Báo cáo khoa học: Sugar and alcohol molecules provide a therapeutic strategy for the serpinopathies that cause dementia and cirrhosis pot

Sugar and alcohol molecules provide a therapeutic strategy for the serpinopathies that cause dementia and cirrhosis Lynda K. Sharp 1, *, Meera Mallya 1, *, Kerri J. Kinghorn 1 , Zhen Wang 1 , Damian C. Crowther 1 , James A. Huntington 2 , Didier Belorgey 1 and David A. Lomas 1 1 Department of Medicine, University of Cambridge, UK 2 Department of Haematology, University of Cambridge, UK Neuroserpin is a serine proteinase inhibitor that is secreted by axons of the central and peripheral nervous systems [1–3]. It is a potent inhibitor of tissue plasmi- nogen activator (tPA) [4–7] and it is likely that neuro- serpin–tPA interactions regulate neuronal and synaptic plasticity [3,8], and play an important role in learning, memory and behaviour [9]. The regulation of tPA by neuroserpin has a role in the pathogenesis of epilepsy [10,11] and limits the tissue damage that results from ischaemic stroke [12,13]. Neuroserpin is a member of the serine proteinase inhibitor or serpin superfamily [14]. Members of this family have > 30% amino acid sequence homology and share a conserved tertiary structure based on three b sheets, nine a helices and an exposed mobile reactive centre loop. This loop presents a peptide sequence as a pseudosubstrate for the target protein- ase. After docking with the enzyme, the reactive loop of the serpin is cleaved and the molecule undergoes a profound conformational transition that swings the proteinase from the upper to the lower pole of the serpin [15]. This is achieved by the cleaved reactive loop snapping into b-sheet A and in most cases the resulting covalently linked complex is stable for many weeks. However, this is not so for the neuro- serpin ⁄ tPA complex which slowly dissociates to Keywords a 1 -antitrypsin; FENIB; neuroserpin; polymerization; serpinopathy Correspondence M. Mallya, Department of Medicine, University of Cambridge, Cambridge Institute for Medical Research, Wellcome Trust ⁄ MRC Building, Hills Road, Cambridge CB2 2XY, UK Fax: +44 1223 336827 Tel: +44 1223 336825 E-mail: mm342@cam.ac.uk Website: http://www.cimr.cam.ac.uk *These authors contributed equally to this study. (Received 14 February 2006, accepted 5 April 2006) doi:10.1111/j.1742-4658.2006.05262.x Mutations in neuroserpin and a 1 -antitrypsin cause these proteins to form ordered polymers that are retained within the endoplasmic reticulum of neurones and hepatocytes, respectively. The resulting inclusions underlie the dementia familial encephalopathy with neuroserpin inclusion bodies (FENIB) and Z a 1 -antitrypsin-associated cirrhosis. Polymers form by a sequential linkage between the reactive centre loop of one molecule and b-sheet A of another, and strategies that block polymer formation are likely to be successful in treating the associated disease. We show here that glycerol, the sugar alcohol erythritol, the disaccharide trehalose and its breakdown product glucose reduce the rate of polymerization of wild-type neuroserpin and the Ser49Pro mutant that causes dementia. They also attenuate the polymerization of the Z variant of a 1 -antitrypsin. The effect on polymerization was apparent even when these agents had been removed from the buffer. None of these agents had any detectable effect on the structure or inhibitory activity of neuroserpin or a 1 -antitrypsin. These data demonstrate that sugar and alcohol molecules can reduce the polymeriza- tion of serpin mutants that cause disease, possibly by binding to and stabil- izing b-sheet A. Abbreviations FENIB, familial encephalopathy with neuroserpin inclusion bodies; tPA, tissue plasminogen activator. 2540 FEBS Journal 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS generate active tPA and inactive, reactive-loop- cleaved neuroserpin [6,7]. Point mutations in neuroserpin can profoundly affect secretion and result in the accumulation of mutant neuroserpin as inclusions (or Collin’s bodies) within the endoplasmic reticulum of neurones in the deep layer of the cerebral cortex [16–18]. These inclu- sions underlie an autosomal-dominant dementia that we have termed familial encephalopathy with neuroser- pin inclusions bodies (FENIB) [17]. Disease-causing mutations perturb b-sheet A of neuroserpin allowing incorporation of the reactive centre loop of a second molecule [17]. This reactive loop–b sheet dimer can then extend to form chains of polymers that are retained within the cell. Polymers of mutant neuroser- pin have been isolated from Collin’s bodies of individ- uals with FENIB [17] and we have shown that mutants of neuroserpin that cause FENIB (Ser49Pro and Ser52Arg) form polymers in vitro [6,19,20] and in cell models of disease [21]. Mutations of neuroserpin that favour intermolecular loop insertion and polymer- ization also allow intramolecular incorporation of the reactive loop and the formation of an inactive latent species [20]. The formation of polymers also underlies diseases associated with point mutations of other members of the serpin superfamily. a 1 -Antitrypsin is secreted from the liver and is the most abundant protease inhibitor in the circulation. The severe Z deficiency variant (Glu342Lys) results in the formation of polymers [22–26] that are retained as inclusions in the rough endoplasmic reticulum of the liver, where they are associated with juvenile hepatitis, cirrhosis and hepato- cellular carcinoma [27,28]. The lack of circulating a 1 -antitrypsin causes early-onset emphysema [29]. Moreover, intrahepatic polymerization of variants of other