Design of new enzyme stabilizers inspired by glycosides of hyperthermophilic microorganisms

Article abstract of DOI:10.1016/j.carres.2008.08.030

In response to stressful conditions like supra-optimal salinity in the growth medium or temperature, many microorganisms accumulate low-molecular-mass organic compounds known as compatible solutes. In contrast with mesophiles that accumulate neutral or zw

Full text of DOI:10.1016/j.carres.2008.08.030

Carbohydrate Research 343 (2008) 3025–3033
Contents lists available at ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Design of new enzyme stabilizers inspired by glycosides
of hyperthermophilic microorganisms
Tiago Q. Faria a, Ana Mingote a, Filipa Siopa b, Rita Ventura b, Christopher Maycock b, Helena Santos a,
*
a Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa, Biology Division, Rua da Quinta Grande 6, Apartado 127, 2780-156 Oeiras, Portugal
b Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa, Chemistry Division, Rua da Quinta Grande 6, Apartado 127, 2780-156 Oeiras, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 7 May 2008
Received in revised form 14 August 2008
Accepted 31 August 2008
Available online 9 September 2008
In response to stressful conditions like supra-optimal salinity in the growth medium or temperature,
many microorganisms accumulate low-molecular-mass organic compounds known as compatible sol-
utes. In contrast with mesophiles that accumulate neutral or zwitterionic compounds, the solutes of
hyperthermophiles are typically negatively charged. (2R)-2-(a-D-Mannopyranosyl)glycerate (herein
abbreviated as mannosylglycerate) is one of the most widespread solutes among thermophilic and hyper-
thermophilic prokaryotes. In this work, several molecules chemically related to mannosylglycerate were
Keywords:
synthesized, namely (2S)-2-(1-O-
a
-D
-mannopyranosyl)propionate, 2-(1-O-a-D-mannopyranosyl)acetate,
Compatible solutes
Negatively charged solutes
Protein stabilization
Thermal inactivation
Protein aggregation
Mannosylglycerate
(2R)-2-(1-O- -glucopyranosyl)glycerate and 1-O-(2-glyceryl)-a-D
a
-
D
-mannopyranoside. The effectiveness
of the newly synthesized compounds for the protection of model enzymes against heat-induced denatur-
ation, aggregation and inactivation was evaluated, using differential scanning calorimetry, light scatter-
ing and measurements of residual activity. For comparison, the protection induced by natural compatible
solutes, either neutral (e.g., trehalose, glycerol, ectoine) or negatively charged (di-myo-inositol-1,30-phos-
phate and diglycerol phosphate), was assessed. Phosphate, sulfate, acetate and KCl were also included in
the assays to rank the solutes and new compounds in the Hofmeister series. The data demonstrate the
superiority of charged organic solutes as thermo-stabilizers of enzymes and strongly support the view
that the extent of protein stabilization rendered by those solutes depends clearly on the specific sol-
ute/enzyme examined. The relevance of these findings to our knowledge on the mode of action of charged
solutes is discussed.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
aggregation, with a number of afflicting human pathologies has
been widely documented in recent years.1–3 In this context, the
Enzymes and other proteins are used in many industrial and
pharmaceutical processes; they are also active components of
numerous food and clinical preparations. Therefore, the preserva-
tion of their native structure is a prerequisite for the usefulness
of the specific industrial process or preparation. Furthermore, the
association of structural instability, and consequent protein
development of reliable strategies to improve protein stability is
of great importance.3–6
Compatible solutes are low-molecular-mass organic com-
pounds that accumulate in the cytoplasm of many halophilic or
halotolerant organisms to counterbalance the osmotic pressure of
the external medium.7–9 They must be highly soluble and usually
belong to one of the following classes of compounds: amino acids,
carbohydrates, polyols, betaines and ectoines. Trehalose, glycerol,
glycine-betaine and ectoine are typical solutes of mesophiles. The
discovery of extreme thermophilic and hyperthermophilic (hereaf-
ter designated hyper/thermophilic) microorganisms in the early
1970s and 1980s, respectively, led to the identification of new
compatible solutes, highly restricted to organisms isolated from
hot habitats.10–12 In contrast with the neutral or zwitterionic nat-
ure of the solutes commonly found in mesophiles, the organisms
with optimal growth temperature above 60 °C accumulate mainly
negatively charged compounds, such as mannosylglycerate and di-
myo-inositol-1,30-phosphate. Furthermore, the concentration of
Abbreviations: MG or mannosylglycerate, (2R)-2-(
acid (potassium salt); ML or mannosyl-lactate, (2S)-2-(1-O -
yl)propionic acid (potassium salt); MGlyc or mannosylglycolate, 2-(1-O -
mannopyranosyl)acetic acid (potassium salt); GG or glucosylglycerate, (2R)-2-(
a
-
D
-mannopyranosyl)glyceric
a-D-mannopyranos-
a-
D-
a-
D-glucopyranosyl)glyceric acid (potassium salt); MGOH or mannosylglycerol, 1-O-
(2-glyceryl)-a-D-mannopyranoside; MGA or mannosylglyceramide, (2R)-2-(a-D-
mannopyranosyl)glyceramide; DIP, di-myo-inositol-1,30-phosphate (potassium
salt); DGP, diglycerol phosphate (potassium salt); HEct, hydroxyectoine; Ect,
ectoine; SNase, recombinant nuclease A from Staphylococcus aureus; MDH, mito-
chondrial malate dehydrogenase from pig heart; LDH, L-lactate dehydrogenase from
rabbit muscle; DSC, differential scanning calorimetry; Tm, melting temperature.
* Corresponding author. Tel.: +351 21 4469828; fax: +351 21 4428766.
E-mail address: santos@itqb.unl.pt (H. Santos).
0008-6215/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2008.08.030

3026
T. Q. Faria et al. / Carbohydrate Research 343 (2008) 3025–3033
these solutes increases not only with the NaCl concentration of the
medium but also with the growth temperature, leading to the view
that they play a role in the protection of proteins and other cellular
components against heat denaturation. In agreement with this
hypothesis, it has been shown that solutes from hyper/thermo-
philes exert superior thermoprotection on a variety of model en-
zymes.13–19
nopyranoside (MGOH) are structural variations of MG (Fig. 1). In
comparison with MG, the non-glycosidic group of ML lacks the pri-
mary hydroxyl group and the lactate has the opposite configura-
tion (S) to the glycerate (R); MGlyc does not possess the CH2OH
group and has lost the asymmetric centre; finally, the carboxylic
group in MG has been replaced by CH2OH to yield a neutral mole-
cule, MGOH.
Mannosylglycerate (MG) is one of the most widespread solutes
in hyper/thermophilic organisms.10 The efficacy of this natural sol-
ute in the protection of protein structures has been amply illus-
trated.15,17,18,20,21 In this work, several compounds structurally
related to MG were tested for their ability to stabilize model
proteins against thermal stress. The glyceric acid group in MG
was replaced by lactic acid, glycolic acid and glycerol groups; in
addition, mannose was replaced by glucose to yield glucosylglycer-
ate. The protecting effect of the synthetic compounds was com-
pared with that rendered by a number of charged and uncharged
natural organic solutes as well as inorganic salts. Different aspects
of the protein stabilization process, that is, the thermodynamic
stability, inhibition of aggregation and retention of the catalytic
activity, were assessed by using differential scanning calorimetry,
light scattering and residual activity measurements. The motiva-
tion behind this study was two-fold: to obtain new compounds
with improved protecting properties and to extend our knowledge
on the structural features that determine the degree of protein
stabilization rendered by a specific compound.
For the synthesis of (2S)-2-(1-O-
a-D-mannopyranosyl)propio-
nate (4, Scheme 1), the -mannose trichloroacetimidate derivative
a
2 was used as the glycosyl donor (Scheme 1). The glycosylation
reaction with (S)-methyl lactate was successfully accomplished
using BF3ꢀOEt2 in dichloromethane, and as expected only the
a-
anomer was obtained. Successive removal of the protecting groups
afforded 4 in good overall yield (53%).
