Molecular interaction of neutral trehalase with other enzymes of trehalose metabolism in the fission yeast Schizosaccharomyces pombe Teresa Soto, Alejandro Franco, S. Padmanabhan, Jero Vicente-Soler, Jose Cansado and Mariano Gacto Department of Genetics and Microbiology, Facultad de Biologı ´ a, University of Murcia, Spain Trehalose metabolism is an essential component of the stress response in yeast cells. In this work we show that the prod- ucts of the principal genes involved in trehalose metabolism in Schizosaccharomyces pombe, tps1 + (coding for trehalose- 6-P synthase, Tps1p), ntp1 + (encoding neutral trehalase, Ntp1p) and tpp1 + (that codes for trehalose-6-P phospha- tase, Tpp1p), interact in vitro with each other and with themselves to form protein complexes. Disruption of the gene tps1 + blocks the activation of the neutral trehalase induced by heat shock but not by osmotic stress. We propose that this association may reflect the Tps1p-dependent requirement for thermal activation of trehalase. Data reported here indicate that following a heat shock the enzyme activity of trehalase is associated with Ntp1p dimers or trimers but not with either Ntp1p monomers or with complexes involving Tps1p. These results raise the possibility that heat shock and osmotic stress activate trehalase differ- entially by acting in the first case through an specific mech- anism involving Tps1p–Ntp1p complexes. This study provides the first evidence for the participation of the cata- bolic enzyme trehalase in the structural framework of a regulatory macromolecular complex containing trehalose- 6-P synthase in the fission yeast. Keywords: neutral trehalase; stress; protein interaction. Synthesis and degradation of the nonreducing disaccharide trehalose is carried out by several enzymes that are widely distributed and conserved among prokaryotes and eukary- otes. One probable reason for this conservation is the ability of trehalose to function as a general stress protectant in living organisms. Recent studies have focused on the role of trehalose in stabilizing cellular structures under conditions like desiccation, osmotic or oxidative stresses, and mild heat shock [1,2]. Studies in vitro have confirmed the exceptional properties of trehalose in protecting biological membranes or enzymes subjected to different types of extreme condi- tions [3,4]. All these findings suggest that the trehalose turnover must occur in a coordinated way for this sugar to play its diverse functional roles during the life cycle. Consequently, the study of the enzymes involved in trehalose metabolism and the subtle regulation of their activities at the molecular and cellular level have received a great deal of attention in the last decade. The best known picture for trehalose synthesis and mobilization in simple eukaryotes has emerged from studies in the budding yeast Saccharomyces cerevisiae, where trehalose synthesis is basically a two-step process: trehalose 6-phosphate synthesis by trehalose-6-P synthase (TPS1) from UDP-glucose and glucose 6-P as substrates, and dephosphorylation of treha- lose-6-P to trehalose by trehalose-6-P phosphatase (TPS2). Studies based on two-hybrid analyses and on Western blot analyses of complexes obtained by gel filtration fractiona- tion concluded that TPS1 and TPS2 together with proteins TSL1andTPS3(whichactasregulatorsofbothsynthase and phosphatase activities) form part of a multimeric protein complex of approximate molecular mass 800 kDa called the trehalose synthase complex [5–7]. In this complex, TPS1, TPS2, and TPS3 subunits interact with each other and among themselves (as dimers or higher order oligo- mers), whereas TSL1 interacts only with TPS1 and TPS2 [6]. Another component of trehalose metabolism is the hydro- lysis of trehalose to glucose. This is catalyzed by the enzyme neutral trehalase (NTH1), whose activity is regulated by phosphorylation of the enzyme protein at serine residues [8,9]. Comparatively much less is known about the regulation of the trehalose synthesis in the evolutionarily distant yeast Schizosaccharomyces pombe. In this fission yeast, the tps1 + gene codes for trehalose-6-P synthase (Tps1p), that synthe- sizes trehalose-6-P as occurs in S. cerevisiae [10]. However, in contrast to the behaviour observed in S. cerevisiae, Dtps1 strains of S. pombe are able to grow on glucose or other readily fermentable carbon sources, although disruption of this gene does prevent spore germination [10]. Recently, we characterized a second gene of the trehalose biosynthetic pathway in S. pombe,namedtpp1 + , which codes for trehalose-6-P phosphatase (Tpp1p) and is responsible for the synthesis of trehalose from trehalose-6-P [11]. In S. pombe, trehalose degradation is due to the action of a 84-kDa neutral trehalase protein (Ntp1p), encoded by ntp1 + gene [12,13] activated by phosphorylation on Correspondence to J. Cansado, Department of Genetics and Microbiology, Facultad de Biologı ´ a, University of Murcia, 30071 Murcia, Spain. Fax: + 34 68 363963, Tel.: + 34 68 364953, E-mail: jcansado@um.es Abbreviations: EMM2, Edinburgh minimal medium; GST, glutathi- one S-transferase; Ha6H, hemagglutinin antigen epitope and six histidines; Ntp1p, neutral trehalase protein; Tps1p, trehalose-6-P synthase protein; Tpp1p, trehalose-6-P phosphatase protein; TTC, triphenyltetrazolium chloride (Received 15 March 2002, revised 6 May 2002, accepted 27 June 2002) Eur. J. Biochem. 