serpins: C1 inhibitor, antithrombin and a 1 -antic- hymotrypsin, cause plasma deficiency that results in conditions as diverse as angio-oedema, thrombosis and emphysema, respectively [30–33]. This common molecular pathology has allowed us to group these conditions together as the serpinopathies [34,35]. A variety of strategies have been developed to reduce polymer formation in an attempt to prevent the associ- ated disease [22,36–45]. Previous studies have shown that the trihydric alcohol glycerol reduced the poly- merization of antithrombin and a 1 -antitrypsin [46] and increased the secretion of the Z variant of a 1 -anti- trypsin in a cell-culture model of disease [39]. The serpinopathies have obvious parallels with other con- formational diseases that result from aberrant b-strand linkage such as Huntington’s disease [47]. This condi- tion can be retarded by feeding Huntington’s mice with the disaccharide trehalose [48]. We report here that glycerol, the larger sugar alcohol erythritol, treha- lose and the monosaccharide glucose (Fig. 1) all reduce the rate of polymerization of mutants of neuroserpin and a 1 -antitrypsin, possibly by binding to and stabil- izing b-sheet A. Results Glycerol, erythritol, trehalose and glucose reduce the rate of polymerization and increase the transition temperature of wild-type neuroserpin when added to the polymerization buffer Glycerol reduced the rate of polymerization of wild- type neuroserpin in a concentration-dependant manner when added directly to the reaction buffer. The find- ings were confirmed by multiple repeats with the max- imal effect being a 2.4-fold reduction in polymerization (n ¼ 5, P ¼ 0.003) with 1.36 m (10% v ⁄ v) glycerol at 45 °C (Fig. 2A). The longer sugar alcohol erythritol had a similar effect, reducing polymerization of wild- type neuroserpin by 2.8-fold (n ¼ 3, P ¼ 0.002) at 1.36 m (Fig. 2A,C). However, unlike glycerol, 0.14 m erythritol caused an increase in the rate of polymeriza- tion when compared with 0.07 or 0.2 m erythritol. This increase was not statistically significant. Trehalose and its breakdown product glucose also reduced the rate of polymerization of wild-type neuroserpin when incuba- ted at 45 °C (Fig. 2B). It was found that 1.36 m glu- cose almost entirely abolished polymerization with most of the monomeric protein being converted to the latent species. The limited solubility of trehalose pre- cluded assessment at the same concentrations as glu- cose, glycerol and erythritol. Nevertheless trehalose also markedly reduced the rate of polymerization of Fig. 1. Structures of glycerol (A), erythritol (B), glucose (C) and trehalose (D). L. K. Sharp et al. A therapeutic strategy for the serpinopathies FEBS Journal 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS 2541 wild-type neuroserpin when added at a final concentra- tion of 1.02 m. The rate of polymerization was so slow that it was difficult to obtain a value in our standard 24 h assay. Even at lower concentrations (0.79 m) both trehalose and glucose decreased the rate of polymeriza- tion of wild-type neuroserpin by approximately four- fold (n ¼ 3, P ¼ 0.003 for both trehalose and glucose). It is possible that the effects of glycerol, erythritol, trehalose and glucose were nonspecific and mediated by their effect on viscosity. This was assessed by measuring the polymerization of wild-type neuroser- pin in 5 and 10% w ⁄ v Ficoll PM70, which have visc- osities of 1.82 and 3.14 cP, respectively [49]. In comparison 1 and 2 m glycerol have viscosities of 1.31 and 1.71 cP, respectively, and 0.6 and 0.8 m tre- halose have viscosities of 1.58 and 2.08 cP, respect- ively [50]. Incubating with 5% w ⁄ v Ficoll PM70 had no effect on the polymerization of wild-type neuro- AB DC Fig. 2. The effect of alcohols and sugars on the polymerization of wild-type neuroserpin when added to the polymerization buffer. (A, B) Increasing concentrations of glycerol, erythritol, glucose or trehalose were added to wild-type neuroserpin in NaCl ⁄ P i (final concentration 0.4 mgÆmL )1 ) and the mixture incubated at 45 °C. The rate of polymerization was assessed by densitometry of the monomeric band on 7.5% w ⁄ v nondenaturing PAGE. The results are the mean and standard error of at least three independent experiments. *P<0.05, **P<0.01 com- pared with the rate without the compounds. X, glycerol; n, erythritol; h, trehalose; e, glucose. (C) 7.5% w ⁄ v acrylamide nondenaturing PAGE to assess the polymerization of wild-type neuroserpin. Neuroserpin was incubated in NaCl ⁄ P i at 0.4 mgÆmL )1 and 45 °C without (upper) or with (lower) the addition of 1.36 M erythritol. The lanes correspond to 0, 1, 2, 3, 4, 5, 6, 7, and 24 h incubation and are representative of three independent experiments. (D) Transition temperatures of wild-type neuroserpin (0.25 mgÆmL )1 ) were determined with and without the alcohols and sugars by monitoring the CD signal at 216 nm between 25 and 90 °C. Solid black line, wild-type neuroserpin; solid grey line, neuro- serpin with 1.36 M erythritol. A therapeutic strategy for the serpinopathies L. K. Sharp et al. 2542 FEBS Journal 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS serpin. However 10% w ⁄ v Ficoll PM70 had a small but significant effect on the polymerization of wild- type neuroserpin (1.23 · 10 )5 Æs )1 compared with 1.6 · 10 )5 Æs )1 for the wild-type protein) but this was still less than the effect of 10% v ⁄ v glycerol (1.01 · 10 )5 Æs )1 ). Thus the increase in viscosity caused by the sugar and short-chain alcohols does not explain the effect of glycerol on the polymerization of neuroserpin and may only account for