The same strategy was employed to synthesize 2-(1-O-
mannopyranosyl)acetate (6, Scheme 2) and 1-O-(2-glyceryl)-
mannopyranoside (8, Scheme 3). Methyl glycolate was the glycosyl
acceptor in the synthesis of 6 and 1,3-di-tert-butyldiphenylsilyl-
glycerol was the glycosyl acceptor in the case of 8.
a
a
-
-
D
-
-
D
For the synthesis of (2R)-2-(1-O-a-D-glucopyranosyl)glycerate
(12,Scheme 4), a different approach was used since the beta isomer
is formed if participating groups, such as acetate, are present in the
2-position. The thioglycoside derivative 9 was used as the glucosyl
donor and it was obtained using a procedure described previ-
ously.22 The glycosylation reaction was successfully accomplished
employing Crich and Smith conditions.23 Successive removal of
protecting groups afforded the a-anomer 12 in 36% overall yield.
2. Results
2.2. Effect of solutes on protein structural stability as
monitored by DSC
2.1. Synthesis of compounds related to (2R)-2-(a-D-
mannopyranosyl)glycerate
The stabilizing properties of different natural solutes and syn-
thetic compounds on the protein structural stability were assessed
for three model enzymes. At the working pH, malate dehydro-
(2S)-2-(1-O-a-D-Mannopyranosyl)propionate (ML), 2-(1-O-a-D-
mannopyranosyl)acetate (MGlyc) and 1-O-(2-glyceryl)-
a-
D
-man-
OH
OH
OH
HO
HO
O
HO
O
O
HO
HO
HO
HO
HO
HO
OH
O
O
O
CO2
ML
OH
CO2
MG
OH
CO2
MGlyc
OH
HO
HO
O
O
O
HO
HO
HO
HO
HO
HO
OH
OH
OH
HO
O
O
O
GG
CO2
MGA
CONH2
MGOH
HO
O
P
HO
HO
HO
OH
OH
HO
OH
O
HO
O
OH
HO
O
O
OH
OH
O
P
OH
OH
O
O
DGP
DIP
OH
COOH
HN
HN
N
COOH
N
Ect
HEct
Figure 1. Chemical structures of the organic solutes whose stabilizing effect was studied. MG, mannosylglycerate; ML, mannosyl-lactate; MGlyc, mannosylglycolate; MGOH,
mannosylglycerol; MGA, mannosylglyceramide; GG, glucosylglycerate; DIP, di-myo-inositol-1,30-phosphate; DGP, diglycerol phosphate; Ect, ectoine; HEct, hydroxyectoine.

T. Q. Faria et al. / Carbohydrate Research 343 (2008) 3025–3033
3027
OH
CO Me
OAc
AcO
OH
HO
a,b,c
2
O
O
AcO
AcO
HO
HO
d
OH
O
NH
D-Mannose
2
1
CCl
3
OH
OAc
HO
AcO
O
e,f
O
HO
HO
AcO
AcO
O
O
4
CO K
CO Me
3
2
2
Scheme 1. Synthesis of (2S)-S-(1-O-a-D-mannopyranosyl)propionate 4 from mannose. Reagents and conditions: (a) Ac2O, pyr, DMAP, 0 °C/rt, 14 h, 99%; (b)
H2NNH2ꢀCH3COOH, DMF, rt, 3 h, 96%; (c) CCl3CN, CH2Cl2, DBU, rt, 2 h, 86%; (d) BF3ꢀOEt2, CH2Cl2, ꢁ78 °C/rt, 74%; (e) MeOH, MeONa 1 N, rt, 88%; (f) KOH 2 N, H2O.
OH
OAc
HO
AcO
HO
CO Me
2
O
b,c
O
2
HO
HO
AcO
AcO
a
O
CO K
2
O
CO Me
2
5
6
Scheme 2. Synthesis of 2-(1-O-
(c) KOH 2 M, H2O.
a-
D-mannopyranosyl)acetate 6 from mannose. Reagents and conditions: (a) BF3ꢀOEt2, CH2Cl2, ꢁ78 °C/rt, 69%; (b) MeOH, MeONa 1 M, rt, 85%;
OH
OH
OAc
AcO
TBDPSO
OTBDPS
HO
b,c
O
2
O
HO
HO
AcO
AcO
a
OH
OTBDPS
O
O
7
8
HO
TBDPSO
Scheme 3. Synthesis of 1-O-(2-glyceryl)-
MeOH, MeONa 1 M, rt, 90%.
a-
D-mannopyranoside 8 from mannose. Reagents and conditions: (a) BF3ꢀOEt2, CH2Cl2, ꢁ78 °C/rt, 85%; (b) TBAF, THF, rt, 96%; (c)
OAc
OAc
b
a
O
O
BnO
BnO
BnO
BnO
OH
OBn
BnO
SEt
O
TBDPSO
OTBDPS
CO Me
10
9
CO Me
2
2
OH
OH
c,d,e
O
O
BnO
BnO
HO
HO
BnO
OH
O
O
OTBDPS
CO Me
OH
CO K
12
11
2
2
Scheme 4. Synthesis of (2R)-2-(a-D-glucopyranosyl)glycerate 12 from glucose. Reagents and conditions: (a) PhSONC5H10, DTBMP, Tf2O, MS, CH2Cl2, 76%; (b) Na, MeOH, 0 °C,
94%; (c) TBAF, THF, 90%; (d) H2, Pd, 35 psi, 99%; (e) KOH 2 M, H2O rt.
genase, SNase and lysozyme have net charges of +3.9, +8.1 and
+8.8, and melting temperature (Tm) values in the absence of solutes
of 48.1, 53.9 and 72.9 °C, respectively. The increment in Tm induced
by the presence of the synthesis compounds and several natural
solutes of hyper/thermophiles and mesophiles is depicted in
Figure 2; the Tm values are presented in Table S1 (Supplementary
data).
Glycerol, ectoine, KCl and hydroxyectoine were poor stabilizers
of the enzymes examined; in contrast, DIP, ML, MG, GG and MGlyc
were highly effective. In particular, DIP and ML were excellent,
inducing increases in Tm greater than 10 °C per 0.5 M solute on
MDH and SNase, respectively. In comparison with MG, the synthe-
sis solute ML was clearly superior in the stabilization of SNase and
lysozyme and had similar effect on MDH. GG was better than MG

3028
T. Q. Faria et al. / Carbohydrate Research 343 (2008) 3025–3033
12
11
10
9
8
7
6
5
4
3
2
1
0
moy
Figure 2. Increment in the melting temperature of malate dehydrogenase, MDH (grey bars, red bars in online), staphylococcal nuclease, SNase (upward diagonal pattern,
green bars in online) and lysozyme (black bars, violet bars in online) in the presence of 0.5 M of different compounds. The melting temperature in the absence of solutes was
48.1 °C for MDH, 53.9 °C for SNase and 72.9 °C for lysozyme. ND, not determined.
for the stabilization of SNase, while MGlyc was similar to MG.
However, MG was slightly better than GG and MGlyc for the stabil-
ization of SNase. On the other hand, the neutral derivative, MGOH,
was substantially less efficient for all the enzymes examined.
The dependence of the increment of Tm (stabilization) on the
concentration of solute was studied using SNase and glycerol,
KCl, acetate, glycerate, sulfate, phosphate, MG and ML. The two lat-
ter solutes were excellent stabilizers of this enzyme, comparable to
the well-known kosmotropic agents, sulfate and phosphate; gly-
cerate and acetate showed a moderate beneficial effect, whereas
glycerol and KCl were poor stabilizers. In fact, while the concentra-
tion of glycerol required to increase the Tm by 3 °C was approxi-
mately 2.5 M, the same increment was produced by 0.15 M MG
or ML (Fig. 3).
