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:
Website:
*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
Acknowledgements
This work was supported by the Medical Research
Council (UK), the Wellcome Trust (UK) and Pap-
worth NHS Trust. We are grateful to Professor R.W.
Carrell, Dr Aiwu Zhou and members of the Lomas
laboratory for helpful discussions.
References
1 Osterwalder T, Contartese J, Stoeckli ET, Kuhn TB &
Sonderegger P (1996) Neuroserpin, an axonally secreted
serine protease inhibitor. EMBO J 15, 2944–2953.
2 Schrimpf SP, Bleiker AJ, Brecevic L, Kozlov SV,
Berger P, Osterwalder T, Krueger SR, Schinzel A &
Sonderegger P (1997) Human neuroserpin (PI12):
cDNA cloning and chromosomal localization to 3q26.
Genomics 40, 55–62.
3 Krueger SR, Ghisu G-P, Cinelli P, Gschwend TP,
Osterwalder T, Wolfer DP & Sonderegger P (1997)
Expression of neuroserpin, an inhibitor of tissue
plasminogen activator, in the developing and adult ner-
vous system of the mouse. J Neurosci 17, 8984–8996.
4 Hastings GA, Coleman TA, Haudenschild CC, Stefans-
son S, Smith EP, Barthlow R, Cherry S, Sandkvist M &
Lawrence DA (1997) Neuroserpin, a brain-associated

inhibitor of tissue plasminogen activator is localized pri-
marily in neurones. J Biol Chem 272 , 33062–33067.
5 Osterwalder T, Cinelli P, Baici A, Pennella A, Krueger
SR, Schrimpf SP, Meins M & Sonderegger P (1998) The
axonally secreted serine proteinase inhibitor, neuroser-
pin, 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 polymers in vitro. J Biol Chem 277,
17367–17373.
7 Barker-Carlson K, Lawrence DA & Schwartz BS (2002)
Acyl–enzyme complexes between tissue type plasmino-
gen activator and neuroserpin are short lived in vitro.
J Biol Chem 277, 46852–46857.
8 Berger P, Kozlov SV, Krueger SR & Sonderegger P
(1998) Structure of the mouse gene for the serine pro-
tease inhibitor neuroserpin (PI12). Gene 214, 25–33.
9 Bradshaw CB, Davis RL, Shrimpton AE, Holohan PD,
Rea CB, Fieglin D, Kent P & Collins GH (2001) Cogni-
tive deficits associated with a recently reported familial
neurodegenerative disease: familial encephalopathy with
neuroserpin inclusion bodies. Arch Neurol 58, 1429–
1434.
10 Takao M, Benson MD, Murrell JR, Yazaki M, Pic-
cardo P, Unverzagt FW, Davis RL, Holohan PD, Law-
rence DA, Richardson R et al. (2000) Neuroserpin
mutation S52R causes neuroserpin accumulation in neu-
rons and is associated with progressive myoclonus epi-

lepsy. J Neuropathol Exp Neurol 59, 1070–1086.
11 Yepes M, Sandkvist M, Coleman TA, Moore E,
Wu JY, Mitola D, Bugge TH & Lawrence DA (2002)
Regulation of seizure spreading by neuroserpin and tis-
sue-type plasminogen activator is plasminogen-indepen-
dent. J Clin Invest 109, 1571–1578.
12 Yepes M, Sandkvist M, Wong MK, Coleman TA,
Smith E, Cohan SL & Lawrence DA (2000) Neuro-
serpin reduces cerebral infarct volume and protects
neurons from ischemia-induced apoptosis. Blood 96,
569–576.
13 Cinelli P, Madani R, Tsuzuki N, Vallet P, Arras M,
Zhao CN, Osterwalder T, Rulicke T & Sonderegger P
(2001) Neuroserpin, a neuroprotective factor in focal
ischemic stroke. Mol Cell Neurosci 18 , 443–457.
14 Silverman GA, Bird PI, Carrell RW, Coughlin PB, Get-
tins PG, Irving JI, Lomas DA, Luke CJ, Moyer RW,
Pemberton PA et al. (2001) The serpins are an expand-
ing superfamily of structurally similar but functionally
diverse proteins: evolution, mechanism of inhibition,
novel functions, and a revised nomenclature. J Biol
Chem 276, 33293–33296.
15 Huntington JA, Read RJ & Carrell RW (2000) Struc-
ture of a serpin–protease complex shows inhibition by
deformation. Nature 407, 923–926.
16 Davis RL, Holohan PD, Shrimpton AE, Tatum AH,
Daucher J, Collins GH, Todd R, Bradshaw C, Kent P,
Feiglin D et al. (1999) Familial encephalopathy with neu-
roserpin inclusion bodies. Am J Pathol 155, 1901–1913.
17 Davis RL, Shrimpton AE, Holohan PD, Bradshaw C,