a small amount of the effect of trehalose. Glycerol, erythritol, trehalose and glucose all increased the transition temperature of wild-type neuroserpin in keeping with increased stability and reduced rates of polymerization (Fig. 2D and Table 1), but had no effect on the far-UV CD spectrum when added directly to the protein (data not shown). Glycerol reduces the rate of polymerization of wild-type neuroserpin even when removed from the polymerization buffer The effect of glycerol, erythritol, trehalose and glu- cose was then assessed after refolding the protein from the Escherichia coli cell pellet in the presence of the alcohol or sugar in buffer D and then removing the compound using nickel agarose and Q-Sepharose chromatography. Refolding in 1.36 m glycerol reduced the rate of polymer formation at 0.4 mgÆmL )1 and 45 °C by 1.7-fold compared with neuroserpin that had been treated identically but which had not been refolded in glycerol (n ¼ 5, P ¼ 0.04). Refolding in 0.68 and 2.72 m glycerol had a very similar effect to 1.36 m glycerol (n ¼ 6, P < 0.05), whereas refolding in 0.14 m glycerol did not significantly alter the poly- merization rate of wild-type neuroserpin (n ¼ 3, P ¼ 0.22). In comparison, refolding neuroserpin in erythri- tol, trehalose or glucose had no significant effect on the rate of polymerization of wild-type neuroserpin (Table 2). The effect of glycerol on the polymerization of neu- roserpin was investigated further by refolding neuro- serpin in buffer C and then adding 1.36 m glycerol for 1 h after filtration. The protein was then purified by nickel chelating and Q-Sepharose chromatography and concentrated into NaCl ⁄ P i as detailed above. The addition of 1.36 m glycerol following refolding still significantly reduced the rate of polymerization of wild-type neuroserpin at 0.4 mgÆmL )1 and 45 °Cby 1.5-fold (n ¼ 3, P ¼ 0.04). Thus even a brief exposure of folded neuroserpin to glycerol can reduce the pro- pensity of the molecule to polymerize, with a similar level of reduction to that seen when the protein was refolded in glycerol. This was despite the protein being subjected to two purification steps and then concentrated using buffers that did not contain any glycerol. Neither refolding in glycerol nor adding glycerol just after refolding had any effect on the CD spectrum or transition temperature of wild-type neuroserpin or the inhibitory kinetics with tPA: k a ¼ 2.1 · 10 4 m )1 Æs )1 (n ¼ 3), 1.1 · 10 4 m )1 Æs )1 (n ¼ 3) and 1.9 · 10 4 m )1 Æs )1 (n ¼ 2), respectively, for neuroserpin refolded in the absence or presence of 1.36 m glycerol or adding 1.36 m glycerol after refolding. Table 1. The effect of glycerol, erythritol, glucose and trehalose on the transition temperature (°C) of wild-type and Ser49Pro neuroserpin and Z a 1 -antitrypsin when added directly to the reaction mixture. NaCl ⁄ P i 1.36 M w ⁄ v glycerol 1.36 M w ⁄ v erythritol 1.36 M w ⁄ v glucose 1.02 M w ⁄ v trehalose Wild-type neuroserpin 59.8 (± 0.4) 61.7 (± 0.3) 62.8 (± 0.6) 66.2 (± 0.4) 65.8 (± 0.9) Ser49Pro neuroserpin 55.7 (± 1.5) 56.7 (± 1.8) 63.1 (± 1.4) 64.5 (± 1.2) 65.9 (± 3.3) at 0.68 M Z a 1 -antitrypsin 60.0 (± 0.5) 62.0 (± 0.9) 63.3 (± 0.9) 64.4 (± 0.7) at 0.68 M 64.7 (± 1.0) at 0.68 M Table 2. Polymerization rates of wild-type and Ser49Pro neuroserpin refolded in glycerol, erythritol, glucose or trehalose. Rates are expressed in s )1 and are the mean and standard deviation of three independent experiments. Protein NaCl ⁄ P i 1.36 M w ⁄ v glycerol 1.36 M w ⁄ v erythritol 1.36 M w ⁄ v glucose 1.02 M w ⁄ v trehalose Wild-type 45 °C 2.35 (± 0.61) · 10 )5 1.37 (± 0.28) · 10 )5 * 2.30 (± 0.78) · 10 )5 2.37 (± 0.43) · 10 )5 2.00 (± 0.35) · 10 )5 Ser49Pro 37 °C 4.80 (± 0.96) x 10 )6 3.49 (± 0.15) · 10 )6 3.20 (± 0.14) · 10 )6 * 4.33 (± 0.48) · 10 )6 5.85 (± 0.28) · 10 )6 Ser49Pro 45 °C 1.89 (± 0.52) · 10 )4 1.73 (± 0.28) · 10 )4 1.07 (± 0.09) · 10 )4 * 1.44 (± 0.18) · 10 )4 1.53 (± 0.29) · 10 )4 *P < 0.05 compared with wild-type or Ser49Pro neuroserpin without glycerol, erythritol, glucose or trehalose. L. K. Sharp et al. A therapeutic strategy for the serpinopathies FEBS Journal 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS 2543 Glycerol, erythritol, trehalose and glucose reduce the rate of polymerization and increase the transition temperature of Ser49Pro neuroserpin that causes FENIB In view of the effect of glycerol, erythritol, trehalose and glucose on the polymerization of wild-type neuro- serpin we also assessed the effect of these compounds on the polymerization of Ser49Pro neuroserpin that causes the dementia FENIB. The rate of polymeriza- tion of Ser49Pro neuroserpin (at 0.4 mgÆmL )1 and 45 °C) was reduced by 1.36 m glycerol, 1.36 m erythri- tol, 1.19 m glucose and 0.84 m trehalose by 2.3-fold (n ¼ 4, P ¼ 0.02), 3.2-fold (n ¼ 4, P ¼ 0.009), 3.6-fold (n ¼ 4, P ¼ 0.01) and 4.9-fold (n ¼ 4, P ¼ 0.006), respectively, when added to the polymerization buffer (Fig. 3A,B). The reduction in polymerization was also observed when the compounds were incubated with Ser49Pro neuroserpin at 37 °C (Fig. 3C,D). We found that 1.36 m glycerol, 1.36 m erythritol, 0.84 m trehalose and 0.84 m glucose reduced polymerization by 2.1-fold (P ¼ 0.002), 2.3-fold (P ¼ 0.006), 2.6-fold (P ¼ 0.007) and 2.8-fold (P ¼ 0.002), respectively (n ¼ 5 for all experiments). In keeping with the results for wild-type neuroserpin, all the compounds increased the trans- ition temperature of Ser49Pro neuroserpin when added directly to the buffer (Table 1). Erythritol reduces the rate of polymerization of Ser49Pro neuroserpin even when removed from the polymerization buffer Ser49Pro