Characterization of a cloned subtilisin-like serine proteinase
from a psychrotrophic
Vibrio
species
Jo
´
hanna Arno
´
rsdo
´
ttir
1,2
,Ru
´
na B. Sma
´
rado
´
ttir
1
,O
´
lafur Th. Magnu
´
sson
2
, Sigrı
´
dur H. Thorbjarnardo
´
ttir

Sequence analysis revealed that especially with respect to
the thermophilic homologues, aqualysin I from Thermus
aquaticus and a proteinase from Thermus strain Rt41A,
the cold-adapted Vibrio-proteinase has a higher content of
polar/uncharged amino acids, as well as aspartate resi-
dues. The thermophilic enzymes had a higher content of
arginines, and relatively higher number of hydrophobic
amino acids and a higher aliphatic index. These factors
may contribute to the adaptation of these proteinases to
different temperature conditions.
Keywords: cold adaptation; psychrotrophic; Vibrio-protein-
ase; proteinase K-like; subtilisin-like proteinase.
Many microorganisms and ectothermic animals live under
environmental temperatures that fluctuate in the range )2
to 10 °C without the opportunity to regulate their cellular
temperatures [1–3]. In fact, cold temperature is the most
widespread physiological stress condition that organisms
have either to adapt to or to avoid. Adaptive changes in
protein structure and function induced by cold are of prime
importance for cold acclimation and survival processes [4].
A common denominator of evolutionary adaptive changes
of proteins appears to be the conservation and optimization
of the functional state of the proteins, such that they are in
Ôcorresponding statesÕ with respect to functionally important
motions, under the different physical conditions to which
the proteins have adapted [5]. It has been suggested that in
order to maintain such Ôcorresponding statesÕ for efficient
biological function at low temperatures, cold-adapted
proteins must have adopted a higher degree of conforma-
tional flexibility [5–11]. As such cold-adaptive strategies

Correspondence to M. M. Kristja
´
nsson, Department of Biochemistry,
Science Institute, University of Iceland, Dunhaga 3, 107 Reykjavik,
Iceland. Fax: + 354 5528911, Tel.: + 354 5254800,
E-mail:
Abbreviations: AQUI, aqualysin I; GdmSCN, guanidinium
thiocyanate; GdmCl, guanidinium chloride; PRK, proteinase K;
Suc-AAPF-NH-Np, succinyl-AlaAlaProPhe-p-nitroanilide;
VPR, Vibrio-proteinase.
Enzymes: aqualysin I (EC 3.4.21 ); proteinase K (EC 3.4.21.64);
Vibrio-proteinase (EC 3.4.21 )
Note: the sequence reported in this paper has been deposited in the
GenBank database (accession number AF521587).
(Received 19 June 2002, revised 11 September 2002,
accepted 16 September 2002)
Eur. J. Biochem. 269, 5536–5546 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03259.x
critical intramolecular interactions that tend to facilitate
increased local or global flexibilities in the protein molecules.
Fewer intra or intersubunit salt-bridges [17,18,21,23,
28,31,33], reduction in aromatic–aromatic interactions
[17,19,31,33], extended surface loops [17,21–23], fewer
prolines in such loops [19,21–23,28,33], lower hydropho-
bicity [18,19,26,34], weaker calcium-binding [17,27,28,34],
improved solvent interactions through additional surface
charges [17,19,22,23], and increased exposure of nonpolar
groups to the solvent [23,28] have all been cited as possible
reasons for increased flexibility and/or decreased thermal
stability of proteins from psychrophiles. In the case of
triose-phosphate isomerase from the psychrophilic bacter-

Bacterial strains and plasmids
The strain used for genomic DNA was Vibrio strain PA44
[25,38]. The E.colistrain used for cloning was TG1 supE,
hsdD5, thi, D(lac-proAB), F¢ (traD36, proAB
+
,lacI
q
,lac
DZM15) [39]. Cloning vectors used were pUC18 and 19,
M13 vectors mp18 and 19 (New England Biolabs). The
vectors pBAD (Invitrogen) and pJOE 3075.3 [40] were used
for gene expression in the E.colistrains Top10 and JM109,
respectively.
Growth conditions and DNA manipulations
Vibrio PA44 was grown as described [25,38] and genomic
DNA was prepared using the CTAB/NaCl method as
described [41]. For cloning of the proteinase gene primers
were designed from the sequence of Vibrio alginolyticus [42]:
5¢-GCG
GAATTCTACACCCGCTACATGTGGCGTCG
CCAT-3¢ and 5¢-CGC
GGATCCTGGGGACTAGATC
GAATC-GACCAACGTAA-3¢. Underlined are restriction
enzyme sites for EcoRI and BamHI, respectively. The
primers were used to amplify about 600 base pairs from the
genomic Vibrio PA44 DNA. The PCR product was cloned
into M13mp19 and sequenced. The sequencing revealed an
EcoRI restriction site on the PCR fragment.
Genomic Vibrio PA44 DNA was digested with several
restriction enzymes (EcoRI, BamHI, HindIII, SalI, SacI,