Feiglin D, Collins GH, Sonderegger P, Kinter J, Becker
LM, Lacbawan F et al. (1999) Familial dementia caused
by polymerization of mutant neuroserpin. Nature 401,
376–379.
18 Davis RL, Shrimpton AE, Carrell RW, Lomas DA,
Gerhard L, Baumann B, Lawrence DA, Yepes M, Kim
TS, Ghetti B et al. (2002) Association between conform-
ational mutations in neuroserpin and onset and severity
of dementia. Lancet 359, 2242–2247.
19 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
seen in the dementia FENIB. Eur J Biochem 271, 3360–
3367.
20 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.
21 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 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS
mers within the endoplasmic reticulum. J Biol Chem
279, 28283–28291.
22 Lomas DA, Evans DL, Finch JT & Carrell RW (1992)
The mechanism of Z a

1
-antitrypsin accumulation in the
liver. Nature 357, 605–607.
23 Sivasothy P, Dafforn TR, Gettins PGW & Lomas DA
(2000) Pathogenic a
1
-antitrypsin polymers are formed
by reactive loop–b-sheet A linkage. J Biol Chem 275,
33663–33668.
24 Janciauskiene S, Dominaitiene R, Sternby NH, Piitulai-
nen E & Eriksson S (2002) Detection of circulating and
endothelial cell polymers of Z and wildtype alpha-1-
antitrypsin by a monoclonal antibody. J Biol Chem 277,
26540–26546.
25 Janciauskiene S, Eriksson S, Callea F, Mallya M, Zhou
A, Seyama K, Hata S & Lomas DA (2004) Differential
detection of PAS-positive inclusions formed by the Z,
Siiyama and Mmalton variants of a
1
-antitrypsin.
Hepatology 40, 1203–1210.
26 An JK, Blomenkamp K, Lindblad D & Teckman JH
(2005) Quantitative isolation of alphalAT mutant Z pro-
tein polymers from human and mouse livers and the
effect of heat. Hepatology 41, 160–167.
27 Eriksson S, Carlson J & Velez R (1986) Risk of cirrho-
sis and primary liver cancer in alpha
1
-antitrypsin defi-
ciency. N Engl J Med 314, 736–739.

28 Sveger T & Eriksson S (1995) The liver in adolescents
with a
1
-antitrypsin deficiency. Hepatology 22, 514–517.
29 Eriksson S (1965) Studies in a
1
-antitrypsin deficiency.
Acta Med Scand Suppl 432, 1–85.
30 Aulak KS, Eldering E, Hack CE, Lubbers YP, Harrison
RA, Mast A, Cicardi M & Davis AE 3rd (1993) A
hinge region mutation in C1-inhibitor (Ala436 fi Thr)
results in nonsubstrate-like behavior and in polymeriza-
tion of the molecule. J Biol Chem 268, 18088–18094.
31 Bruce D, Perry DJ, Borg JY, Carrell RW & Wardell
MR (1994) Thromboembolic disease due to thermola-
bile conformational changes of antithrombin Rouen-VI
(187 Asn fi Asp). J Clin Invest 94, 2265–2274.
32 Gooptu B, Hazes B, Chang WS, Dafforn TR, Carrell
RW, Read RJ & Lomas DA (2000) Inactive conforma-
tion of the serpin a
1
-antichymotrypsin indicates two-
stage insertion of the reactive loop: implications for
inhibitory function and conformational disease. Proc
Natl Acad Sci USA 97, 67–72.
33 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 multimeri-
zation. J Biol Chem 270, 2579–2587.
34 Lomas DA & Carrell RW (2002) Serpinopathies and

the conformational dementias. Nat Rev Genet 3, 759–
768.
35 Lomas DA & Mahadeva R (2002) a
1
-Antitrypsin poly-
merization and the serpinopathies: pathobiology and
prospects for therapy. J Clin Invest 110, 1585–1590.
36 Schulze AJ, Baumann U, Knof S, Jaeger E, Huber R &
Laurell C-B (1990) Structural transition of a
1
-antitryp-
sin by a peptide sequentially similar to b-strand s4A.
Eur J Biochem 194, 51–56.
37 Lomas DA, Evans DL, Stone SR, Chang WS & Carrell
RW (1993) Effect of the Z mutation on the physical
and inhibitory properties of a
1
-antitrypsin. Biochemistry
32, 500–508.
38 Sidhar SK, Lomas DA, Carrell RW & Foreman RC
(1995) Mutations which impede loop ⁄ sheet polymerisa-
tion enhance the secretion of human a
1
-antitrypsin defi-
ciency variants. J Biol Chem 270, 8393–8396.
39 Burrows JA, Willis LK & Perlmutter DH (2000) Chemi-
cal chaperones mediate increased secretion of mutant
a
1
-antitrypsin (a