269, 3847–3855 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03082.x supplementing cultures with glucose and a nitrogen source. Activation of this enzyme protein appears to be mediated by protein kinases Pka1p and Sck1p [14,15]. In addition, mRNA levels of ntp1 + rise when S. pombe cultures are subjected to thermal, osmotic or oxidative stresses [12,16]. The increased expression of ntp1 + is accompanied by a rise in enzyme activity, which is also regulated by the Pka1p/Sck1p pathway during osmotic and oxidative stress [16,17]. In a previous work, we showed that mutants of S. pombe disrupted in trehalose-6-P synthase function were unable to increase neutral trehalase activity under heat shock conditions or after the addition of glucose or nitrogen source. However, these Dtps1 strains still respond to osmotic stress by increasing trehalase levels [18]. Thus, in S. pombe, trehalose-6-P synthase appears to be involved in the regulation of the thermal- and nutrient-induced activa- tion of neutral trehalase but not during activation on osmotic stress. These observations raised the question of whether trehalose-6-P synthase and neutral trehalase would interact in vivo. In this paper we demonstrate that both Tps1p and Ntp1p interact in vivo and are part of distinct trehalose-6-P synthase/trehalase complexes. Moreover, trehalose-6-P phosphatase (Tpp1p) is likely a member of these complexes suggesting that regulation of trehalose synthesis and breakdown may be integrated mechanisms. These findings indicate a divergence in the molecular designs controlling trehalose metabolism in S. pombe from those in S. cerevisiae. MATERIALS AND METHODS Strains and culture media The S. pombe strains employed in this study are listed in Table 1. They were routinely grown with shaking at 28 °C in YES [19] or EMM2 with or without thiamin (5 mgÆL )1 ) [12]. Culture media were supplemented with adenine, leucine, histidine or uracil (100 mgÆL )1 , all obtained from Sigma Chemical Co.) depending on the requirements for each particular strain. For sporulation of diploids, MEL medium was employed [19]. Solid media were made by the addition of 2% (w/v) bacto-agar (Difco Laboratories). Transformation of S. pombe strains was performed by the lithium acetate method as described elsewhere [19]. Escher- ichia coli DH5aF¢ was employed as a host to propagate plasmids. It was grown at 37 °C in Luria–Bertani medium plus 50 lgÆmL )1 ampicillin. Construction of Ntp1p-, Tps1p-, and Tpp1p-tagged strains A5¢ truncated version of ntp1 + ORF was amplified by PCR employing the 5¢ oligonucleotide NTP-1 (CCG CTCGAG TCGAATATCTGCCGGAAG, which hybridizes at posi- tions 701–718 in the ntp1 + ORF and contains an internal XhoIsite),andthe3¢ oligonucleotide TAG-3 (CTAC G GCGGCCGCCATTTTTATGAATGGAAA, which hybridizes at the 3¢ end of ntp1 + ORF and incorporates a NotI site placed immediately upstream of the TAA stop codon). The restriction sites in both oligonucleotides are underlined. PCR amplification employing the Expand high- fidelity system (Roche Molecular Biochemicals) generated a 1.6 kb fragment that was cleaved with XhoIandNotIand cloned into plasmids pIH-ura and pIH-LEU. These are integrative plasmids derived from plasmid pDS472a [20], without nmt promoter and ars1 sequences, and with ura4 + or LEU2 as selectable markers, which allow the construc- tion of vectors with Ha6H tag fusions at the C-terminus. The resulting plasmids were digested at the unique SfiIsite within the ntp1 + coding region (at position 1665) and the linear fragments transformed into haploid strains MM1 and MM2, to target integration at the ntp1 + locus. Uracil or leucine prototrophs were selected for strains MM1 and MM2, respectively. The identification of strains C3 and C69, with one copy of Ntp1p–Ha6H expressed from the genomic ntp1 + promoter, was verified by Southern blot analysis and immunoblot of whole-cell extracts with anti- Ha Ig (see below). To obtain Tps1p-tagged strains, a 5¢ truncated version of tps1 + ORF was amplified by PCR employing the 5¢ oligonucleotide TPSINT-1 (CCG CTCGAGCCTAACGG TGTGGAATAC, which hybridizes at positions 715–732 in the tps1 + ORF and shows an internal XhoIsite),andthe3¢ oligonucleotide TAF-3 (CTACG GCGGCCGCCCGAGC TAGAATTCATCGA, which hybridizes at the 3¢ end of tps1 + ORF and incorporates a NotI site placed immediately upstream of the TAA stop codon). The 0.7 kb amplified PCR fragment was cleaved with XhoIandNotI and cloned into integrative plasmids pIH-ura and pIG-ura (fusion to a GST tag at C-terminus). The resulting plasmids were digested at the unique NcoI site within tps1 + (at position 1297) and the linear fragments transformed into haploid strain MM1, to target integration at the tps1 + locus. Uracil prototrophs were selected and the identification of strains C4 and C5, with one copy of Tps1p–GST or Tps1p–Ha6H Table 1. S. pombe strains used in this study. Strain Genotype Source/Reference MM1 h + ade6-M216 leu 1-32 ura4-D18 M. Yamamoto MM2 h – ade6-M210 leu 1-32 ura4-D18 A. Dura ´ n C3 h + ade6-M216 leu 1-32 ura4-D18 ntp1 + :Ha6H (ura4 + ) This study C69 h – ade6-M210 leu 1-32 ura4-D18 ntp1 + :Ha6H (LEU2) This study C4 h ) ade6-M216 leu 1-32 ura4-D18 tps1 + :GST (ura4 + ) This study C5 h ) ade6-M216 leu 1-32 ura4-D18 tps1 + :Ha6H (ura4 + ) This study C694 h + ade6-M216 leu 1-32 ura4-D18 ntp1 + :Ha6H (LEU2) tps1 + :GST (ura4 + ) This study MMPI-3a h + ade6-M216 leu 1-32 ura4-D18 tpp1 + :Ha6H (ura4 + ) [11] MMPI-3b h – ade6-M210 leu 1-32 ura4-D18 tpp1 + :Ha6H (ura4 + ) [11] C33 h – ade6-M210 leu 1-32 ura4-D18 tpp1 + :Ha6H (ura4 + ) ntp1 + :Ha6H (ura4 + ) tps1 + :Ha6H (ura4 + ) This study 3848 T. Soto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 expressed from the genomic tps1 + promoter, was verified by Southern blot analysis and immunoblot of whole-cell extracts with anti-GST or anti-Ha Ig, respectively. The double-tagged strain C694 (Ntp1p–Ha6H, Tps1p– GST) was constructed by mating strains C69 and C4, and selecting diploids in EMM2 medium without supplements. Sporulation was performed in MEL medium and the spores purified by glusulase treatment [21] were allowed to germinate in minimal medium plus adenine. Strains with the double-tagged genotype were identified by Southern and immunoblot analysis with anti-HA and anti-GST Ig. Double-tagged strain C33 (Ntp1p–Ha6H, Tpp1p–Ha6H) was constructed by mating strains C3 and MMPI-3b, and selecting diploids in EMM2 medium with leucine. After spore purification and germination, the double-tagged strains were identified by Southern and immunoblot analysis with anti-Ha Ig. The triple-tagged strain C335 (Ntp1p–Ha6H, Tps1p–Ha6H, Tpp1p–Ha6H) was