¼ 0.7. At this stage CaCl
2
was added to
the culture to a final concentration of 10 m
M
, expression
was induced with 0.01%
L
-arabinose and cells were then
grown at 22 °CtoanD
600
¼ 1.2 for up to 48 h. Cells were
harvested by centrifugation (8000 g for 15 min) and lysed in
50 m
M
Tris,pH 8,10 m
M
CaCl
2
with 1 mgÆmL
)1
lysozyme,
sonication and repeated freezing and thawing in liquid
nitrogen. The cell lysate was treated with 5 lgÆmL
)1
DNase
and 5 lgÆmL
)1
RNase at 4 °C overnight and cell debris was
removed by centrifugation (20 000 g for 20 min). Proteins

)1
Æcm
)1
and 34 295
M
)1
Æcm
)1
for the 40.6
Ó FEBS 2002 Characterization of a cloned psychrotrophic proteinase (Eur. J. Biochem. 269) 5537
and 29.7 kDa forms of the protein, respectively, determined
as described by Pace et al. [43]. Mass spectral data was
obtained on a Reflex III MALDI-TOF spectrometer
(Bruker), operated in a linear mode. The matrix used was
sinnepic acid (Bruker) in a 50 : 50 acetonitrile/water mix-
ture, containing 0.1% trifluoroacetic acid, using the dried
droplet method.
Kinetic and thermal stability measurements
Enzymatic activity of VPR was assayed using SucAAPF-
NH-Np as a substrate as described previously [25], using
a thermoregulated Unicam UV1 spectrometer. Kinetic
parameters for activity were determined at 25 °Cbyfitting
the rate data measured at substrate concentrations between
0.05 and 1 m
M
, to the Michaelis–Menten equation by
nonlinear regression using the software
KCAT
(Biometallics,
Inc., Princeton, NJ, USA).

as the temperature at which the rate of inactivation
corresponded to 50% loss of original enzyme activity after
30 min.
GdmSCN denaturation experiments
For measuring denaturation curves of VPR, PRK and
AQUI as a function of GdmSCN concentration, the
proteinases were first inhibited by incubation with 10 m
M
phenylmethylsulfonyl fluoride for at least 15 min, before
the samples were applied to a Sephadex G-25 column,
equilibrated with 25 m
M
Tris, pH 8.0, containing 15 m
M
CaCl
2
and 1 m
M
EDTA. The collected protein peaks
were diluted with the buffer containing the different
concentrations of GdmSCN. After incubation for up
to 24 h at 25 °C, the degree of unfolding was moni-
tored by measuring the fluorescence intensity using a
Spex Fluoromax spectrofluorometer. Reversal of inhibi-
tion of the proteinases during incubation was minor and
would not affect the pretransition baselines in the
denaturation curves. The excitation wavelengths were
275 nm for VPR and AQUI and 280 nm for PRK, but
emissions were monitored at 355 nm for VPR, at 335 nm
for PRK and at 320 nm in the case of AQUI. Excitation

place during unfolding. The relationship between unfolding
and autolysis is usually ill-defined and in most cases it is not
clear to what extent global or local unfolding of the protein
molecule has to take place to trigger autolysis. In order to
obtain an estimate of the conformational stability of the
three related proteinases of this comparative study, we also
determined the denaturation curves of the enzymes inhibited
by phenylmethanesulfonyl fluoride, as a function of denat-
urant concentration. Several subtilases have been reported
to exhibit high stability towards protein denaturants, such
as urea and GdmCl [44]. In this study, the powerful
denaturant GdmSCN was used in most experiments as
neither urea nor GdmCl unfolded aqualysin I at concen-
trations set by the upper limit of the aqueous solubility of
the denaturants. Normalized denaturation curves for VPR,
PRK and AQUI as a function of GdmSCN concentration
at 25 °C and pH 8.0 are shown in Fig. 1. In line with the
previous results on thermal stability, a significant difference
was observed in the conformational stability of the
proteinases, following the order of their respective temper-
atures of adaptation. The [GdmSCN]
1/2
-values obtained
from the curves were 0.55
M
,1.5
M
and 3.2
M
for the VPR,