1
-AT) Z: a potential pharmacological
strategy for prevention of liver injury and emphysema
in a
1
-AT deficiency. Proc Natl Acad Sci USA 97, 1796–
1801.
40 Chow MK, Devlin GL & Bottomley SP (2001) Osmo-
lytes as modulators of conformational changes in ser-
pins. Biol Chem 382, 1593–1599.
41 Devlin GL, Parfrey H, Tew DJ, Lomas DA & Bottom-
ley SP (2001) Prevention of polymerization of M and Z
a
1
-antitrypsin (a
1
-AT) with trimethylamine N-oxide.
Implications for the treatment of a
1
-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 patho-
genic serpin conformation and blocks polymerisation.
Implications for the prevention of Z a
1
-antitrypsin rela-
ted cirrhosis. J Biol Chem 277, 6771–6774.
43 Parfrey H, Mahadeva R, Ravenhill N, Zhou A, Dafforn
TR, Foreman RC & Lomas DA (2003) Targeting a sur-

face cavity of a
1
-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 a
1
-antitrypsin related
emphysema. Am J Respir Cell Mol Biol 31, 133–
139.
45 Zhou A, Stein PE, Huntington JA, Sivasothy P, Lomas
DA & Carrell RW (2004) How small peptides block
and reverse serpin polymerization. J Mol Biol 342, 931–
941.
46 Zhou A, Stein PE, Huntington JA & Carrell RW (2003)
Serpin polymerisation is prevented by a hydrogen-bond
network which is centred on His 334 and stabilised by
glycerol. J Biol Chem 278, 15116–15122.
47 Carrell RW & Lomas DA (1997) Conformational dis-
eases. Lancet 350, 134–138.
48 Tanaka M, Machida Y, Niu S, Ikeda T, Jana NR, Doi
H, Kurosawa M, Nekooki M & Nukina N (2004) Tre-
halose alleviates polyglutamine-mediated pathology in a
mouse model of Huntington disease. Nature Med 10,
148–154.
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 2551
49 Wenner JR & Bloomfield VA (1999) Crowding effects
on EcoRV kinetics and binding. Biophys J 77, 3234–

3241.
50 Uribe S & Sampedro JG (2003) Measuring solution
viscosity and its effect on enzyme activity. Biol Proceed
Online 5, 108–115.
51 Elliott PR, Lomas DA, Carrell RW & Abrahams J-P
(1996) Inhibitory conformation of the reactive loop of
a
1
-antitrypsin. Nat Struct Biol 3, 676–681.
52 Dafforn TR, Mahadeva R, Elliott PR, Sivasothy P &
Lomas DA (1999) A kinetic mechanism for the poly-
merisation of a
1
-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
a
1
-antitrypsin variant. Biochemical mapping of the deg-
radative event among compartments in the secretory
pathway. J Biol Chem 265, 14001–14007.
55 Dziegielewska KM, Evans CA, Malinowska DH,
Mollgard K, Reynolds JM, Reynolds ML & Saunders
NR (1979) Studies of the development of brain barrier
systems to lipid insoluble molecules in fetal sheep. J
Physiol 292, 207–231.

56 Zhou A, Faint R, Charlton P, Dafforn TR, Carrell RW
& Lomas DA (2001) Polymerization of plasminogen
activator inhibitor-1. J Biol Chem 276, 9115–9122.
57 Lawrence DA, Olson ST, Palaniappan S & Ginsburg D
(1994) Engineering plasminogen activator inhibitor 1
mutants with increased functional stability. Biochemistry
33, 3643–3649.
58 Dafforn TR, Della M & Miller AD (2001) The molecu-
lar interactions of heat shock protein 47 (Hsp47) and
their implications for collagen biosynthesis. J Biol Chem
276, 49310–49319.
59 Briand C, Kozlov SV, Sonderegger P & Gru
¨
tter MG
(2001) Crystal structure of neuroserpin: a neuronal ser-
pin involved in a conformational disease. FEBS Lett
25171, 1–5.
A therapeutic strategy for the serpinopathies L. K. Sharp et al.
2552 FEBS Journal 273 (2006) 2540–2552 ª 2006 The Authors Journal compilation ª 2006 FEBS