neuroserpin was refolded from the E. coli cell pellet in 1.36 m glycerol, 1.36 m erythritol, 1.02 m treha- lose or 1.36 m glucose in buffer D and the compounds then removed by nickel agarose and Q-Sepharose chro- matography. Only erythritol caused a reduction in the Fig. 3. The effect of alcohols and sugars on the polymerization of Ser49Pro neuroserpin when added to the polymerization buffer. Increasing concentrations of glycerol, erythritol, glucose or trehalose were added to Ser49Pro neuroserpin (final concentration 0.4 mgÆmL )1 ) and the mixture incubated in NaCl ⁄ P i at 45 °C (A, B) or 37 °C (C, D). The rate of polymerization was assessed by densitometry of the monomeric band on 7.5% w ⁄ v nondenaturing PAGE. The results are the mean and standard error of at least three independent experiments. *P<0.05, **P<0.01 compared with the rate without the compounds. X, glycerol, n, erythritol; h, trehalose; e, glucose. A therapeutic strategy for the serpinopathies L. K. Sharp et al. 2544 FEBS Journal 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS rate of polymerization of neuroserpin when incubated at either 37 or 45 °C (almost twofold, see Table 2; P ¼ 0.04, n ¼ 4). None of the compounds had any effect on the CD spectrum, transition temperature, unfold- ing profile on transverse urea gradient gel or inhibitory kinetics with tPA (k a ¼ 0.22 · 10 4 m )1 Æs )1 , n ¼ 2) of Ser49Pro neuroserpin. We then investigated whether erythritol could mediate its effect on the folded protein, by adding 1.36 m erythri- tol for 1 h after refolding but before the protein was purified over two columns. Again this reduced the rate of polymerization of Ser49Pro neuroserpin by 1.5- (n ¼ 3, P ¼ 0.04) and 1.6-fold (n ¼ 5, P ¼ 0.002) at 37 and 45 °C, respectively, but did not alter the CD spectrum, transition temperature, unfolding profile on transverse urea gradient gel or inhibitory kinetics of Ser49Pro neuroserpin with tPA (k a ¼ 0.19 · 10 4 m )1 Æs )1 , n ¼ 2). Glycerol, erythritol, trehalose and glucose reduce the polymerization and increase the transition temperature of Z a 1 -antitrypsin when added to the polymerization buffer The finding that refolding in glycerol and erythritol reduced the rate of polymerization of wild-type and Ser49Pro neuroserpin, respectively, prompted an assessment of the effect of alcohols and sugars on the Z variant of a 1 -antitrypsin that also causes disease by polymerization. Glycerol is known to enhance the secretion of Z a 1 -antitrypsin in a cell-culture model of the disease [39] and similarly 1.36 m glycerol reduced the polymerization of Z a 1 -antitrypsin by 2.9-fold at 41 °C(n ¼ 3, P ¼ 0.003; Fig. 4A). Because erythritol has a greater effect on mutant rather than on wild-type neuroserpin, the effect of erythritol on Z a 1 -antitrypsin was also investigated. It was found that 1.36 m erythri- tol reduced the rate of polymerization of Z a 1 -antitryp- sin fourfold at 41 °C(n ¼ 3, P ¼ 0.001; Fig. 4A,C) and 5.3-fold when incubated at 37 °C(n ¼ 3, P ¼ Concentration of compound (M) 0 0 1 2 2.5 0.2 0.5 1.5 0.4 0.6 0.8 1 1.2 1.4 1.6 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0.20.1 0.3 0.50.4 0.6 0.80.7 Polymerisation rate (s –1 x 10 –6 ) Polymerisation rate (s –1 x 10 –6 ) A B C Concentration of compound (M) 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 ** ** ** ** ** ** ** ** ** ** * * * Fig. 4. The effect of alcohols and sugars on the polymerization of Z a 1 -antitrypsin when added to the polymerization buffer. Increas- ing concentrations of glycerol, erythritol, glucose or trehalose were added to Z a 1 -antitrypsin (final concentration 0.1 mgÆmL )1 ) and the mixture incubated in NaCl ⁄ P i at 41 °C (A, B). The rate of polymer- ization was assessed by densitometry of the monomeric band on 7.5% w ⁄ v nondenaturing PAGE. The results are the mean and standard error of at least three independent experiments. *P<0.05, **P<0.01 compared with the rate without the com- pounds. X, glycerol; n, erythritol; h, trehalose; e, glucose. (C) 7.5% w ⁄ v acrylamide nondenaturing PAGE to assess the polymer- ization of Z a 1 -antitrypsin. Z a 1 -Antitrypsin was incubated in NaCl ⁄ P i at 0.1 mgÆmL )1 and 41 °C without (upper) or with (lower) the add- ition of 1.36 M erythritol. The lanes correspond to 0, 1, 2, 3, 4, 5, 6, and 7 days incubation and are representative of three independent experiments. L. K. Sharp et al. A therapeutic strategy for the serpinopathies FEBS Journal 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS 2545 0.004). As for wild-type neuroserpin, there was a non- significant increase in the rate of polymerization at 0.14 m erythritol that was not apparent at 0.07 or 0.2 m erythritol. The addition of 0.68 m trehalose reduced the rate of polymerization of Z a 1 -antitrypsin at 41 °C by 7.7-fold (n ¼ 3, P ¼ 0.001), whereas the addition of 0.68 m glucose reduced the rate by 7.4-fold (n ¼ 3, P ¼ 0.001) (Fig. 4B). All these compounds increased the transition temperature of Z a 1 -antitrypsin (Table 1) but did not change the CD spectrum of the protein (data not shown). In order to assess the effect of viscosity we performed the same polymerization experiments with 5 and 10% w ⁄ v Ficoll PM70; 5% w ⁄ v Ficoll PM70 had no effect on the polymerization of Z a 1 -antitrypsin. However 10% w ⁄ v Ficoll PM70 reduced the rate of polymerization of Z a 1 -antitrypsin from 1.72 · 10 )6 to 1.01 · 10 )6 Æs )1 (n ¼ 3, P ¼ 0.015) but this was less than the effect seen with 10% v ⁄ v gly- cerol (5.93 · 10 )7 Æs )1 ). Effect of refolding Z a 1 -antitrypsin with erythritol or glucose Two milligrams of Z a 1 -antitrypsin was unfolded for 