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
Trang 1Sugar and alcohol molecules provide a therapeutic
strategy for the serpinopathies that cause dementia and cirrhosis
Lynda K Sharp1,*, Meera Mallya1,*, Kerri J Kinghorn1, Zhen Wang1, Damian C Crowther1,
James A Huntington2, Didier Belorgey1and David A Lomas1
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 a1-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 a1-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 a1-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 a1-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.
Trang 2generate 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 a1-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
a1-antitrypsin causes early-onset emphysema [29]
Moreover, intrahepatic polymerization of variants of
other serpins: C1 inhibitor, antithrombin and a1
-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 a1-antitrypsin [46] and
increased the secretion of the Z variant of a1
-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 a1-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 (n¼ 5, P ¼ 0.003) with 1.36 m (10% v ⁄ v) glycerol at
45C (Fig 2A) The longer sugar alcohol erythritol had a similar effect, reducing polymerization of wild-type neuroserpin by 2.8-fold (n¼ 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 45C (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).
Trang 3wild-type neuroserpin when added at a final
concentra-tion of 1.02 m The rate of polymerizaconcentra-tion 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 (n¼ 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
D C
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)1and 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.
Trang 4serpin However 10% w⁄ v Ficoll PM70 had a small
but significant effect on the polymerization of
wild-type neuroserpin (1.23· 10)5Æs)1 compared with
1.6· 10)5Æs)1 for the wild-type protein) but this was
still less than the effect of 10% v⁄ v glycerol
(1.01· 10)5Æs)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
45C by 1.7-fold compared with neuroserpin that
had been treated identically but which had not been
refolded in glycerol (n¼ 5, P ¼ 0.04) Refolding in
0.68 and 2.72 m glycerol had a very similar effect to
1.36 m glycerol (n¼ 6, P < 0.05), whereas refolding
in 0.14 m glycerol did not significantly alter the poly-merization rate of wild-type neuroserpin (n¼ 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⁄ Pi 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 45C by 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: ka ¼ 2.1 · 104m)1Æs)1 (n¼ 3), 1.1 · 104m)1Æs)1(n¼ 3) and 1.9 · 104m)1Æs)1 (n¼ 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 a1-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
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)1and are the mean and standard deviation of three independent experiments.
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.
Trang 5Glycerol, 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
45C) 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
(n¼ 4, P ¼ 0.02), 3.2-fold (n ¼ 4, P ¼ 0.009), 3.6-fold
(n¼ 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 37C (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 (P¼ 0.002), 2.3-fold (P ¼ 0.006), 2.6-fold (P ¼ 0.007) and 2.8-fold (P¼ 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.
Trang 6rate of polymerization of neuroserpin when incubated at
either 37 or 45C (almost twofold, see Table 2; P ¼
0.04, n¼ 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 (ka¼ 0.22 · 104m)1Æs)1, n¼ 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- (n¼
3, P¼ 0.04) and 1.6-fold (n ¼ 5, P ¼ 0.002) at 37 and
45C, 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 (ka¼ 0.19 · 104m)1Æs)1, n¼ 2)
Glycerol, erythritol, trehalose and glucose reduce
the polymerization and increase the transition
temperature of Z a1-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 a1-antitrypsin that also causes disease by
polymerization Glycerol is known to enhance the
secretion of Z a1-antitrypsin in a cell-culture model of
the disease [39] and similarly 1.36 m glycerol reduced
the polymerization of Z a1-antitrypsin by 2.9-fold at
41C (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 a1-antitrypsin
was also investigated It was found that 1.36 m
erythri-tol reduced the rate of polymerization of Z a1 -antitryp-sin fourfold at 41C (n ¼ 3, P ¼ 0.001; Fig 4A,C) and 5.3-fold when incubated at 37C (n ¼ 3, P ¼
0 0 1
2 2.5
0.2 0.5 1.5
0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
0.2
A
B
C
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 a1-antitrypsin when added to the polymerization buffer
Increas-ing concentrations of glycerol, erythritol, glucose or trehalose were
added to Z a1-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 a1-antitrypsin Z a1-Antitrypsin was incubated in NaCl ⁄ P i
at 0.1 mgÆmL)1and 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.