obtained after mating strains C33 and C5, and Southern and immunoblot analysis of germinated spores, as described above. Expression of Tps1p–GST and Ntp1p–GST fusions The tps1 + ORF was amplified by PCR with oligonucleotides TPS-5 (CCG CTCGAGGAATCTTTGTTTTGCTGA, which hybridizes at sequences upstream of the ATG start codon in the tps1 + ORF and shows an internal XhoIsite), and TAF-3. The 1.5 kb product was cloned into the XhoI/ NotI sites of plasmid pDS472M, which is a derivative of plasmid pDS472a with an attenuated version of the nmt1 promoter and LEU2 selectable marker, to create plasmid pTGST, which expresses trehalose-6-P synthase (Tps1) fused to GST at the C-terminus under the control of the medium strength thiamin-repressible promoter. pTGST and control plasmid pDS472M (unfused GST) were transformed into strains C3, C5 and MMPI-3, and leucine prototrophs selected in EMM2 medium plus adenine. For Ntp1p–GST expression, ntp1 + ORF was amplified by PCR with oligonucleotides NTP-5 (CCG CTCGAGG CTATCATTCGTGAATAG, whichhybridizesatsequences upstream of the ATG start codon in the ntp1 + ORF and shows an internal XhoI site), and TAG-3. The 1.5 kb product was cloned as above into pDS472M to create plasmid pNGST, containing an in-frame fusion wherethe 3¢ endofthe ntp1 + ORF is followed by the GST epitope, and whose expression is under the control of the medium strength nmt1 thiamin-repressible promoter. Both pNGST and control plasmid pDS472M were transformed into strains C3, C5 and MMPI-3, and leucine prototrophs were selected. Purification of Ha6H- and GST-tagged proteins by affinity chromatography Total cell homogenates were prepared under native condi- tions employing chilled acid-washed glass beads and lysis buffer (10% glycerol, 50 m M Tris/HCl pH 7.5, 150 m M NaCl, 0.1% Nonidet NP-40, plus an specific protease inhibitor cocktail for fungal and yeast extracts obtained from Sigma Chemical Co.). The lysate was removed and cleared by centrifugation at 10 000 g for 30 min. Ha6H- tagged proteins were purified by using Ni 2+ -nitrilotriacetic acid-agarose beads (Qiagen Inc.) whereas GST and GST- tagged proteins were precipitated using glutathione glutathi- one–Sepharose beads (Amersham-Pharmacia) following the procedures described by Shiozaki & Russell [22]. Immunoprecipitation of Ha6H- and GST-tagged proteins For immunoprecipitation of Ha6H-tagged proteins, the extracts were incubated for 12 h at 4 °C with monoclonal mouse anti-Ha Ig (clone 12CA5, Roche Molecular Bio- chemicals), and the immunocomplexes were adsorbed with protein A–agarose (Roche Molecular Biochemicals) for 4 h at 4 °C. The immunoprecipitation of GST-tagged proteins was performed with a polyclonal sheep anti-GST Ig (Amersham-Pharmacia) and protein G–agarose (Roche Molecular Biochemicals). In all cases the suppliers’ recom- mendations were followed in terms of incubation times and washing of the complexes. Detection of neutral trehalase activity in gel Affinity purified Ntp1p-Ha and Tps1p–GST proteins were mixed with loading buffer (100 m M Mes pH 6, 15% glycerol and 0.01% bromophenol blue), and resolved at 4 °C in native 6% polyacrylamide gels (200 V for 8 h), employing Tris/borate pH 7.5 as running buffer. The gels were then washed with 100 m M Mes pH 6.0 for 10 min, and incubated with 100 m M Mes pH 6 plus 200 m M trehalose (Sigma Chemical Co.) for 2 h at 30 °C. After washing with distilled water, active neutral trehalase proteins were detected in situ by incubating the gels with 0.1% TTC in 0.5 M sodium hydroxide at 80 °C. Color development was stopped with a 7.5% acetic acid solution. Gel filtration A Superdex-200 column (Amersham-Pharmacia) equili- brated with buffer A (10 m M Mes, pH 6.0, 150 m M NaCl) was used for size-exclusion analysis in an AKTA HPLC system (Amersham-Pharmacia). Lower salt concentrations were not used in order to minimize nonspecific electrostatic interactions with the column matrix. The column was calibrated using vitamin B 12 (1.3 kDa), cytochrome c (12.4 kDa), carbonic anhydrase (29 kDa), ovalbumin (43 kDa), BSA (66 kDa), yeast alcohol dehydrogenase (150 kDa), b-amylase (200 kDa), apoferritin (430 kDa) and thyroglobulin (670 kDa) (all from Sigma Chemical Co.) at the concentrations recommended by the manufac- turer. One-hundred microliters (1 mg total protein) of the supernatant from ultracentrifuged extract obtained from triple-tagged strain C335 grown to mid-log phase, after either a heat shock (40 °C, 1 h) or an osmotic shock (0.75 M NaCl, 2 h) were applied to the column with buffer A plus a yeast protease inhibitor cocktail (Sigma Chemical Co.). A flow rate of 0.4 mLÆmin )1 was used and the elution was tracked by absorbance at 280, 235 and 220 nm. Blue Dextran (2000 kDa; Sigma) and vitamin B 12 were used for void volume (V o )andtotalbedvolume(V t ) determinations, respectively. Molecular masses were estimated from a calibration curve (correlation coefficients ‡ 0.98) generated by linear regressions of the elution volume for each protein using SIGMAPLOT (Jandel Scientific). To determine the fractions containing Ntp1p-, Tps1p-, and Tpp1p–Ha6H tagged proteins, 100 lL of each fraction (0.4 mL) were Ó FEBS 2002 Neutral trehalase complexes in fission yeast (Eur. J. Biochem. 