2 h in 6 m GuHCl, 100 mm dithiothreitol, 50 mm Tris pH 7.8, before refolding overnight in buffer containing 5mm dithiothreitol, 50 mm Tris pH 7.8 and either 1.36 m erythritol or 0.68 m glucose. All attempts to refold Z a 1 -antitrypsin either with or without the compounds were unsuccessful as Z a 1 -antitrypsin spon- taneously formed polymers. Glycerol, erythritol, trehalose and glucose reduce the rate of polymerization of Z a 1 -antitrypsin even when removed from the polymerization buffer Purified Z a 1 -antitrypsin was incubated with 1.36 m glycerol, 1.36 m erythritol, 0.68 m glucose or 0.68 m trehalose at 4 °C for 1 h and then the Z a 1 -antitrypsin was dialysed into NaCl ⁄ P i . Pre-incubating Z a 1 -anti- trypsin with glycerol, erythritol, trehalose or glucose reduced the rate of polymerization at 41 °C by 1.9-fold (n ¼ 3, P ¼ 0.016), 2.2-fold (n ¼ 3, P ¼ 0.010), 2.4-fold (n ¼ 4, P ¼ 0.004) and 1.9-fold (n ¼ 3, P ¼ 0.014), respectively. The brief exposure to glycerol, erythritol and glucose had no effect on the CD spec- trum, inhibitory activity or transition temperature of Z a 1 -antitrypsin but incubation with trehalose resulted in a small decrease in ellipticity between 195 and 212 nm on CD and a small but significant increase in transition temperature (from 60.0 to 61.1 °C, n ¼ 3, P ¼ 0.046). Discussion Previous studies have shown that glycerol is able to bind to b-sheet A of antithrombin [46] and increase the secre- tion of Z a 1 -antitrypsin from cell models of disease [39]. We show here that glycerol is also able to stabilize and reduce the polymerization of wild-type neuroserpin and the Ser49Pro neuroserpin mutant that causes the demen- tia FENIB. Moreover, glycerol has a similar effect on the Z mutant of a 1 -antitrypsin that polymerizes within hepatocytes to cause liver disease. In view of these find- ings, we assessed the longer sugar alcohol erythritol (Fig. 1) and demonstrated that this molecule was also able to block the polymerization of wild-type and Ser49- Pro neuroserpin and Z a 1 -antitrypsin. Polymer formation results from the sequential link- age between the reactive centre loop of one molecule and b-sheet A of another [17,22,23,51,52]. The mole- cular pathology that underlies this conformational transition is now well defined and has been used as a paradigm for other diseases that result from aberrant b-strand linkage and tissue deposition [34,47]. These include Alzheimer’s disease, Huntington’s disease, Par- kinson’s disease and the amyloidoses. As such, inter- ventions that are effective in blocking b-strand linkages in one of these disorders may also be effective in others. The progression of Huntington’s disease can be slowed in mouse models by feeding the mice with the disaccharide trehalose [48]. We therefore assessed the effect of both trehalose and its metabolite glucose on serpin polymerization. Both of these agents were effective in stabilizing wild-type and Ser49Pro neuro- serpin and Z a 1 -antitrypsin (as evidenced by increased transition temperature) and blocking polymerization. There is an inverse relationship between the melting temperature of serpins and the rate of polymerization [52]. This is seen again here with the addition of alcoh- ols and sugars to wild-type and Ser49Pro neuroserpin. However, in contrast to other serpins such as a 1 -anti- trypsin [52], heating neuroserpin results in an increase rather than a decrease in CD signal [6,19]. This implies that, rather than measuring melting, the assay is reporting an increase in secondary structure. The most likely explanation is that neuroserpin is rapidly form- ing polymers and that the effect of the compounds is to increase the activation temperature required for polymerization. The striking effects of glycerol, erythritol, trehalose and glucose may be nonspecific and result from the increased viscosity. This would decrease the diffusion rates, and, assuming a diffusion-limited reaction, pre- dictably slow polymerization. This explanation is un- likely as the reduction in polymerization rate tends to A therapeutic strategy for the serpinopathies L. K. Sharp et al. 2546 FEBS Journal 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS plateau rather than decrease progressively, as would be expected if the effect were mediated by increasing vis- cosity (see Figs 2–4). Moreover, incubation of neuro- serpin and Z a 1 -antitrypsin with Ficoll PM70, at concentrations that cause an increase in viscosity com- parable with that of the highest concentration of gly- cerol, had no effect on the rate of polymerization. An alternative explanation is that the alcohols and sugars may be able to exert their effect by a specific interac- tion with either neuroserpin or a 1 -antitrypsin. Support for this hypothesis comes from the demonstration that merely flash-cooling antithrombin crystals with gly- cerol as a cryoprotectant was sufficient to allow a gly- cerol molecule to bind to b-sheet A [46] (Fig. 5A). To address this question the compounds were added to neuroserpin either during, or for 1 h after, refolding and then removed by two chromatography columns and dialysis into NaCl ⁄ P i . The brief exposure of wild- type neuroserpin to glycerol and Ser49Pro neuroserpin to erythritol significantly reduced the rate of polymer- ization without affecting the other biochemical pro- perties of the protein. Refolding experiments were impossible with Z a 1 -antitrypsin as it immediately formed polymers. Nevertheless, incubating any of these compounds with Z a 1 -antitrypsin for only 1 h (and then removal by dialysis into NaCl ⁄ P i ) was sufficient to reduce the rate of