Trang 70.004) 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 a1-antitrypsin
at 41C 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
(n¼ 3, P ¼ 0.001) (Fig 4B) All these compounds
increased the transition temperature of Z a1-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 a1-antitrypsin However 10% w⁄ v Ficoll PM70
reduced the rate of polymerization of Z a1-antitrypsin
from 1.72· 10)6 to 1.01· 10)6Æs)1 (n¼ 3, P ¼ 0.015)
but this was less than the effect seen with 10% v⁄ v
gly-cerol (5.93· 10)7Æs)1)
Effect of refolding Z a1-antitrypsin with erythritol
or glucose
Two milligrams of Z a1-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
5 mm dithiothreitol, 50 mm Tris pH 7.8 and either
1.36 m erythritol or 0.68 m glucose All attempts to
refold Z a1-antitrypsin either with or without the
compounds were unsuccessful as Z a1-antitrypsin
spon-taneously formed polymers
Glycerol, erythritol, trehalose and glucose reduce
the rate of polymerization of Z a1-antitrypsin
even when removed from the polymerization
buffer
Purified Z a1-antitrypsin was incubated with 1.36 m
glycerol, 1.36 m erythritol, 0.68 m glucose or 0.68 m
trehalose at 4C for 1 h and then the Z a1-antitrypsin
was dialysed into NaCl⁄ Pi Pre-incubating Z a1
-anti-trypsin with glycerol, erythritol, trehalose or glucose
reduced the rate of polymerization at 41C by 1.9-fold
(n¼ 3, P¼ 0.016), 2.2-fold (n¼ 3, P¼ 0.010),
2.4-fold (n¼ 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 a1-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 61.1C, n ¼ 3,
P¼ 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 a1-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 a1-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 a1-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 a1-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 a1 -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
Trang 8plateau 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 a1-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 a1-antitrypsin Support
for this hypothesis comes from the demonstration that
merely flash-cooling antithrombin crystals with
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⁄ Pi 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 a1-antitrypsin as it immediately formed polymers Nevertheless, incubating any of these compounds with Z a1-antitrypsin for only 1 h (and then removal by dialysis into NaCl⁄ Pi) 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,
4 and 5 of b-sheet A.
Trang 9this 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
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 a1-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 associattenu-ated 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 a1-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
dis-rupted by sonication The cell pellet was washed three times
Tri-ton X-100 and then once more with buffer B alone before
guanidinium-HCl pH 8 The protein was refolded overnight
150 mm NaCl, pH 7.8) or buffer D (20 mm imidazole,
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
Trang 10Where specified, 1.36 m glycerol or erythritol was added for
1 h at the mixing stage A 300 mm imidazole solution in
bound protein, giving a single peak This fraction was diluted
and then loaded onto a HiTrap Q-Sepharose column and
eluted with a NaCl gradient (20 mm to 1 m) in 20 mm
had been extensively rinsed with distilled water to wash out
any residual glycerol The resulting protein was then
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 a1-antitrypsin and
refolding/incubation with compounds
homo-zygotes as described previously [37] and migrated as a single
band on SDS, nondenaturing and transverse urea gradient
in 50 mm Tris, 5 mm dithiothreitol pH 7.8 or in 50 mm Tris,
5 mm 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
1.36 m glycerol, 1.36 m erythritol, 0.68 m glucose or 0.68 m
assessed in assays of polymerization
Polymerization of wild-type neuroserpin,
Ser49Pro neuroserpin and Z a1-antitrypsin
Polymerization of wild-type neuroserpin, Ser49Pro
PAGE Wild-type or Ser49Pro neuroserpin were incubated
The samples were overlaid with oil to prevent evaporation
and 2 lg of protein for each time point was loaded onto a
added to the buffer at a range of concentrations (0–1.36 m)
Protease Inhibitor Cocktail Tablet for neuroserpin (Roche,
dilution) The proteins were visualized by staining with
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
Activity assays of serpins Rate constants for the inhibition of tPA by wild-type or Ser49Pro neuroserpin, and of bovine a-chymotrypsin by
sugars were determined as described previously [6,37]
Circular dichroism
J-810 spectropolarimeter Thermal unfolding experiments were performed by monitoring the CD signal at 216 nm
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
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
hydrogen bonds to s5A residues Superposition and analysis
of potential hydrogen bonding were conducted using the program xtalview, and figures were prepared using