269) 3849 trichloroacetic-acid-precipitated, washed with cold acetone, air-dried, resuspended in SDS gel sample buffer, and resolved in SDS/PAGE (10%) gels. Immunoreactive bands were detected by Western blot analysis with anti-Ha Ig (see below). SDS/PAGE and Western blotting Proteins were resolved in 8 or 10% SDS/PAGE gels as previously described [11], transferred to nitrocellulose filters (Amersham-Pharmacia), and incubated with mouse anti- Ha or sheep anti-GST Ig. The immunoreactive bands were revealed with HRP-conjugated secondary Ig [anti-(mouse IgG) Ig or anti-(sheep IgG) Ig; Sigma Chemical Co.] and the ECL system (Amersham-Pharmacia). Enzyme assays and trehalase activation Trehalase activity was assayed after cell breakage as described previously [23]. Activation of trehalase by heat treatment or osmotic shock was carried out as indicated earlier [17]. Enzyme activity in eluates was expressed as nmol glucose produced per min. All trehalase determina- tions were repeated at least three times with consistent results. Representative results are shown. Specific activity of trehalase in slab gels was expressed as enzyme units per mg protein. Protein determination was performed by absor- bance measurement at 280 and 205 nm according to the method described in previously [24]. RESULTS Neutral trehalase and trehalose-6- P synthase association in vitro In order to analyse possible interactions between treha- lose-6-P synthase and neutral trehalase interaction in vitro, we constructed strains C3 and C69, which express a C-terminal Ha6H-tagged version of neutral trehalase (Ntp1p, 84 kDa) under the control of the genomicpromoter.Thesestrainsshowedthesamegrowth behavior and neutral trehalase activation pattern as the parental strains MM1 and MM2 for a variety of conditions. Strain C3 was further transformed with plasmid pTGST, which expresses trehalose-6-P synthase (55 kDa) fused to a GST epitope (25 kDa) at its C-terminus under the regulation of an attenuated version of the thiamin-repressible promoter [20]. Several transfor- mants were cultured in minimal medium plus thiamin, and then transferred to the same medium without thiamin to allow synthesis of the Tps1p–GST fusion protein. The cells were collected at mid-log phase of growth, and Tps1p–GST was affinity-purified employing glutathione (glutathione)–Sepharose beads. As shown in Fig. 1 (lane 2), significant levels of Tps1p–GST were detected employing anti-GST Ig under these conditions. However, if Tps1p interacts with Ntp1p in vitro, Ntp1p–Ha6H should be detectable with anti-Ha Ig after Tps1p–GST purification. As shown in Fig. 1 (lane 6), this appears to be the case, and a clear band of the size expected for Ntp1p–Ha6H fusion was visible after hybridization with anti-Ha Ig. Besides, the Ntp1p–Ha6H signal was absent when strain C3 was transformed with a control plasmid that expresses unfused GST (Fig. 1, lane 5). Taken together, these results indicate that Ntp1p and Tps1p interact in vitro in S. pombe, and that the nature of this association is specific and independent of the presence of the GST domain fused to Tps1p. Ntp1p–Tps1p association takes place during normal yeast growth and thermal shock, but might not occur when the cells are stressed by an osmotic upshift [18]. To test this possibility, we performed the same experiment described above by subjecting strain C3 plus plasmid pTGST either to osmotic stress or to a thermal one. As shown in Fig. 1, Ntp1p–Tps1p association was observed not only in control, exponentially growing cells but also for both thermal as well as osmotic stresses (lanes 7 and 8). Thus, in S. pombe, Ntp1p–Tps1p interaction does not appear to be transient but is stable, and is maintained even under conditions where the presence of Tps1p is not needed for neutral trehalase activation (osmotic shock). These results, however, do not exclude by themselves the possibi- lity that some Ntp1p might be in a free, nonassociated state (see below). We employed a medium-strength thiamin-repressible promoter for Tps1p–GST expression in Ntp1p–Ha6H cells to achieve low levels of Tps1p–GST synthesis. Although Ntp1p–Ha6H was not detectable in the absence of the Tps1p–GST fusion, it was conceivable that the described Tps1p–Ntp1p interaction could be due, in part, to the presence of nonphysiological levels of Tps1p–GST. In order to clarify this point, we constructed the S. pombe double- tagged strain C694. This strain expresses Ntp1p and Tps1p fused at their C-terminus to Ha6H and GST epitopes, respectively, and in both cases the synthesis is regulated by their own genomic promoters. The Tps1p fusion protein is active because strain C694 synthesizes trehalose at normal level. This strain was grown in rich medium to mid-log phase, and after obtaining the corresponding extracts, Tps1p–GST and Ntp1p–Ha6H were immunoprecipitated with anti-GST and anti-Ha Ig, respectively. As shown in Fig. 2 (lane 2), the Ntp1p–Ha6H band was clearly detected Fig. 1. Ntp1p–Tps1p association takes place in vivo in growing cells of S. pombe, and in cells subjected to heat and osmotic stresses. Strain C3, with a Ha6H epitope-tagged version of Ntp1p, was transformed with plasmids pDS472a (unfused GST; lanes 1 and 5) or pTGST (Tps1p– GST fusion; lanes 2, 3, 4, 6, 7 and 8). GST and Tps1p–GST fusions were expressed using the medium-strength thiamin-regulated promoter for 24 h. Yeast lysates prepared from exponentially growing cells (lanes 2 and 6), or from either heat- (lanes 3 and 7) or osmotically shocked cells (lanes 4 and 8), were adsorbed with glutathione–Sepha- rose beads. After extensive washing in lysis buffer, the proteins bound to the beads were analyzed by Western blot using anti-GST (lanes 1–4) and anti-Ha Ig (lanes 5–8). 3850 T. Soto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 with anti-Ha Ig in the complexes obtained from Tps1p– GST immunoprecipitation, whereas Tps1p–GST was visi- ble with anti-GST Ig after Ntp1p–Ha6H immunoprecipi- tation (lane 5). In this strain, the existence of Ntp1p–Tps1p interaction was evident not only in growing cells, but also during heat or osmotic stress (not shown). These data strongly suggest the existence of an in vivo association between Ntp1p and Tps1p in S. pombe over a variety of environmental conditions. To ascertain the functional significance of this interac- tion, we developed a gel assay for neutral trehalase activity (see Materials and methods) and determined the enzyme activity of Ntp1p bounded to Tps1p in exponen- tially growing cells from double-tagged strain C694 and in cultures subjected to heat or osmotic stress. As shown in Fig. 2B, affinity-purified Ntp1p–Ha6H from growing cells displayed a typical pattern of neutral trehalase activity, with low level of enzyme activity in unstressed cells that increases strongly upon stress. Notably, we were unable to