2.3. Effect on protein aggregation as monitored by light
scattering
Thermodynamic data obtained by calorimetry were comple-
mented with light scattering measurements to assess the effective-
300
MGOH
250
200
72
Tre
Phosphate
HEct
ML
70
68
66
64
62
60
58
56
54
52
Sulfate
150
Ect
MG
NS
100
Glycerate
Acetate
KCl
Glycerol
Sulfate
Acetate
ML
50
0
MGlyc
Gly
DGP
KCl
MGA
MG
DIP
Pi
-10
0
10
20
30
40
50
60
70
80
90
100
Glycerol
2.5
MDH scattering (%)
Figure 4. Effect of different compounds at 0.5 M concentration upon malate
dehydrogenase (MDH) and lactate dehydrogenase (LDH) light scattering. 100% is
defined as the value measured in the absence of solutes after incubation of MDH
(0.2 mg/mL) or LDH (0.2 mg/mL) at the final temperature for 50 or 30 min,
respectively. Ionic compounds are represented by open circles (s) and non-charged
compounds are shown as open squares (h). Abbreviations: Pi, potassium phos-
phate; Gly, potassium glycerate; Tre, trehalose; other abbreviations as in the legend
of Figure 1.
0.0
0.5
1.0
1.5
2.0
3.0
Concentration (M)
Figure 3. Dependence of staphylococcal nuclease melting temperature on the
concentration of glycerol (–Ç–), KCl (–s–), acetate (–j–), glycerate (–h–), MG
(–.–), sulfate (–d–), ML (–}–) and phosphate (–4–). Abbreviations as in the legend
of Figure 1.

T. Q. Faria et al. / Carbohydrate Research 343 (2008) 3025–3033
3029
ness of solutes to prevent protein aggregation. In this case, LDH
and MDH were used as model enzymes. All the compounds exam-
ined reduced the aggregation of MDH but to different extents
(Fig. 4 and Table S2). The most efficient compounds, inducing a de-
crease of light scattering over 85%, were MG, MGA, DGP, DIP,
MGOH, phosphate, glycerate and acetate. In the case of LDH, a
few compounds, trehalose, ectoine, hydroxyectoine and MGOH,
appeared to promote aggregation (increased the light scattered),
while all the others reduced protein aggregation. DIP, MG and
phosphate were the best inhibitors of LDH aggregation and also ex-
erted excellent protection on MDH. Curiously, a few solutes had
opposite effects upon MDH or LDH; for example, MGOH was very
effective in reducing MDH aggregation, but increased considerably
the scattering observed with LDH. On the other hand, ML and glyc-
erol prevented aggregation of both enzymes but to a modest ex-
tent. The effect of MG concentration on LDH aggregation was
also studied. In the presence of 0.1 M MG, the scattering was only
slightly lower than in the assay without solutes (84% vs 100%).
However, with 0.25 M MG the light scattering was reduced to
37%, and 0.5 M MG led to a strong suppression of scattering (4%).
trend emerging from the large data set presented. This trend is
even more explicit in a plot of the increment in the Tm of MDH ver-
sus those of SNase and lysozyme, which shows that ionic com-
pounds (except for KCl) cluster in the upper-right corner of the
graph, while non-charged compounds appear in the lower-left cor-
ner (Fig. 5). This general pattern is also corroborated by the light
scattering experiments that reveal a good correlation between
charge and the ability to prevent protein aggregation (Fig. 4); curi-
ously, the uncharged solute, mannosylglyceramide, is a notable
exception, inhibiting protein aggregation comparable to mannosyl-
glycerate, a closely related charged solute, and conferring a much
greater protection than mannosylglycerol, a related uncharged
compound. On the other hand, in respect to structural stabilization,
mannosylglyceramide falls well in the low performance region,
typical of neutral solutes (Fig. 5).
Considering the simpler phenomenon of structural stabilization
and the subset of charged compounds, it is clear that the extent of
stabilization upon a given protein varies with the structure of the
solute, but the scatter observed is difficult to rationalize in mole-
cular terms. Potassium glycerate is significantly less efficient upon
SNase than the related compounds mannosylglycerate and gluco-
sylglycerate, suggesting that the presence of the sugar moiety
plays an important role. However, the discrepancy between the
increment in Tm induced by mannosylglycerate and glucosylglycer-
ate is difficult to explain, given the high similarity of the mannose
and glucose groups (Fig. 2). In another example, the replacement of
a CH2OH group in mannosylglycerate by a CH3 as in mannosyl-
lactate had a considerable beneficial impact on the stabilization
of SNase and lysozyme, but not of MDH. These findings also
indicate that attempts to correlate the stabilizing effect with the
polarity of the solute are probably too simplistic.
2.4. Effect of solutes on the protection of MDH against thermal
inactivation
When a protein is subjected to heat stress, processes like dena-
turation, aggregation, oxidation of some amino acid residues,
reduction of disulfide bonds, deamidation of glutamine or aspara-
gine residues and hydrolysis of peptide bonds may occur, which
lead to the irreversible loss of function.24 The time-course of
MDH inactivation was studied to evaluate the inhibitory effect of
the different compounds on those irreversible processes (Table
1). Sulfate and MG exerted a high protective effect on MDH,
increasing the half-life for inactivation more than 24- and 18-fold,
respectively. In contrast, DIP and hydroxyectoine had a deleterious
effect, inducing a decrease of approximately 50% in the half-life for
MDH inactivation. Unexpectedly, ML, MGlyc and glycerate totally
inhibited MDH activity, thereby precluding further investigation
on their effect on the thermal inactivation profile.
The Hofmeister series has been used as a ranking of ions based
on their ‘salting-out’ ability. Initially, this series was based on the
minimum concentration of salts required to precipitate egg white
proteins.25–27 However, the accumulation of data on other fields
of protein science has broadened the scope of application;28 for
example, the influence of ions on protein thermal unfolding
appears to follow the Hofmeister series.29,30 Our results show
that chloride < acetate < sulfate ꢂ phosphate in respect to their
efficiency to increase the Tm of SNase, and this order agrees well
3. Discussion
In the search for better protein stabilizers, mannosylglycerate, a
very efficient protector widely distributed in organisms thriving in
hot environments, was used as a lead compound for the design of
new molecules. The effectiveness of the newly synthesized com-
pounds in the protection of model enzymes against heat-induced
denaturation, aggregation and inactivation was evaluated. The
greater stabilization rendered by charged compounds is a clear
12
10
8
8
6
4
2
0
6
4
Table 1
Half-life for thermal inactivation of pig heart mitochondrial malate dehydrogenase in
the presence of different compounds (at a final concentration of 0.5 M)
2
Half-life (min)
0
Hydroxyectoine
Di-myo-inositol-1,30-phosphate
Trehalose
No solute
Glycerol
1.6 0.3
1.6 0.6
2.6 0.2
3.2 0.6
3.9 0.3
4.3 0.3
8.7 0.1
13.2 0.3
32.5 5.0
50.2 8.2
53.0 1.7
56.5 4.0
72.8 11.3
0
2
4
6
8
10
12
ΔTm of MDH (°C)
Figure 5. Stabilizing effect of different compounds against thermal denaturation of
MDH, SNase and lysozyme. In the abscissa axis, the increment in the melting
temperature of MDH induced by 0.5 M of the various compounds, and in the
ordinates axis the increment in the melting temperature of SNase (solid symbols)
and lysozyme (open symbols) are plotted. The increment value is defined as the
difference between the melting temperature in the presence of compound and the
value without addition. Ionic compounds are represented by circles (d,s) and non-
charged compounds by squares (j,h). Abbreviations as in the legend of Figures 1
and 4.
