. After centrifugation at 20 000 g for 30 min, the supernatants were collected and subjected to sample preparation for HPLC analysis [40].

Supporting information
The following supplementary material is available: Fig. S1. Tissue specificity of expression of PvGp, PvTps a ⁄ b, PvTpp and PvTreh in P. vanderplanki larvae. Fig. S2. Immunostaining of PvTPS protein in desiccating larvae. Doc S1. Experimental procedures for supplementary data. Table S1. Primers for 5¢-and 3¢-RACE, and for the determination of PvTps gene structure. Table S2. Primers for real-time PCR. This supplementary material can be found in the online version of this article.
Please note: As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.