polymerization by approximately twofold. Once again, this had no effect on the other biochemical properties of the protein. Taken together these data argue that small mole- cules are able to bind specifically to wild-type and mutant serpins and slow down conformational transi- tions. They have no effect on association rate con- stants as the energy that is released on reactive loop cleavage is sufficient to overcome the binding of small molecules. The critical region in stabilizing the serpin molecule is the shutter domain that controls the opening and closing of b-sheet A. Mutations in Fig. 5. Potential binding site for polyols in b-sheet A of neuroserpin based on the struc- ture of glycerol bound to antithrombin. (A) Glycerol (magenta rods) bound in the P8 position of antithrombin (shown in the stand- ard orientation with the yellow RCL and green P1 Arg placed on the top) was observed with a peptide (cyan) bound at the top of b-sheet A (red). A close up of the region containing the glycerol molecule (right-hand panel) reveals hydrogen bonding interactions (broken green rods) with strands 4A and 5A. (B) A similar placement of gly- cerol in neuroserpin allows the preserv- ation of the hydrogen bonds described above for antithrombin. (C) Placement of erythritol preserves the interactions observed for glycerol and creates additional hydrogen bonds which can bridge strands 3, 4and5ofb-sheet A. L. K. Sharp et al. A therapeutic strategy for the serpinopathies FEBS Journal 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS 2547 this region have profound effects on the serpin mole- cule and favour the formation of loop–sheet polymers or the inactive latent conformer [20,34,53]. Indeed, all four neuroserpin mutants that cause the inclusion body dementia FENIB are located in the shutter domain with the destabilizing effect of each mutation being directly proportional to the rate of polymeriza- tion [6,18,19,21]. Analysis of the structure of glycerol bound to antithrombin [46] provides information on the likely site of binding of these compounds (Fig. 5). Glycerol binds to the shutter domain in the position that would be occupied by P8 threonine when the loop is inserted into b-sheet A during com- plex formation [46] (Fig. 5A). This P8 position is therefore the most likely place for compounds to bind to neuroserpin (Fig. 5B), because this region is highly conserved across the serpin family. The hydro- xyl groups of glycerol could form stabilizing hydro- gen bonds with the d-nitrogen of the His334 imidazole, the carbonyl group of Phe333 and the car- bonyl group of the P9 residue (Fig. 5B). Thus gly- cerol is able to cross-link a partially inserted reactive loop (the first step on the pathway to polymerization) [32] to b-sheet A and stabilize the shutter domain against opening further, thereby preventing the incor- poration of the reactive loop of another molecule and hence the formation of polymers. The effect of erythritol on the polymerization rate of neuroserpin might also be due to its binding to the shutter domain of b-sheet A (Fig. 5C). Our previous work has shown that Ser49Pro neuroserpin mutation causes the molecule to adopt a polymerogenic conformation that is intermediate between wild-type protein and fully formed polymers [19]. This conformer has a partially inserted reactive centre loop and a patent b-sheet A [19,20,32]. As well as preserving the in- teractions described in the glycerol-bound model, erythritol could participate in additional hydrogen bonding to strand 3, involving the carbonyl groups of Asn186 and Leu184, and also binds Ser56 of helix B. It is likely that the larger erythritol molecule is required to form sufficient hydrogen bonds between strands 3 and 5 to stabilize the mutant pro- tein against polymerization. The Z mutation of a 1 -antitrypsin is located, not in the shutter domain, but at the head of strand 5 and the base of the reactive centre loop. The mutation for- ces open the gap between strands 3 and 5 of b-sheet A to allow partial loop insertion and a patent lower b-sheet A that can act as a receptor for the loop of another molecule and hence form polymers [42,51]. This patent b-sheet A can also accept exogenous pep- tides that can block polymerization [42,44,46] and, as demonstrated here, is also able to bind small molecules with similar results. It is likely that small molecules will ultimately prove effective in stabilizing polymerogenic serpins to attenu- ate the associated disease. We have assessed four compounds and have shown that they can reduce poly- merization even when exposed only briefly to the serpin molecule. The effects are relatively small but this may be sufficient to treat the associated disease. For example, only 1% of Z a 1 -antitrypsin is retained as intracellular polymers with the majority of the protein being targeted for degradation [54]. Thus only a small shift to stabilize the monomer may be sufficient to prevent the accumula- tion of toxic polymers that cause cell death and disease. Trehalose and erythritol are particularly exciting as lead compounds for treatment of the serpinopathies as they are both well absorbed from the gut and can cross the blood–brain barrier [48,55]. Moreover, any secreted pro- tein will retain its inhibitory activity against the target proteinase. Experimental procedures Materials Ni-NTA agarose was from Qiagen (Crawley, UK), HiTrap Q-Sepharose and Ficoll PM70 were from Amersham Bio- sciences (Little Chalfont, UK), tPA was from Calbio- chem (CN Biosciences UK, Nottingham, UK) and S-2288 (H-d-Ile-pro-Arg-para-nitroanilide) was from Chromogenix (Quadratech, Epsom, UK). Expression