Ectoine
KCl
Mannosylglyceramide
Potassium phosphate
Diglycerol phosphate
Potassium acetate
Mannosylglycerate
Sodium sulfate

3030
T. Q. Faria et al. / Carbohydrate Research 343 (2008) 3025–3033
with the Hofmeister series, suggesting that the increment in the Tm
of SNase might be useful to infer the position of the hyper/thermo-
philes solutes and of the synthetic compounds in the Hofmeister
series. In this context, the order of efficiency found for SNase,
Clꢁ < acetate < glycerate < MG ꢂ MGlyc < DGP < sulfate ꢂ DIP < GG ꢂ
phosphate < ML illustrates the point that the solutes from hyper/
thermophiles appear close to the best stabilizing ions of the Hofmei-
ster series, that is, sulfate and phosphate; in addition, the synthetic
compounds ML and GG are placed at the head of the series, being
better or comparable to phosphate. It should be pointed out,
however, that the order of efficiency depends on the enzyme studied,
and for MDH the following series was found: Clꢁ < acetate <
DGP < phosphate < MGlyc < GG ꢂ ML ꢂ sulfate < MG < DIP. In this
case, phosphate is significantly worse than sulfate, and MG and DIP
are better than the synthetic compounds.
25
20
15
10
5
120
100
80
60
40
20
0
0
If this analysis is extended to include the data on the prevention
of protein aggregation, the salts commonly represented in the Hof-
meister series are ordered as Clꢁ < sulfate < acetate < phosphate,
with the remarkable inversion of positions between sulfate and
acetate. In regard to the solutes of hyper/thermophiles, DIP and
MG fall into the front region regardless of the enzyme examined,
but the synthetic compounds had very distinct effects upon MDH
and LDH, thus precluding their positioning in the effectiveness
ranking. As a general conclusion of the calorimetric and light scat-
tering experiments, the natural solutes of hyper/thermophiles
(MG, DIP and DGP) had stabilizing abilities similar to those of the
most efficient inorganic ions, while the ranking of the newly syn-
thesized compounds depends on the assay examined: from the
melting temperature experiments ML can be placed among the
best stabilizers while it has a poor performance as inhibitor of pro-
tein aggregation. It seems as though an increment in structural sta-
bility does not translate necessarily into a decrease in protein
aggregation, probably due to the distinct nature of the two pro-
cesses: protein aggregation is controlled at the kinetic level, while
Tm is a thermodynamic property.
In practical terms, enzyme thermostabilization should be pri-
marily assessed from the ability to preserve activity, that is, from
the inactivation profiles at the operational temperature. Therefore,
we deemed it interesting to correlate structural stabilization (Tm
values), inhibition of protein aggregation and half-life values for
inactivation (Fig. 6). If the abnormal behaviour of DIP is ignored
for the moment (see below), there is a good correlation between
the increment of Tm and the operational stabilization. In fact, the
extra structural stabilization rendered by a solute (Tm increment)
is clearly more important than its ability to prevent aggregation.
The patterns of MGA and KCl are good illustrations of this view:
both compounds inhibited MDH aggregation to a large extent,
yet their protection against thermal inactivation was modest. On
the other hand, the natural solutes MG and DGP were excellent
activity protectors, contrasting with the peculiar destabilizing
effect of DIP, the solute that induced the greatest increase in Tm.
In fact, a similar destabilizing effect has been observed for DIP on
a different enzyme, LDH.17 A possible explanation could be the
ability of DIP to stabilize non-productive oligomeric forms of the
enzymes. The reasons for the unexpected inhibition of MDH by
the synthetic compounds ML, MGlyc and GG (results not shown)
remain elusive. A possible negative effect of a contaminant is ruled
out by the high purity of the compounds used in the assay.
In conclusion, besides a sound demonstration of the superiority
of charged organic solutes to act as thermo-stabilizers of enzymes,
this work strongly supports the view that the extent of protein
stabilization rendered by organic solutes depends clearly on the
specific solute/enzyme examined. Moreover, the magnitude of
the stabilizing effect also depends on the particular step of the
inactivation process that is targeted in the assays, for example,
denaturation, as assessed by Tm measurements, or aggregation, as
Figure 6. Comparison between the effect of several compounds on MDH half-life
for thermal inactivation (upward diagonal pattern, green bars in online), increment
in the melting temperature (black bars, orange bars in online) and aggregation
inhibition (grey bars, blue bars in online). Thermal inactivation values are depicted
as the ratio between the half-life value in the presence of solutes and that without
additions. The increment in Tm is defined as the difference between the Tm values
with and without solutes. HEct, hydroxyectoine; DIP, di-myo-inositol-1,30-phos-
phate; Tre, trehalose; Ect, ectoine; MGA, mannosylglyceramide; DGP, diglycerol
phosphate; MG, mannosylglycerate.
assessed by light scattering. These data reinforce the concept that
the mechanisms underlying stabilization by charged solutes may
depend to a greater extent on specific protein/solute interactions
than on general alterations of the water properties that seem to
characterize the mode of action of neutral solutes.31–33
Our efforts to produce better solutes were reasonably success-
ful, if we consider the complexity of the mechanisms involved in
enzyme protection by organic solutes. In fact, mannosyl-lactate
was the best stabilizer of SNase and lysosyme in regard to thermal
unfolding. Interestingly, the best wide-ranging protector was
mannosylglycerate, a canonical solute among hyper/thermophiles.
On a physiological point of view, it is notable that hyper/thermo-
philic cells have evolved to hold high levels of negatively charged
solutes, a feature that enables them to benefit from the most effec-
tive stabilizers against heat stress.
4. Experimental
4.1. Materials
Mannosylglycerate (MG), mannosylglyceramide (MGA), digly-
cerol phosphate (DGP), di-myo-inositol-1,30-phosphate (DIP),
hydroxyectoine (HEct) and ectoine (Ect) were obtained from bitop
AG, Witten, Germany. Potassium glycerate was prepared from
hemicalcium glycerate from Sigma. Calcium was removed on a
H+-activated Dowex column (Bio-Rad) and ultra-pure potassium
hydroxide (FIXANAL from Riedel-de Haën AG) was used to titrate
the glyceric acid solution. Purity and concentration of the com-
pounds was assessed by 1H NMR spectra recorded on a Bruker
AMX300 spectrometer with a 5 mm inverse probe head. For quan-
tification purposes, spectra were acquired with a repetition delay
of 60 s with formate as concentration standard. Only samples with
purity higher than 98% were used. Trehalose, potassium chloride,
glycerol, sodium and potassium phosphate, sodium sulfate and
potassium acetate were of the highest purity available. Mitochon-
drial malate dehydrogenase from pig heart (MDH) and rabbit mus-
cle L-lactate dehydrogenase (LDH) were purchased from Roche;
hen egg white lysozyme was from Fluka. These enzymes were used
without further purification. Recombinant staphylococcal nuclease

T. Q. Faria et al. / Carbohydrate Research 343 (2008) 3025–3033
3031
A (SNase) was produced and purified from Escherichia coli cells as
described by Faria et al.21 Protein concentration was determined
from UV absorbance at 280 nm, using 0.28 (mg/mL)ꢁ1 cmꢁ1 for
the extinction coefficient of MDH,34 1.49 (mg/mL)ꢁ1 cmꢁ1 for
LDH,35 2.64 (mg/mL)ꢁ1 cmꢁ1 for lysozyme36 and 0.93 (mg/
mL)ꢁ1 cmꢁ1 for SNase.37
of KOH 2 M (8.0 mL). The reaction was stirred for 4 h. The pH was
adjusted to neutral with HCl 10% and the solvent evaporated to
afford 4 (4.68 g, potassium salt) as a colourless viscous foam, which
was purified as described above. ½a D20
ꢄ
+30.5 (c 2.2, H2O). 1H NMR
(D2O): d 4.86 (1H, d, J = 1.5 Hz), 4.06 (1H, q, J = 6.8 Hz), 3.89 (1H,
dd, J = 3.4 Hz, J = 1.7 Hz), 3.81 (1H, dd, J = 8.6 Hz, J = 3.2 Hz), 3.70–
3.57 (4H, m), 1.29 (3H, d, J = 6.8 Hz). 13C NMR (D2O): d 182.2,
101.9, 77.0, 75.7, 73.1, 72.9, 69.1, 63.2, 20.0. FT-IR 1587.5 cmꢁ1
(C@O).