detect in situ any neutral trehalase activity associated to native Tps1–GST protein purified with glutathione– Sepharose beads in samples from either unstressed, osmotic- or heat-shocked cells. A quantitative estimation of neutral trehalase activity gave similar results (Fig. 2B, lower panel). Because a significant fraction of Ntp1p protein is bounded in vitro toTps1p(seeFigs1and2A), these results demonstrate that in S. pombe the active form of neutral trehalase exists in a free, non Tps1p-associated state, independently of the environmental condition used for stress. We focussed then our attention on the likelihood that Tps1p or Ntp1p undergo self-association. This has been previously reported for the Tps1p homologue in S. cerevisiae [6]. Using the same experimental procedure used in Fig. 1, we observed that self-association does indeed take place for both proteins. As can be seen in Fig. 3A (lane 4) and Fig. 3B (lane 4), both Ntp1p– Ha6H and Tps1p–Ha6H were detected employing anti- Ha Ig after Ntp1p–GST and Tps1p–GST purification, respectively. These and other results (see below) support Fig. 2. Ntp1p and Tps1p coimmunoprecipitate and associate to form complexes devoid of neutral trehalase activity. (A) The double-tagged strain C694 (Ntp1p–Ha6H, Tps1p–GST) was grown to mid-log phase, and Tps1p–GST and Ntp1p–Ha6H were immunoprecipitated from the corresponding extracts with anti-Ha Ig (lanes 1 and 5) or anti-GST Ig (lanes 2 and 4) Ig. After incubation with protein A–agarose (anti-Ha immunoprecipitations) or protein G–agarose (anti-GST immunoprecipitations) and extensive washing, the immunocomplexes were resolved by SDS/PAGE and analyzed by Western blot using anti- Ha (lanes 1, 2 and 3) or anti-GST Ig (lanes 3, 5 and 6). Lanes 3 and 6 correspond to negative controls without immunoprecipitation but incubated with protein A–agarose (lane 3) or protein G–agarose (lane 6) to account for possible nonspecific protein binding to the matrix. (B) Upper panel: Affinity-purified Ntp1p–Ha6H and Tps1p–GST were prepared from exponentially growing cells of strain C694 prior (lanes 1 and 4) and after osmotic (lanes 2 and 5) or heat stress (lanes 3 and 6). Samples were resolved by native PAGE and spots developed for neutral trehalase activity. Lower panel: quantitative estimation of enzyme activity (as trehalase units per mg protein) in each sample. Fig. 3. Ntp1p and Tps1p self-interact in vivo. (A) Ntp1p–Ntp1p interaction. The Ntp1p–Ha6H epitope-tagged strain C3 was transformed with plasmids pDS472a (unfused GST; lanes 1 and 3) or pNGST (Ntp1p–GST fusion; lanes 2 and 4). GST and Ntp1p–GST fusions were expressed using the medium-strength thiamin-regulated promoter for 24 h. Yeast lysates were adsorbed with glutathione–Sepharose beads and after washing in lysis buffer the proteins bound to the beads were analyzed by SDS/PAGE and Western blotting using anti-GST Ig (lanes 1 and 2) and anti-Ha Ig (lanes 3 and 4). (B) Tps1p–Tps1p interaction. The Tps1p–Ha6H epitope-tagged strain C5 was transformed with plasmids pDS472a (unfused GST; lanes 1 and 3) or pTGST (Tps1p–GST fusion; lanes 2 and 4). GST and Tps1p–GST fusions were expressed using the medium-strength thiamin- regulated promoter for 24 h. Yeast lysates were processed as described above, and the proteins bound to glutathione beads were analyzed after SDS/PAGE using anti-GST Ig (lanes 1 and 2) and anti-Ha Ig (lanes 3 and 4). Ó FEBS 2002 Neutral trehalase complexes in fission yeast (Eur. J. Biochem. 269) 3851 the existence of a multiprotein complex involved in the regulation of trehalose synthesis and breakdown in S. pombe. Trehalose- 6P phosphatase (Tpp1) is a member of the Ntp1–Tps1 complex Recently a third member of the trehalose metabolism pathway in S. pombe,thetpp1 + gene, which codes for trehalose-6-P phosphatase, has been isolated and charac- terized [11]. The tpp1 + gene has considerable sequence homology to S. cerevisiae TPS2, which encodes trehalose- 6-P phosphatase. S. cerevisiae Tps2 has been shown to interact, among others, with Tps1, and form part of the trehalose synthase complex in this yeast [6,7]. Based on these precedents, we examined if Tpp1p interacts with both Tps1p and Ntp1p in S. pombe. The strain MMPI-3a, that expresses a Ha6H-tagged version of Tpp1p [11], was transformed separately with plasmids pTGST and pNGST (expressing Tps1p and Ntp1p fused to GST, respectively), and the GST fusions were purified with glutathione– Sepharose beads. In either case (Fig. 4A,B), the 100 kDa Tpp1p–HA6H protein coprecipitated with the purified GST fusion proteins, whereas it was absent in control experi- ments (with plasmids expressing unfused GST). These results clearly suggest that Tpp1p may also participate in vivo to form Ntp1p–Tps1p complexes. However, as for Tps1p–Ntp1p complexes (see Fig. 2B), trehalase activity wasabsentinTpp1p–Ntp1passemblieswhenassayedongel slabs (data not shown). Analysis of the Ntp1p–Tps1p–Tpp1p complex by HPLC-gel filtration The results obtained thus far suggest the existence of complexes in S. pombe formed by mutual interaction among Ntp1p, Tps1p, and Tpp1p. To gain additional information on their physical nature in terms of size and composition, we constructed the triple-tagged strain C335 by mating double-tagged strain C33 (Ntp1p–Ha6H, Tpp1p–Ha6H) and Tps1p–Ha6H tagged strain C5, which affords simultaneous detection of Tps1p, Ntp1p and Tpp1p by Western blot analysis using anti-Ha Ig. Protein extracts obtained from exponentially growing cells of strain C335 were fractionated in native solution conditions by gel filtration employing a Superdex-200 HPLC column. The fractions eluted from the column were subjected to SDS/ PAGE, blotted, and subsequently analyzed by Western blot with anti-Ha Ig. As shown in Fig. 5A, the elution profiles for Tps1p, Ntp1p and Tpp1p are quite intricate, with Tps1p being detected at higher levels than either Ntp1p or Tpp1p. Moreover, Tps1p was detected over a wide range of elution volumes extending from that expected for the monomeric protein (fractions 18–20) to