and purification of recombinant proteins Wild-type and Ser49Pro neuroserpin were expressed with a His-tag in the pQE81L vector in E. coli SG13009 (pREP4) cells (Qiagen) as described previously [6]. Cells containing the plasmid coding for wild-type or Ser49Pro neuroserpin were collected by centrifugation, resuspended in buffer A (20 mm imidazole, 0.5 m NaCl, 20 mm Na 2 HPO 4 , pH 7.8) and dis- rupted by sonication. The cell pellet was washed three times with buffer B (50 mm Tris ⁄ HCl, 100 mm NaCl, 10 mm EDTA, 20 mm imidazole pH 7.8) containing 0.05% v ⁄ v Tri- ton X-100 and then once more with buffer B alone before solubilization in 10 mm Tris ⁄ HCl, 100 mm NaH 2 PO 4 ,6m guanidinium-HCl pH 8. The protein was refolded overnight at 4 °C in buffer C (20 mm imidazole, 20 mm Na 2 HPO 4 , 150 mm NaCl, pH 7.8) or buffer D (20 mm imidazole, 20 mm Na 2 HPO 4 , 150 mm NaCl, containing either 1.36 m glycerol, 1.36 m erythritol, 1.36 m glucose or 1.02 m trehalose pH 7.8). It was then filtered through a 45 lm membrane before being mixed on a roller ⁄ shaker at 4 °C for 1 h with Ni-NTA chelating agarose precharged with 0.1 m NiSO 4 . A therapeutic strategy for the serpinopathies L. K. Sharp et al. 2548 FEBS Journal 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS Where specified, 1.36 m glycerol or erythritol was added for 1 h at the mixing stage. A 300 mm imidazole solution in 20 mm Na 2 HPO 4 ,20mm NaCl pH 7.8 was used to elute the bound protein, giving a single peak. This fraction was diluted three times in 20 mm Tris ⁄ HCl, 20 mm NaCl pH 7.4 buffer and then loaded onto a HiTrap Q-Sepharose column and eluted with a NaCl gradient (20 mm to 1 m)in20mm Tris ⁄ HCl pH 7.4. Monomeric protein was collected and con- centrated into NaCl ⁄ P i with a Vivaspin concentrator that had been extensively rinsed with distilled water to wash out any residual glycerol. The resulting protein was then snap- frozen and stored at )80 °C. Purified neuroserpin migrated as a single band on SDS ⁄ PAGE and > 90% was in a mono- meric form when assessed by nondenaturing and transverse urea gradient PAGE [56]. Purification of Z a 1 -antitrypsin and refolding/incubation with compounds Z a 1 -antitrypsin was purified from the plasma of PiZ homo- zygotes as described previously [37] and migrated as a single band on SDS, nondenaturing and transverse urea gradient PAGE. Z a 1 -antitrypsin was denatured by incubation in 6 m guanidinium-HCl, 100 mm dithiothreitol, 50 mm Tris pH 7.8 for 2 h at 4 °C and then refolded overnight at 4 °C in 50 mm Tris, 5 mm dithiothreitol pH 7.8 or in 50 mm Tris, 5mm dithiothreitol pH 7.8 with either 1.36 m erythritol or 0.68 m glucose. The refolded protein was then loaded onto a HiTrap Q-Sepharose column (Amersham Biosciences) and eluted in a 0–250 mm NaCl gradient in 50 mm Tris pH 8.0. To assess the effect of stabilizing compounds on the folded protein, Z a 1 -antitrypsin was incubated at 0.1 mgÆmL )1 with 1.36 m glycerol, 1.36 m erythritol, 0.68 m glucose or 0.68 m trehalose for 60–90 min at 4 °C. The compounds were then removed by dialysing 2 · 2 h and then overnight in NaCl ⁄ P i at 4 °C, and the Z a 1 -antitrypsin was concentrated and then assessed in assays of polymerization. Polymerization of wild-type neuroserpin, Ser49Pro neuroserpin and Z a 1 -antitrypsin Polymerization of wild-type neuroserpin, Ser49Pro neuro- serpin and Z a 1 -antitrypsin was assessed by nondenaturing PAGE. Wild-type or Ser49Pro neuroserpin were incubated at 0.4 mgÆmL )1 in NaCl ⁄ P i at 45 °C for wild-type or 37 °C and 45 °C for Ser49Pro neuroserpin. Z a 1 -antitrypsin was incubated at 0.1 mg ÆmL )1 in NaCl ⁄ P i at 37 °Cor41°C. The samples were overlaid with oil to prevent evaporation and 2 lg of protein for each time point was loaded onto a 7.5% w ⁄ v nondenaturing gel. Alcohols or sugars were added to the buffer at a range of concentrations (0–1.36 m) with either sodium azide (0.1% w ⁄ v) for Z a 1 -antitrypsin or Protease Inhibitor Cocktail Tablet for neuroserpin (Roche, one tablet dissolved in 2 mL NaCl ⁄ P i , used at a 1 in 50 dilution). The proteins were visualized by staining with Gel- Code Ò Blue Stain Reagent from Pierce (Tattenhall, UK). The density of the complex bands was determined by densi- tometry scanning with the data being analysed by a semilog plot against time using the software quantity one (Bio- Rad, Hercules, CA). Measurement of the rate of polymer- ization was performed on at least three occasions for each concentration of alcohol or sugar with either wild-type or Ser49Pro neuroserpin or Z a 1 -antitrypsin. Activity assays of serpins Rate constants for the inhibition of tPA by wild-type or Ser49Pro neuroserpin, and of bovine a-chymotrypsin by a 1 -antitrypsin in the presence or absence of alcohols or sugars were determined as described previously [6,37]. Circular dichroism CD experiments were performed in NaCl ⁄ P i using a JASCO J-810 spectropolarimeter. Thermal unfolding experiments were performed by monitoring the CD signal at 216 nm (neuroserpin) or 222 nm (a 1 -antitrypsin) between 25 and 90 °C using a heating rate of 1 °CÆmin )1 at a concentration of 0.25 mgÆmL )1 for wild-type neuroserpin and Z a 1 -anti- trypsin, 0.8 mgÆ mL )1 for Ser49Pro neuroserpin, and 0.18 mgÆmL )1 for Z a 1 -antitrypsin that had been incubated with sugars for 1 h before dialysing into NaCl ⁄ Pi. The trans- ition points were calculated using an expression for a two state transition as described previously [57,58]. The results are the mean and standard deviation of three experiments. Glycerol, erythritol, trehalose or glucose was added to the reaction mixture at the concentrations