4.2. Chemical synthesis
Mannosyl-lactate (ML), mannosylglycerol (MGOH), mannosyl-
glycolate (MGlyc) and glucosylglycerate (GG) were obtained by
chemical synthesis. All the reactions were carried out under argon
and stirring, unless otherwise stated. Preparative TLC was run on
silica gel Merck 60 GF254 plates and analytical TLC on alumin-
ium-backed silica gel Merck 60 GF254 plates. Silica Gel 60H
was used for the medium pressure column chromatography.
Tetrahydrofuran (THF) was dried by distillation over Na/benzo-
phenone under argon and dry CH2Cl2 was obtained by distillation
over phosphorus pentoxide. Other solvents and reagents were
distilled as required. The reaction products were analyzed by
NMR; spectra were recorded in CDCl3 and D2O depending on
the solubility of the products, at 300 MHz (1H NMR) and
75 MHz (13C NMR). The desired compounds were purified by an-
ion exchange chromatography on a QAE-Sephadex A25 column
(Amersham-Pharmacia Biotech) eluted with sodium carbonate/
bicarbonate buffer, pH 9.8 (from 5 mM to 1 M concentration).
The fractions containing the pure compounds were pooled,
lyophilized and loaded onto a H+-activated DOWEX 50W-X8
column (Bio-Rad) eluted with water. The fractions containing
the compounds were pooled, degassed under vacuum and the
pH was adjusted to 4.5–5.0 with ultra-pure potassium hydroxide
(FIXANAL from Riedel-de Haën AG). Purity and quantification
were evaluated as described above.
4.2.3. Methyl 2-(2,3,4,6-tetra-O-acetyl-1-O-a-D-
mannopyranosyl)acetate 5
To a solution of 2 (4.18 g, 8.5 mmol) in CH2Cl2 (20 mL) was
added methyl glycolate (0.741 g, 10.2 mmol). The solution was
cooled to ꢁ20 °C and BF3ꢀOEt2 (1.0 mL, 8.5 mmol) was slowly
added. The reaction mixture was stirred, while the temperature
was allowed to rise, until no starting material was detected by
TLC. Saturated aqueous NaHCO3 solution (20 mL) was added and
the aqueous phase was extracted with CH2Cl2 (3 ꢃ 20 mL). The
combined organic phases were dried (MgSO4) and concentrated.
The residue was purified by column chromatography (AcOEt/hex-
ane 2:8 to 4:6) to afford 5 (2.33 g, 69%) as a colourless viscous
foam. ½a 2D0
ꢄ
+55.1 (c 0.82, CH2Cl2). 1H NMR (CDCl3): d 5.40–5.27
(3H, m), 4.95 (1H, s), 4.31–4.26 (2H, m), 4.20–4.13 (2H, m), 4.10
(1H, dd, J = 12.1 Hz, J = 2.2 Hz), 3.77 (3H, s), 2.16 (3H, s), 2.10 (3H,
s), 2.05 (3H, s), 2.00 (3H, s). 13C NMR (CDCl3): d 170.6, 170.0,
169.8, 169.7, 169.6, 169.5, 97.9, 69.2, 69.1, 68.8, 65.9, 64.6, 62.3,
52.0, 20.8, 20.7, 20.6, 20.5. FT-IR 1750.9 cmꢁ1 (C@O).
4.2.4. 2-(1-O-a-D-Mannopyranosyl)acetic acid, potassium salt 6
To a solution of 5 (4.7 g, 11 mmol) in dried MeOH (40 mL) was
added a solution of MeONa in MeOH (1 N, 7.6 mL) at rt. After 3 h,
the solution was diluted with MeOH, and Dowex 50 WX8 resin
was added until neutral pH. Filtration of the resulting suspension
and evaporation of the solvents afforded a viscous residue, which
was diluted with water (10 mL) and washed with CH2Cl2 twice
(2 ꢃ 5 mL). The aqueous phase was concentrated under vacuum
4.2.1. Methyl (2S)-2-(2,3,4,6-tetra-O-acetyl-1-O-a-D-
mannopyranosyl)propionate 3
To a solution of trichloroacetimidate 238 (15 g, 30 mmol) in
CH2Cl2 (60 mL) was added (S)-methyl lactate (3.49 mL, 36 mmol).
The solution was cooled to ꢁ20 °C and BF3ꢀOEt2 (3.90 mL,
30 mmol) was slowly added. The reaction mixture was stirred,
while the temperature was allowed to rise to 0 °C, for 2 h (no start-
ing material was detected by analytical TLC). Saturated aqueous
NaHCO3 solution (50 mL) was added followed by extractions with
CH2Cl2 (3 ꢃ 60 mL). The combined organic phases were dried
(MgSO4) and concentrated. The residue was purified by column
chromatography (AcOEt/hexane 2:8) to afford 3 (9.80 g, 74%) as a
to furnish methyl 2-(1-O-a-D-mannopyranosyl) acetate (2.22 g,
80%) as a colourless viscous foam. To a solution of this intermedi-
ate (2.22 g, 8.8 mmol) in H2O (10.2 mL), was added a 2 M KOH
aqueous solution (4.5 mL). The pH was adjusted to 7 with HCl
1 M and the solvent evaporated under vacuum to afford
6
(2.07 g) as a colourless viscous foam, which was purified as de-
scribed above. ½a D20
ꢄ
+60.4 (c 1.88, H2O). H NMR (D2O): d 4.81
1
(1H, d, J = 1.6 Hz), 4.07 (1H, d, J = 16 Hz), 3.99 (1H, d, J = 16 Hz),
3.97–3.95 (1H, m), 3.82–3.79 (3H, m), 3.70–3.66 (2H, m), 3.59–
3.57 (3H, m). 13C NMR (D2O): d 176.4, 99.6, 72.9, 70.4, 69.9, 66.7,
65.2, 60.8. FT-IR 1594.1 cmꢁ1 (C@O).
colourless viscous foam. ½a D20
ꢄ
+22.8 (c 1.91, CH2Cl2). 1H NMR
(CDCl3): d 5.35–5.26 (3H, m), 4.90 (1H, d, J = 1.3 Hz), 4.33–4.24
(2H, m), 4.17 (1H, q, J = 6.8 Hz), 4.00 (1H, dd, J = 12.1 Hz,
J = 1.8 Hz), 3.75 (3H, s), 2.16 (3H, s), 2.10 (3H, s), 2.05 (3H, s),
2.00 (3H, s), 1.44 (3H, d, J = 6.8 Hz). 13C NMR (CDCl3): d 172.4,
170.6, 170.0, 169.8, 169.7, 98.3, 74.6, 69.7, 69.0, 68.9, 65.8, 62.0,
52.1, 20.8, 20.7, 20.6, 20.5, 18.2. FT-IR 1746.6 cmꢁ1 (C@O).
4.2.5. 2,3,4,6-Tetra-O-acetyl-1-O-(1,3-O-di-tert-
butyldiphenylsilyl-2-glyceryl)-a-D-mannopyranoside 7
To a solution of 2 (25 g, 0.051 mmol) in CH2Cl2 (100 mL) was
added 1,3-di-O-tert-butyldiphenylsilylglycerol (34.2 g, 0.061 mol).