high molecular mass complexes close to 700–800 kDa (fractions 4–5), as well as intermediate-sized complexes. Neutral trehalase (Ntp1p) was also detected in virtually all the fractions, but was most pronounced in the 700–800 kDa complex fractions and in those correspond- ing to sizes between 80 kDa (free monomer) and 300 kDa. Finally, Tpp1p showed an elution profile similar to Ntp1p and, like Tps1p and Ntp1p, was also found in the 800 kDa complex. These data indicate that, in addition to the monomeric forms, at least two distinct complexes exist containing significant amounts of all three proteins. One corresponds to the high molecular mass complex (700–800 kDa) and the other to a more diffusely spread population corresponding to lower molecular mass range (80–250 kDa, fractions 12–17). To clarify which of these protein complexes exhibit neutral trehalase activity, we prepared log-phase cultures from strain C335 and subjected them to thermal (40 °C, 1 h) or osmotic (0.75 M NaCl, 2 h) stress. The corresponding cell extracts were then fraction- ated by gel filtration under the conditions described above. No significant differences in the elution profile of Ntp1p, Tps1p and Tpp1p were found under these circumstances, despite an increase in the overall signal (as the three proteins behave as heat shock proteins). Surprisingly, when the eluted fractions were assayed for neutral trehalase, no enzyme activity was found associated with Ntp1p mono- mers or in fractions corresponding to the high molecular mass trehalose synthase–trehalase complex (fractions 4–5). Instead, the heat-activated neutral trehalase was mainly detected in complexes of about 250 kDa, whereas neutral Fig. 4. Tpp1p associates with Ntp1p and Tps1p in vivo . (A) Tpp1p–Tps1p interaction. The Tpp1p–Ha6H epitope-tagged strain MMPI-3a was transformed with plasmids pDS472a (unfused GST; lanes 1 and 3) or pTGST (Tps1p–GST fusion; lanes 2 and 4). GST and Tps1p–GST fusions were expressed using the medium-strength thiamin-regulated promoter for 24 h. Yeast lysates were then adsorbed with glutathione–Sepharose beads, and after extensive washing the proteins bound to the beads were analyzed by Western blot using anti-GST Ig (lanes 1 and 2) and anti-Ha Ig (lanes 3 and 4). (B) Ntp1p–Tpp1p interaction. Strain MMPI-3a was transformed with plasmids pDS472a (unfused GST; lanes 1 and 3) or pNGST (Ntp1p–GST fusion; lanes 2 and 4). GST and Tps1p–GST fusions were expressed using the medium-strength thiamin-regulated promoter for 24 h, purified as described above, and analyzed using anti-GST (lanes 1 and 2) and anti-Ha Ig (lanes 3 and 4). 3852 T. Soto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 trehalase activated by osmotic shock peaked mostly in fractions corresponding to somewhat lower molecular masses. These enzyme activities fit the expected elution for trehalase trimers and dimers, respectively. DISCUSSION Two main lines of evidence lead us to propose that in the fission yeast S. pombe neutral trehalase (Ntp1p) interacts in vitro with other proteins involved in trehalose metabo- lism. First, the activation of Ntp1p by heat shock or nutritional stimuli only takes place in the presence of trehalose-6-P synthase (Tps1p) [18]. Secondly, in contrast to the neutral trehalase in S. cerevisiae [25], all attempts to activate S. pombe Ntp1p in vitro have been unsuccessful to date. Taken together, these results support the idea that in S. pombe the regulation of Ntp1p activity under stress relies on the existence of some kind of interaction between Ntp1p and Tps1p, or other elements, that may be lost under the conditions of the in vitro assay for activation. In the present work, we have performed experiments including affinity chromatography, immunoprecipitation, and gel filtration to demonstrate that not only Tps1p, but also Tpp1p, interact with Ntp1p and that macromolecular complexes, involving trehalose synthase/phosphatase and trehalase, might indeed exist in this yeast. The demonstration of interactions between Ntp1p–Tps1p and Ntp1p–Tpp1p (Figs 1, 2 and 4) is, to our knowledge, the first description of an association between the hydrolytic enzyme trehalase and proteins involved in synthesis of trehalose, thus revealing a novel relationship in trehalose metabolism. In E. coli, trehalases are monomeric enzymes [26]. However, in S. cerevisiae the functional cytoplasmic neutral trehalase (NTH1) is probably a dimer, as deduced from gel filtration experiments with active enzyme [25]. No evidence has been reported demonstrating that NTH1 participates in the formation of the well-characterized trehalose synthase complex in the budding yeast [6,7]. In S. pombe, we find that part of Ntp1p is apparently present as free monomeric protein that does not display trehalase activity in vitro (Fig. 5). The elution profile of trehalase activity in gel filtration indicated that after heat or osmotic shock only fractions corresponding to protein sizes between 170 and 300 kDa contain active enzyme. On the other hand, Ntp1p eluted mainly in two peaks, either as part of a 700–800 kDa complex, together with Tps1p and Tpp1p, or, as expected, in the fractions containing trehalase activity, which again showed the coexistence of Tps1p and Tpp1p (Fig. 5). Because Ntp1p molecules also associate among themselves (Fig. 2), these results alone would not allow us to establish whether trehalase activity correlates with self-assembly of Ntp1p homomultimers or with the formation of hetero- meric complexes involving Tps1p and Tpp1p in addition to Ntp1p. However, the observation that Tps1p–(Tpp1p)– Ntp1p complexes lack trehalase activity (Fig. 2B) strongly favors the first interpretation. Active trehalase might thus arise from the specific self-assembly of a discrete number of Ntp1p molecules to form small oligomers (probably dimers or trimers, considering the molecular mass of each putative subunit, 84 kDa). An intriguing fact is that there is a small but reproducible shift in the elution behavior of trehalase depending on the nature of the activation