specified in the figures or text. Statistical analysis Rates of polymerization of Z a 1 -antitrypsin and wild-type and Ser49Pro neuroserpin were compared using Student’s t-test. Structural analysis Models of neuroserpin with bound glycerol and erythritol were built using the published structure of cleaved mouse neuroserpin (1JJO) [59] superimposed on the structure of antithrombin bound to glycerol (1LK6) [46]. Placement of glycerol in neuroserpin is identical to that observed in the antithrombin structure, and a conservative placement of erythritol was made by preserving the important O 2 and O 3 hydrogen bonds to s5A residues. Superposition and analysis of potential hydrogen bonding were conducted using the program xtalview, and figures were prepared using bobscript and raster3d. L. K. Sharp et al. A therapeutic strategy for the serpinopathies FEBS Journal 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS 2549 [...]... RW, Lomas DA, Gerhard L, Baumann B, Lawrence DA, Yepes M, Kim TS, Ghetti B et al (2002) Association between conformational mutations in neuroserpin and onset and severity of dementia Lancet 359, 2242–2247 Belorgey D, Sharp LK, Crowther DC, Onda M, Johansson J & Lomas DA (2004) Neuroserpin Portland (Ser52Arg) is trapped as an inactive intermediate that rapidly forms polymers: implications for the epilepsy... detection of PAS-positive inclusions formed by the Z, Siiyama and Mmalton variants of a1 -antitrypsin Hepatology 40, 1203–1210 An JK, Blomenkamp K, Lindblad D & Teckman JH (2005) Quantitative isolation of alphalAT mutant Z protein polymers from human and mouse livers and the effect of heat Hepatology 41, 160–167 Eriksson S, Carlson J & Velez R (1986) Risk of cirrhosis and primary liver cancer in alpha1-antitrypsin... and the serpinopathies: pathobiology and prospects for therapy J Clin Invest 110, 1585–1590 A therapeutic strategy for the serpinopathies 36 Schulze AJ, Baumann U, Knof S, Jaeger E, Huber R & Laurell C-B (1990) Structural transition of a1 -antitrypsin by a peptide sequentially similar to b-strand s 4A Eur J Biochem 194, 51–56 37 Lomas DA, Evans DL, Stone SR, Chang WS & Carrell RW (1993) Effect of the. .. conformational disease Proc Natl Acad Sci USA 97, 67–72 Eldering E, Verpy E, Roem D, Meo T & Tosi M (1995) COOH-terminal substitutions in the serpin C1 inhibitor that cause loop overinsertion and subsequent multimerization J Biol Chem 270, 2579–2587 Lomas DA & Carrell RW (2002) Serpinopathies and the conformational dementias Nat Rev Genet 3, 759– 768 Lomas DA & Mahadeva R (2002) a1 -Antitrypsin polymerization and. .. Zhou A, Dafforn TR, Foreman RC & Lomas DA (2003) Targeting a surface cavity of a1 -antitrypsin to prevent conformational disease J Biol Chem 278, 33060–33066 44 Parfrey H, Dafforn TR, Belorgey D, Lomas DA & Mahadeva R (2004) Inhibiting polymerisation: new therapeutic strategies for Z a1 -antitrypsin related emphysema Am J Respir Cell Mol Biol 31, 133– 139 45 Zhou A, Stein PE, Huntington JA, Sivasothy... (a1 -AT) with trimethylamine N-oxide Implications for the treatment of a1 -AT deficiency Am J Respir Cell Mol Biol 24, 727–732 42 Mahadeva R, Dafforn TR, Carrell RW & Lomas DA (2002) Six-mer peptide selectively anneals to a pathogenic serpin conformation and blocks polymerisation Implications for the prevention of Z a1 -antitrypsin related cirrhosis J Biol Chem 277, 6771–6774 43 Parfrey H, Mahadeva R, Ravenhill... seen in the dementia FENIB Eur J Biochem 271, 3360– 3367 Onda M, Belorgey D, Sharp LK & Lomas DA (2005) Latent S49P neuroserpin spontaneously forms polymers: identification of a novel pathway of polymerization and implications for the dementia FENIB J Biol Chem 280, 13735–13741 Miranda E, Romisch K & Lomas DA (2004) Mutants ¨ of neuroserpin that cause dementia accumulate as poly- FEBS Journal 273 (2006)... P, Baici A, Pennella A, Krueger SR, Schrimpf SP, Meins M & Sonderegger P (1998) The axonally secreted serine proteinase inhibitor, neuroserpin, inhibits plasminogen activators and plasmin but not thrombin J Biol Chem 237, 2312–2321 6 Belorgey D, Crowther DC, Mahadeva R & Lomas DA (2002) Mutant neuroserpin (Ser49Pro) that causes the familial dementia FENIB is a poor proteinase inhibitor and readily forms... Carrell RW & Wardell MR (1994) Thromboembolic disease due to thermolabile conformational changes of antithrombin Rouen-VI (187 Asn fi Asp) J Clin Invest 94, 2265–2274 Gooptu B, Hazes B, Chang WS, Dafforn TR, Carrell RW, Read RJ & Lomas DA (2000) Inactive conformation of the serpin a1 -antichymotrypsin indicates twostage insertion of the reactive loop: implications for inhibitory function and conformational... mechanism for the polymerisation of a1 -antitrypsin J Biol Chem 274, 9548– 9555 53 Stein PE & Carrell RW (1995) What do dysfunctional serpins tell us about molecular mobility and disease? Nat Struct Biol 2, 96–113 54 Le A, Graham KS & Sifers RN (1990) Intracellular degradation of the transport-impaired human PiZ a1 -antitrypsin variant Biochemical mapping of the degradative event among compartments in the secretory . Sugar and alcohol molecules provide a therapeutic strategy for the serpinopathies that cause dementia and cirrhosis Lynda K. Sharp 1, *, Meera Mallya 1, *, Kerri J. Kinghorn 1 , Zhen Wang 1 ,. Miranda E, Ro ¨ misch K & Lomas DA (2004) Mutants of neuroserpin that cause dementia accumulate as poly- A therapeutic strategy for the serpinopathies L. K. Sharp et al. 2550 FEBS Journal. strands 3 and 5 of b-sheet A to allow partial loop insertion and a patent lower b-sheet A that can act as a receptor for the loop of another molecule and hence form polymers [42,51]. This patent