The solution was cooled to ꢁ20 °C and BF3ꢀOEt2 (6.7 mL,
0.051 mol) was slowly added. The reaction mixture was stirred,
while the temperature was allowed to rise to 0 °C, for 2 h. Satu-
rated aqueous NaHCO3 solution (60 mL) was added followed by
extractions with CH2Cl2 (3 ꢃ 80 mL). The combined organic phases
were dried (MgSO4) and concentrated. The residue was purified by
column chromatography (AcOEt/hexane 1:9 to 2:8) to afford 7
4.2.2. (2S)-2-(1-O-a-D-Mannopyranosyl)propionic acid,
potassium salt 4
To a solution of 3 (8.00 g, 18 mmol) in MeOH (60 mL) was added
MeONa (1 N in MeOH, 12.4 mL) at rt. After 3 h, all the starting
material had been consumed. The solution was diluted with MeOH,
and Dowex 50 WX8 resin was added until pH 6. Filtration and
evaporation of the solvents afforded a viscous residue, which was
diluted with water (10 mL) and washed with CH2Cl2 twice
(2 ꢃ 10 mL). The aqueous phase was concentrated to afford the
(38.8 g, 85%) as a colourless viscous foam. ½a D20
ꢄ
+26.1 (c 2.99,
CH2Cl2). 1H NMR (CDCl3): d 7.65–7.59 (8H, m), 7.44–7.32 (12H,
m), 5.40 (1H, dd, J = 10.0 Hz, J = 3.4 Hz), 5.35–5.25 (2H, m), 5.11
(1H, s), 4.21–4.11 (2H, m), 4.01 (1H, m), 3.89 (1H, dd, J = 12.2 Hz,
J = 1.8 Hz), 3.72 (4H, t, J = 5.9 Hz), 2.12 (3H, s), 2.04 (3H, s), 1.99
(3H, s), 1.96 (3H, s), 1.02 (9H, s), 0.99 (9H, s). 13C NMR (CDCl3): d
crude methyl (2S)-2-(1-O-a-D-mannopyranosyl)propionate as a
very colourless viscous foam. To a solution of this compound
(4.27 g, 15.4 mmol) in H2O (16 mL) was added an aqueous solution

3032
T. Q. Faria et al. / Carbohydrate Research 343 (2008) 3025–3033
170.7, 169.8, 169.7, 169.6, 135.5, 135.4, 133.1, 133.0, 132.9, 129.8,
129.7, 127.8, 127.8, 127.7, 96.6, 77.7, 69.5, 69.3, 68.2, 65.9, 63.8,
63.1, 62.2, 26.8, 26.7, 20.8, 20.7, 20.6, 19.1. FT-IR 1752.6 cmꢁ1
(C@O).
(7.5 mL) at 0 °C. After 30 min, saturated aqueous NH4Cl (15 mL)
was added. The aqueous phase was extracted with AcOEt
(3 ꢃ 20 mL), and the combined organic extracts were dried
(MgSO4), filtered and the solvent removed. Purification by medium
pressure column chromatography (10:90 to 20:80 AcOEt/hexane)
afforded the product 11 as very colourless viscous foam (1.79 g,
4.2.6. 1-O-(2-Glyceryl)-a-D-mannopyranoside 8
To a solution of 7 (35 g, 0.039 mol) in THF (120 mL) was added
TBAF (25 g, 0.098 mol) at rt. After all the starting material had been
consumed, H2O (60 mL) was added. The aqueous phase was ex-
tracted with AcOEt (3 ꢃ 80 mL), the combined organic phases were
dried (MgSO4) and then concentrated. The obtained residue was
used in the next reaction without further purification. In order to
characterize this compound, only a small sample was purified by
preparative TLC (50:50 AcOEt/hexane), and 2,3,4,6-tetra-O-acetyl-
94%). ½a 2D0
ꢄ
+17.2 (c 1.49, CH2Cl2). 1H NMR (CDCl3): d 7.70–7.67
(5H, m), 7.44–7.24 (25H, m), 5.14 (1H, d, J = 3.5 Hz), 4.99 (1H, d,
J = 10.8 Hz), 4.90 (1H, d, J = 11.5 Hz), 4.88 (1H, d, J = 11.2 Hz), 4.78
(1H, d, J = 10.1 Hz), 4.71 (1H, d, J = 11.7 Hz), 4.62 (1H, d,
J = 11.3 Hz), 4.46 (1H, dd, J = 6.1 Hz, J = 4.3 Hz), 4.13–3.95 (3H, m),
3.35–3.79 (1H, m), 3.74 (3H, s), 3.70–3.45 (5H, m), 1.03 (9H, s).
13C NMR (CDCl3): d 170.3, 138.8, 138.4, 138.2, 135.6, 135.5,
133.0, 132.8, 129.7, 128.3, 128.1, 128.0, 127.7, 127.6, 127.5, 94.9,
81.5, 79.3, 76.9, 75.7, 74.8, 74.7, 72.1, 71.2, 64.6, 61.6, 51.9, 26.7,
19.1. FT-IR 1752.8 cmꢁ1 (C@O). HR-MS: Calcd for C47H54O9Si
[M]+: 790.3531; found: 790.3500.
1-O-(2-glyceryl)-a-D-mannopyranoside (96%) was isolated as a
colourless viscous foam. ½a D20
ꢄ
+33.8 (c 0.66, CH2Cl2). 1H NMR
(CDCl3): d 5.35 (1H, dd, J = 10.0 Hz, J = 3.3 Hz), 5.31–5.26 (2H, m),
5.05 (1H, d, J = 1.6 Hz), 4.25–4.15 (3H, m), 3.82–3.73 (5H, m) 2.16
(3H, s), 2.11 (3H, s), 2.06 (3H, s), 2.01 (3H, s). 13C NMR (CDCl3): d
170.7, 170.2, 170.1, 169.7, 96.5, 81.2, 69.7, 69.0, 66.2, 63.0, 62.7,
62.1, 20.8, 20.7. FT-IR 1745.6 cmꢁ1 (C@O). To a solution of the pre-
vious compound (16 g, 0.038 mol) in MeOH (50 mL) was added
MeONa (1 N in MeOH, 24.7 mL) at rt. After 3 h, all the starting
material had been consumed. The solution was diluted with MeOH,
and Dowex 50 WX8 resin was added until pH 6. Filtration and
evaporation of the solvents afforded a viscous residue, which was
diluted with water (20 mL) and washed with CH2Cl2 twice
(2 ꢃ 10 mL). The aqueous phase was concentrated and furnished
8 (9.6 g, 90%) as a colourless viscous foam, which was purified as
4.2.9. (2R)-2-(a-D-Glucopyranosyl)glycerate, potassium salt 12
To a solution of 11 (3.00 g, 3.8 mmol) in THF (15 mL) at rt was
added Bu4NF (1.20 g, 4.6 mmol). The reaction mixture was stirred
for 3 h and then water was added. The mixture was extracted with
AcOEt (3 ꢃ 20 mL), dried (MgSO4) and concentrated to furnish a
yellow viscous residue. Purification by medium pressure column
chromatography (50:50 to 100% AcOEt) afforded the product
methyl
(2R)-2-(2,3,4-tri-O-benzyl-a-D-glucopyranosyl)glycerate
as a very colourless viscous foam (1.89 g, 90%). ½a D20
ꢄ
+49.5 (c
1.16, CH2Cl2). 1H NMR (CDCl3): d 7.43–7.26 (18H, m), 5.20 (1H, d,
J = 3.6 Hz), 5.00 (1H, d, J = 10.8 Hz), 4.92–4.80 (3H, m), 4.70 (1H,
d, J = 11.4 Hz), 4.62 (1H, d, J = 10.8 Hz), 4.36 (1H, t, J = 4.3 Hz),
4.03 (1H, t, J = 9.3 Hz), 3.93 (2H, d, J = 4.5 Hz), 3.76 (3H, s), 3.75–
3.52 (4H, m), 2.09 (2H, br s). 13C NMR (CDCl3): d 170.1, 138.6,
137.9, 137.7, 128.5, 128.4, 128.3, 127.9, 94.9, 81.3, 79.4, 77.1,
75.7, 75.1, 74.6, 72.4, 71.6, 63.3, 61.7, 52.1. HR-MS: Calcd for
C31H36O9 [M]+: 552.2354; found: 552.2333. Methyl (2R)-2-(2,3,4-
described above. ½a D20
ꢄ
+30.3 (c 1.5, H2O). 1H NMR (D2O): d 4.89
(1H, d, J = 1.6 Hz), 3.84 (1H, q, J = 1.8 Hz), 3.74–3.48 (10H, m). 13C
NMR (D2O): d 99.3, 78.0, 72.9, 70.4, 70.2, 66.7, 61.3, 60.9, 60.1.