stress (Fig. 5B). This shift might argue against the participation of a unique set of interactions involving association of Ntp1p molecules and could be taken to indicate a specific event occurring during activation by hyperthermia that is absent in the response elicited by osmotic stress. However, the elution pattern of trehalase cannot be exclusively interpreted in Fig. 5. Analysis of Tps1p–Tpp1p–Ntp1p complexes by gel filtration. (A) A high-speed supernatant of cell-free extracts from exponentially growing, triple Ha6H-tagged strain C335, was size fractionated by gel filtration through a Superdex-200 column. The proteins from 100 lLofeachfraction were concentrated by trichloroacetic acid precipitation, separated by SDS/PAGE, transferred to nitrocellulose and incubated with anti-Ha Ig. After incubation with an HRP-conjugated secondary antibody [anti-(mouse IgG) Ig], the signals specific for Ha-tagged proteins were visualized with the ECL system. V 0 indicates the void volume. (B) Elution profile of activated neutral trehalase in S. pombe. Exponentially growing cultures of strain C335 were subjected to either heat shock (open circles) or osmotic stress (closed circles), and the correspondent protein extracts size fractionatedas described for (A). Neutral trehalase activity (expressed as nmol glucose produced per min) was assayed in 250 lLofeachfraction. Ó FEBS 2002 Neutral trehalase complexes in fission yeast (Eur. J. Biochem. 269) 3853 terms of changes in the complex composition. Because gel filtration is responsive to Stokes radius, one should also consider that the conformation has merely changed upon activation depending on the triggering stimulus while the composition is unaltered. Tps1p is probably the most abundant protein related to trehalose metabolism present in S. pombe (Fig. 5), although we can not exclude that the differential relative expression of Ntp1p, Tps1p and Tpp1p in eluted extracts might be an experimental artifact arising from differences in the acces- sibility of the antibody to the Ha tag. However, there is a parallel in S. cerevisiae, where the Tps1p homologue (TPS1) is also present at levels higher than those of the other constituents of the trehalose synthase complex [7]. In S. cerevisiae, trehalose is synthesized by the 800 kDa trehalose synthase complex as well as by free TPS1 [7]. It is unknown whether Tps1p activity in S. pombe is present in the high molecular mass trehalose synthase–trehalase com- plex or linked to forms of lower molecular mass as for neutral trehalase, although accumulation of trehalose by ntp1-deleted strains seems to indicate that Tps1p function does not require association with Ntp1p. In any case, our data suggest the existence of different types of Ntp1p complexes in S. pombe that might be specifically activated as a function of the external stimulus. If so, an attractive explanation for the above results would be that S. pombe harbors at least two forms of Ntp1p that are capable of being activated, one as a Ntp1p–Ntp1p homodimer ( 170 kDa), that becomes activated during osmotic shock, and the other as a Tps1p–(Tpp1p)–Ntp1p heterocomplex ( 250 kDa) activated during heat shock. Apparently, trehalase requires association with Tps1p while being activated by heat shock acquires enzymatic activity only when detached from the heat-induced activation complex. Other possibilities may exist, but some experimental data are consistent with this hypothesis. For instance, in contrast to heat shock, Ntp1p activation during osmotic shock is independent of the presence of Tps1p [18]. Also, the activation of neutral trehalase induced by heat shock is largely a post-translational event that exhibits considerably faster kinetics than the osmotically induced activation [17,23]. In the latter case, the Pka/Sck1p-mediated activa- tion of Ntp1p appears to occur at the level of the enzyme synthesized de novo [17]. Finally, the Tps1p–Ntp1p associ- ation can occur under any growth condition or stress treatment (Fig. 1), which lends physiological significance to this association. In this context it should be mentioned that Tpp1p disruption mutants show heat-shock activation of Ntp1p [11], which supports the view that Tpp1p, but not Tps1p, may be dispensable for this activation of trehalase. The advantages of a complex bearing enzymes separately implicated in synthesis and breakdown of trehalose could provide improved control efficiency of the respective enzymatic activities as has been suggested in the case of the synthesizing Tps1–Tps2 complex in S. cerevisiae [7]. On the other hand, although this work has demonstrated the existence of Tps1p–(Tpp1p)–Ntp1p complexes, one or more of these may also interact with, as yet, unknown elements to form the final complex. In this context, an exhaustive search for homologues to Tps1p, Tpp1p and Ntp1p in the databases of the S. pombe sequencing project (Sanger Center, Cambridge, UK), revealed the existence of three putative ORFs (SPAC3G6.09c, SPAC2F8.05 and SPACUNK.16c) encoding proteins that show amino acid identity with Tps1p and Tpp1p ranging from 34 to 42%, and that may be good candidates for additional interac- tions. More remarkably, the degree of identity of these ORF products is also high with respect to TSL1 and TPS3, two members of the trehalose synthase complex in S. cerevisiae that regulate both trehalose synthase and phosphatase activities. In particular, the protein sequence deduced from ORF SPAC2F8.05 indicated 37% identity with TSL1, whereas that of SPACUNK.16c showed a 38% of identity with TPS3. Hence, as described for the TPS1–TPS2 complex in S. cerevisiae, it is tempting to speculate that in S. pombe these ORFs code for proteins that somehow regulate trehalose metabolism by interacting with some members of the enzyme proteins described here. The data presented in this report reveal a crucial difference between the two organisms. Unlike previously reported for S. cerevisiae [25], both the anabolic and catabolic enzymes might be integrated in some instances into a regulatory complex in S. pombe. ACKNOWLEDGEMENTS T. S. and A. F. contributed equally to this work. We thank Prof F. J. Murillo for generous access to the HPLC equipment and F. Garro for expert technical assistance. We are indebted to Profs. M. Yamamoto, S. L. Forsburg and A. Duran for kindly providing plasmids and strains. A. F. is a predoctoral fellow of PFPI from the University of Murcia. This work was supported in part by grant BMC2001-0135 from MCYT, Spain. REFERENCES 1. Eroglu, A., Russo, M.J., Bieganski, R., Fowler, A., Cheley, S., Bayley, H. & Toner, M. (2000) Intracellular trehalose improves the survival of cryopreserved mammalian cells. Nat. Biotechnol. 