FT-IR 3413 cmꢁ1 (O–H).
4.2.7. Methyl 3-tert-butyldiphenylsilyl-(2R)-2-(6-O-acetyl-2,3,4-
tri-O-benzyl-
a
-D
-glucopyranosyl)glycerate 10
tri-O-benzyl-a-D-glucopyranosyl)glycerate (1.89 g, 3.4 mmol) in
To a stirred solution containing the thioglycoside 922 (2.86 g,
5.3 mmol), 1-benzenesulfinyl piperidine (BSP, 1.12 g, 5.3 mmol),
2,6-di-tert-butyl-4-methylpyridine (DTBMP, 2.14 g, 10.6 mmol)
and powdered 4 Å sieves in CH2Cl2 (25 mL) at ꢁ78 °C was added
Tf2O (1.00 mL, 5.8 mmol). After 5 min, a solution of methyl (2R)-
3-tert-butyldiphenylsilylglycerate (2.86 g, 7.9 mmol) in CH2Cl2
(10 mL) was added. The reaction mixture was stirred 10 min at
ꢁ78 °C and 10 min at 0 °C and quenched with saturated aqueous
NaHCO3 (40 mL). The aqueous phase was extracted with CH2Cl2
(3 ꢃ 25 mL), the combined organic phases were dried (MgSO4), fil-
tered and the solvent removed under vacuum. Purification by med-
ium pressure column chromatography (5:95 to 10:90 AcOEt/
hexane) afforded the product 10 as colourless viscous foam
AcOEt/EtOH (20 mL/10 mL) was hydrogenated at 35 psi in the
presence of a catalytic amount of Pd/C 10% (0.05 equiv). After
3 h, the reaction mixture was filtrated, the solvent was evaporated
and the residue dried under vacuum to afford methyl (2R)-2-(a-D-
glucopyranosyl)glycerate as very colourless viscous foam (0.95 g,
99%). No further purification was needed and we proceeded to
the next step. To a solution of methyl (2R)-2-(a-D-glucopyranos-
yl)glycerate (0.95 g, 3.4 mmol) in H2O (10 mL) was added a 2 M
KOH aqueous solution (1.65 mL). The reaction was stirred over-
night at rt. The pH was neutralized with HCl 10% and the solvent
evaporated to afford 12 as a very colourless viscous foam in the
form of its potassium salt. Its NMR data were identical to those
of the natural sample. ½a D20
ꢄ
+97.6 (c 2.08, H2O). 1H NMR (D2O): d
(4.44 g, 76%).
½
a 2D0
ꢄ
+45.2 (c 1.45, CH2Cl2). FT-IR (film)
5.00 (1H, d, J = 3.7 Hz), 4.39 (1H, t, J = 3.5 Hz), 3.89–3.87 (2H, m),
3.78–3.62 (4H, m), 3.49 (1H, dd, J = 9.8 Hz, J = 3.9 Hz), 3.34 (1H, t,
J = 9.1 Hz). 13C NMR (D2O): d 176.2 (C@O), 100.1 (C-1), 78.1, 75.3,
75.0, 73.9, 72.0, 65.1, 63.0. FT-IR
1744.3 cmꢁ1 (C@O). 1H NMR (CDCl3): d 7.38–7.37 (5H, m), 7.37–
7.22 (25H, m), 5.16 (1H, d, J = 3.3 Hz), 5.02 (1H, d, J = 10.6 Hz),
4.89 (2H, dd, J = 11.9 Hz, J = 3.1 Hz), 4.76 (1H, d, J = 10.5 Hz), 4.70
(1H, d, J = 11.8 Hz), 4.54 (1H, d, J = 11.2 Hz), 4.49 (1H, dd,
J = 6.3 Hz; J = 4.3 Hz), 4.15–3.94 (6H, m), 3.73 (3H, s), 3.60 (1H,
dd, J = 9.7 Hz, J = 3.5 Hz), 3.48 (1H, t, J = 9.5 Hz), 1.99 (3H, s), 1.03
(9H, s). 13C NMR (CDCl3): d 170.6, 170.2, 138.7, 138.1, 138.0,
135.6, 135.5, 133.0, 132.8, 129.8, 129.8, 128.4, 128.3, 128.2,
128.2, 128.1, 127.8, 127.7, 127.6, 94.8, 81.6, 75.8, 74.8, 72.0, 71.9,
69.0, 65.8, 64.6, 62.9, 51.9, 26.7, 20.8, 19.1. HR-MS: Calcd for
C49H56O10Si [M]+: 559.2130; found: 559.2125.
4.3. Differential scanning calorimetry
Differential scanning calorimetry (DSC) was performed on a
MicroCal VP-DSC MicroCalorimeter controlled by the VP-viewer
program and equipped with 0.51 mL cells. Calibration of tempera-
ture and heat-flow was carried out according to MicroCal instruc-
tions. Protein stock solutions of nuclease or MDH were prepared in
phosphate buffer (10 mM of sodium phosphate, pH 7.5), and lyso-
zyme was prepared in citrate buffer (20 mM sodium citrate,
55 mM NaCl, pH 6.0). These stock solutions were extensively dia-
lyzed against the same buffer before the assays. A volume of
2 mL of buffer with or without solutes was prepared and dispensed
4.2.8. Methyl 3-tert-butyldiphenylsilyl-(2R)-2-(2,3,4-tri-O-
benzyl-a-D-glucopyranosyl)glycerate 11
A solution (7.6 mL) of Na (0.046 g) in MeOH (10 mL) was added
to a stirred solution of the acetate 10 (2.0 g, 2.4 mmol) in MeOH

T. Q. Faria et al. / Carbohydrate Research 343 (2008) 3025–3033
3033
into two Eppendorf tubes, 1 mL each. An aliquot of the concen-
trated protein solution, or of buffer, was added to each Eppendorf,
and these solutions were used to fill up the sample and reference
cells, respectively. Both solutions were degassed for 8 min under
vacuum, prior to the calorimetric experiments. DSC scans were
run at a constant heating rate of 1 °C/min with an overpressure
of about 30 psi. The initial temperature of the DSC scans was
20 °C but the final temperature depended on the protein/solute
pair under study. For example, in the case of SNase in the presence
of KCl or glycerol the sample was heated up to 80 °C, while in the
presence of MG, DGP or DIP scans were run up to 90 °C. Protein
MDH looses 50% of the initial activity. Results are averages of at
least two independent experiments.
Acknowledgements
This work was supported by the European Commission, Con-
tract COOP-CT-2003-508644 and by Fundação para a Ciência e a
Tecnologia (FCT), Portugal, Project POCI/BIA-PRO/57263/2004.
TQF acknowledges a grant from FCT, SFRH/BPD/20352/2004. We
thank Professor A. L. Maçanita, Instituto Superior Técnico, Univer-
sidade Técnica de Lisboa for providing access to the fluorimeter.
concentrations approximately 5
lM were typically used for MDH
and approximately 20 M for SNase and lysozyme. Unless other-
l
Supplementary data
wise stated, solute concentration was 0.5 M. The dependency of
the melting temperature of SNase on the concentration of MG,
ML, phosphate, sulfate, acetate, glycerate (up to 1 M), KCl and glyc-
erol (up to 3 M) was studied. Sodium sulfate was used due to the
low solubility of the potassium salt. Potassium was the counterion
in all other salts examined.
A single endothermic peak was detected in the thermograms of
the three proteins studied, indicating that thermal unfolding of
these proteins occurred in a single step. Very good reproducibility
of the baselines, determined under identical conditions to those of
the assay, was always observed. Raw calorimetric data were con-
verted to heat capacity by subtracting the buffer baseline and by
dividing the scan rate and the protein concentration in the sample.
The protein melting temperature (the temperature at which equal
amounts of native and denatured forms are present) was deter-
mined using the software supplied with the instrument.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.carres.2008.08.030.
References
1. Selkoe, D. J. Neurochem. Res. 2003, 28, 1705–1713.
2. Dobson, C. M. Nature 2003, 426, 884–890.
3. Bernier, V.; Lagacé, M.; Bichet, D. G.; Bouvier, M. Trends Endocrinol. Metab. 2004,
15, 222–228.
4. van den Burg, B. Curr. Opin. Microbiol. 2003, 6, 213–218.
5. Rothschild, L. J.; Mancinelli, R. L. Nature 2001, 409, 1092–1101.
6. Vieille, C.; Zeikus, G. J. Microbiol. Mol. Biol. Rev. 2001, 65, 1–43.
7. Brown, A. D. Bacteriol. Rev. 1976, 40, 803–846.
8. da Costa, M. S.; Santos, H.; Galinski, E. A. Adv. Biochem. Eng. Biotechnol. 1998, 61,
117–153.
9. Oren, A. Microbiol. Mol. Biol. Rev. 1999, 63, 334–348.
10. Santos, H.; Lamosa, P.; Faria, T. Q.; Borges, N.; Neves, C. The Physiological Role,
Biosynthesis and Mode of Action of Compatible Solutes from
(Hyper)thermophiles. In Physiology and Biochemistry of Extremophiles; Gerday,
C., Glansdorff, N., Eds.; ASM Press: Washington, 2007; pp 86–103.
11. Santos, H.; da Costa, M. S. Environ. Microbiol. 2002, 4, 501–509.
12. Martins, L. O.; Santos, H. Appl. Environ. Microbiol. 1995, 61, 3299–3303.
13. Hensel, R.; König, H. FEMS Microbiol. Lett. 1988, 49, 75–79.
14. Scholz, S.; Sonnenbichler, J.; Schäfer, W.; Hensel, R. FEBS Lett. 1992, 306, 239–
242.
15. Ramos, A.; Raven, N. D. H.; Sharp, R. J.; Bartolucci, S.; Rossi, M.; Cannio, R.;
Lebbink, J.; van der Oost, J.; de Vos, W. M.; Santos, H. Appl. Environ. Microbiol.
1997, 63, 4020–4025.
16. Shima, S.; Hérault, D. A.; Berkessel, A.; Thauer, R. K. Arch. Microbiol. 1998, 170,
469–472.
4.4. Light scattering experiments
Protein aggregation was monitored by measuring the light scat-
tering in a Spectrofluorometer Spex-Fluorolog 1680, with excita-
tion and emission at 320 nm. Assays were performed with MDH
(0.2 mg/mL) and LDH (0.2 mg/mL) in 50 mM phosphate buffer,
pH 7.6, in the absence and presence of solutes. All solutes were
used at a concentration of 0.5 M; in the case of LDH, concentrations
of 0.1 M and 0.25 M MG were also examined. The protein solutions
17. Borges, N.; Ramos, A.; Raven, N. D. H.; Sharp, R. J.; Santos, H. Extremophiles
2002, 6, 209–216.
18. Faria, T. Q.; Knapp, S.; Ladenstein, R.; Maçanita, A. L.; Santos, H. ChemBioChem
2003, 4, 734–741.
19. Lamosa, P.; Burke, A.; Peist, R.; Huber, R.; Liu, M. Y.; Silva, G.; Rodrigues-
Pousada, C.; LeGall, J.; Maycock, C.; Santos, H. Appl. Environ. Microbiol. 2000, 66,
1974–1979.
20. Pais, T. M.; Lamosa, P.; dos Santos, W.; LeGall, J.; Turner, D. L.; Santos, H. FEBS J.
2005, 272, 999–1011.
(with or without solutes) were filtered (0.2
lm pore size) and
placed in a quartz cuvette thermostated at 25 °C. Temperature
was increased rapidly (heating rate approximately 10 °C/min) to
40 °C for MDH and 50 °C for LDH, and light scattering was mea-
sured as a function of time. Results are expressed as percentages
of the value measured in the absence of solutes after incubation
at the final temperature for 50 or 30 min for MDH or LDH,
respectively.
21. Faria, T. Q.; Lima, J. C.; Bastos, M.; Maçanita, A. L.; Santos, H. J. Biol. Chem. 2004,
279, 48680–48691.
22. Ray, A. K.; Roy, N. Carbohydr. Res. 1990, 196, 95–100.
23. Crich, D.; Smith, M. J. Am. Chem. Soc. 2001, 123, 9015–9020.
24. Mozhaev, V. V. Trends Biotechnol. 1993, 11, 88–95.
25. Cacace, M. G.; Landau, E. M.; Ramsden, J. J. Q. Rev. Biophys. 1997, 30, 241–277.
26. Ramos, C. H. I.; Baldwin, R. L. Protein Sci. 2002, 11, 1771–1778.
27. Kunz, W.; Lo Nostro, P.; Ninham, B. W. Curr. Opin. Colloid Interface Sci. 2004, 9,
1–18.
4.5. Activity assays
To evaluate the long-term thermostability of MDH, an enzyme
solution (1 U/mL in 20 mM sodium phosphate buffer, pH 7.6)
was incubated at 45 °C with or without solutes, and the residual
activity was measured at different time intervals as described be-
low. The solutes were used at a final concentration of 0.5 M.
MDH activity was assessed by following the conversion of NADH
to NAD+ at 340 nm with a Beckman spectrophotometer DU 70. A
cuvette containing 1 mL of reaction mixture (100 mM sodium
phosphate buffer, pH 7.6, 0.12 mM NADH and 0.25 mM oxaloace-
tate) was thermostated at 25 °C for 2 min and the reaction started
by the addition of 0.025 U of MDH. The results are presented as
half-life times, that is, as the incubation period during which
28. Parsegian, V. A. Nature 1995, 378, 335–336.
29. von Hippel, P. H.; Wong, K. Y. J. Biol. Chem. 1965, 240, 3909–3923.
30. Tadeo, X.; Pons, M.; Millet, O. Biochemistry 2007, 46, 917–923.
31. Timasheff, S. N. Annu. Rev. Biophys. Biomol. Struct. 1993, 22, 67–97.
32. Timasheff, S. N. Biochemistry 2002, 41, 13473–13482.
33. Liu, Y.; Bolen, D. W. Biochemistry 1995, 34, 12884–12891.
34. Thorne, C. J. Biochim. Biophys. Acta 1962, 59, 624–633.
35. Jaenicke, R.; Knof, S. Eur. J. Biochem. 1968, 4, 157–163.
36. Aune, K. C.; Tanford, C. Biochemistry 1969, 8, 4579–4585.
37. Fuchs, S.; Cuatrecasas, P.; Anfinsen, C. B. J. Biol. Chem. 1967, 242, 4768–
4770.
38. Schmidt, R. R.; Jung, K.-H. In Preparative Carbohydrate Chemistry; Hanessian, S.,
Ed.; Dekkar: New York, 1997; pp 283–312.