18, 163–167. 2. Guo, N., Puhlev, I., Brown, D.R., Mansbridge, J. & Levine, F. (2000) Trehalose expression confers desiccation tolerance on human cells. Nat. Biotechnol. 18, 168–171. 3. Colaco, C., Sen, S., Thangavelu, M., Pinder, S. & Roser, B. (1992) Extraordinary stability of enzymes dried in trehalose: simplified molecular biology. Biotechnology 10, 1007–1011. 4. Singer, M.A. & Lindquist, S. (1998) Multiple effects of trehalose on protein folding in vitro and in vivo. Mol. Cell 1, 639–648. 5. Vuorio, O.E., Kalkkinen, N. & Londesborough, J. (1993) Cloning of two related genes encoding the 56-kDa and 123-kDa subunits of trehalose synthase from the yeast Saccharomyces cerevisiae. Eur. J. Biochem. 216, 849–861. 6. Reinders, A., Burckert, N., Hohmann, S., Thevelein, J.M., Boller, T., Wiemken, A. & De Virgilio, C. (1997) Structural analysis of the subunits of the trehalose-6-phosphate synthase/phosphatase complex in Saccharomyces cerevisiae and thsir function during heat shock. Mol. Microbiol. 24, 687–695. 7.Bell,W.,Sun,W.,Hohmann,S.,Wera,S.,Reinders,A.,De Virgilio, C., Wiemken, A. & Thevelein, J.M. (1998) Composition and functional analysis of the Saccharomyces cerevisiae trehalose synthase complex. J. Biol. Chem. 273, 33311–33319. 8. Kopp, M., Mu ¨ ller, H. & Holzer, H. (1993) Molecular analysis of the neutral trehalase gene from Saccharomyces cerevisiae. J. Biol. Chem. 268, 4766–4774. 9. Wera, S., De Schrijver, E., Geyskens, I., Nwaka, S. & Thevelein, J.M. (1999) Opposite roles of trehalase activity in heat-shock recovery and heat-shock survival in Saccharomyces cerevisiae. Biochem. J. 343, 621–626. 3854 T. Soto et al. (Eur. J. Biochem. 269) Ó FEBS 2002 10. Blazquez, M.A., Stucka, R., Feldmann, H. & Gancedo, C. (1994) Trehalose-6-P synthase is dispensable for growth on glucose but not for spore germination in Schizosaccharomyces pombe. J. Bacteriol. 176, 3895–3902. 11. Franco, A., Soto, T., Vicente-Soler, J., Valero Guillen, P., Cansado, J. & Gacto, M. (2000) Characterization of tpp1+ as encoding a main trehalose-6-P phosphatase in the fission yeast Schizosaccharomyces pombe. J. Bacteriol. 182, 5880–5884. 12. Cansado,J.,Soto,T.,Fernandez,J.,Vicente-Soler,J.&Gacto,M. (1998) Characterization of mutants devoid of neutral trehalase activity in the fission yeast Schizosaccharomyces pombe: partial protection from heat shock and high-salt stress. J. Bacteriol. 180, 1342–1345. 13. Soto, T., Fernandez, J., Dominguez, A., Vicente-Soler, J., Cansado, J. & Gacto, M. (1998) Analysis of the ntp1 + gene, encoding neutral trehalase in the fission yeast Schizosaccharomyces pombe. Biochim. Biophys. Acta 1443, 225–229. 14. Soto,T.,Fernandez,J.,Cansado,J.,Vicente-Soler,J.&Gacto,M. (1995) Glucose-induced, cyclic-AMP-independent signalling pathway for activation of neutral trehalase in the fission yeast Schizosaccharomyces pombe. Microbiology 141, 2665–2671. 15. Soto,T.,Fernandez,J.,Cansado,J.,Vicente-Soler,J.&Gacto,M. (1997) Protein kinase Sck1 is involved in trehalase activation by glucose and nitrogen source in the fission yeast Schizosaccharo- myces pombe. Microbiology 143, 2457–2463. 16. Fernandez, J., Soto, T., Franco, A., Vicente-Soler, J., Cansado, J. & Gacto, M. (1998) Enhancement of neutral trehalase activity by oxidative stress in the fission yeast Schizosaccharomyces pombe. Fungal Genet. Biol. 25, 79–86. 17. Fernandez,J.,Soto,T.,Vicente-Soler,J.,Cansado,J.&Gacto,M. (1997) Osmo-stress-induced changes in neutral trehalase activity of the fission yeast Schizosaccharomyces pombe. Biochim. Biophys. Acta 1357, 41–48. 18. Cansado, J., Vicente-Soler, J., Soto, T., Fernandez, J. & Gacto, M. (1998) Trehalose-6-P synthase is essential for trehalase activation triggered by glucose, nitrogen source or heat shock, but not by osmostress. Schizosaccharomyces pombe. Biochim. Biophys. Acta. 1381, 271–278. 19. Moreno, S., Klar, A. & Nurse, P. (1991) Molecular genetic ana- lysis of the fission yeast Schizosaccharomyces pombe. Methods Enzymol. 194, 795–823. 20. Forsburg, S.L. & Sherman, D.A. (1997) General purpose tagging vectors for fission yeast. Gene 191, 191–195. 21. Kelly, T.J., Martin, G.S., Forsburg, S.L., Stephen, R.J., Russo, A. & Nurse, P. (1993) The fission yeast cdc18+ gene product couples S-phase to start and mitosis. Cell 74, 371–382. 22. Shiozaki, K. & Russell, P. (1997) Stress-activated protein kinase pathwayincellcyclecontroloffissionyeast.Methods Enzymol. 283, 506–520. 23. Carrillo, D., Vicente-Soler, J. & Gacto, M. (1994) Cyclic AMP signalling pathway and trehalase activation in the fission yeast Schizosaccharomyces pombe. Microbiology 140, 1467–1472. 24. Stoscheck, C.M. (1990) Quantitation of protein. Methods Enzy- mol. 182, 50–68. 25. Londesborough, J. & Varimo, K. (1984) Characterization of two trehalases in baker’s yeast. Biochem. J. 219, 511–518. 26. Uhland, K., Mondigler, M., Spiess, C., Prinz, W. & Ehrmann, M. (2000) Determinants of translocation and folding of TreF, a trehalase of Escherichia coli. J. Biol. Chem. 275, 23439–23445. Ó FEBS 2002 Neutral trehalase complexes in fission yeast (Eur. J. Biochem. 269) 3855 . Molecular interaction of neutral trehalase with other enzymes of trehalose metabolism in the fission yeast Schizosaccharomyces pombe Teresa. enzyme trehalase in the structural framework of a regulatory macromolecular complex containing trehalose- 6-P synthase in the fission yeast. Keywords: neutral