1/2
for PRK was
5.3
M
(data not shown).
Sequence analysis
The sequencing of the vpr gene revealed a 1593-bp sequence
encoding a protein of 530 amino acid residues with a
calculated molecular mass of 55.7 kDa (Fig. 2). The
deduced sequence has high sequence identity to enzymes
of the proteinase K family of subtilisin-like serine protein-
ases (Fig. 3), confirming its former classification as a
member of that family [25]. As for other enzymes belonging
to the proteinase K family [42,45], the VPR sequence
consists of three parts; an N-terminal prosequence, a
proteinase or catalytic domain, and a C-terminal prose-
quence. The N-terminal prosequence of VPR consists of 139
residues and most probably functions as a molecular
chaperone for correct folding, but that is subsequently
cleaved off by autolysis to give the active proteinase [46–50].
Thus, VPR is isolated from cultures of Vibrio strain PA44 as
a 40.6-kDa protein, without the 139 residue N-terminal
sequence. The enzyme undergoes further autolysis under
relatively mild conditions where a 100-residue C-terminal
extension is cleaved off to give a 29.7-kDa proteinase that
remains fully active [25]. Both the 40.6-kDa and the
29.7-kDa forms of the enzyme have the same N-terminal
sequence, based on amino acid sequencing and MALDI-
TOF mass spectrometry of the two enzyme forms indicates
that the peptide bond cleaved between the catalytic domain

the polypeptide chain within these protein structures. When
compared with the thermophilic counterparts, there appar-
ently is a trend to higher content of Asn and Gln in the cold-
adapted enzyme and also is the number of acidic amino acid
residues, especially Asp, higher in the three Vibrio protein-
ases (Table 1). The thermophilic enzymes, however have
higher content of Arg, and relatively higher numbers of
hydrophobic amino acids, as well as a higher aliphatic index
than the psychro- or mesophilic enzymes (Table 2). When
each of the amino acid exchanges that occur between VPR
and the other related enzymes were studied, a trend to an
increased number of Ser in the cold-adapted enzyme was
apparent. Especially was Ala fi Ser a frequent thermophi-
lic-to-psychrophilic exchange. Eight such Ala to Ser
exchanges occur between the structures of AQUI and
VPR and seven between the Thermus Rt41 proteinase
and VPR. Six of the exchanges that occur between VPR and
AQUI are also present, however, in the mesophilic enzyme
Fig. 2. Nucleotide sequence and deduced amino acid sequence of the vpr
gene coding for the subtilisin-like proteinase precursor from Vibrio strain
PA44. The N- and C-terminal residues of the proteinase domain are
underlined and residues of the catalytic triad are enlarged in bold
letters.
Ó FEBS 2002 Characterization of a cloned psychrotrophic proteinase (Eur. J. Biochem. 269) 5539
from V. alginolyticus, but all together there are four
Ala fi Ser exchanges between VPR and the mesophilic
enzyme. Comparison of the sequences of the two mesophilic
proteinases from V. alginolyticus and V. cholerae (69%
sequence identity) and between the thermophilic AQUI and
Thermus Rt41 proteinase (71% sequence identity) indicated

ase (VPR), proteinase from Vibrio alginolyti-
cus, aqualysin I (AQUI) from Thermus
aquaticus YT-1 and proteinase K (PRK) from
Tritirachium album. Arrows indicate the
N-terminal of VPR as determined by amino
acid sequence analysis [25] and the beginning of
the C-terminal extension based on results from
MALDI-TOF mass spectrometry. Residues of
the catalytic triad are indicated by an asterisk.
Secondary structural elements based on known
crystal structures of PRK are indicated by h
(helix) and s (strand) and calcium binding
ligands (P175, V177 and D200 at the
Ca1 site and T16 and D260 at the Ca2 site,
according to the numbering of the PRK
sequence) are denoted by C.
5540 J. Arno
´
rsdo
´
ttir et al. (Eur. J. Biochem. 269) Ó FEBS 2002
[17]. We have also observed a strong dependence on calcium
binding for the stability of VPR, as well as for PRK and
AQUI (M. M. Kristjansson, unpublished results). Two
calcium sites with different binding affinities have been
identified in the high resolution crystal structures of PRK
[57,58]. At the stronger Ca1-site, calcium ion is coordinated
in the form of pentagonal bipyramid by O
d1
and O

Ala 25 8.6 33 11.5 28 9.8 36 12.9 40 14.2 33 11.8
Arg 9 3.1 8 2.8 11 3.8 10 3.6 15 5.3 12 4.3
Asn 23 7.9 20 7.0 26 9.1 14 5.0 19 6.8 17 6.1
Asp 21 7.2 20 7.0 21 7.3 13 4.7 13 4.6 13 4.7
Cys 6 2.1 6 2.1 4 1.4 4 1.4 4 1.4 5 1.8
Gln 10 3.4 9 3.1 13 4.5 8 2.9 5 1.8 7 2.5
Glu 5 1.7 3 1.0 5 1.7 1 0.4 3 1.1 5 1.8
Gly 40 13.7 7 12.9 35 12.2 31 11.2 37 13.2 33 11.8
His 4 1.4 4 1.4 5 1.7 7 2.5 5 1.8 4 1.4
Ile 9 3.1 8 2.8 13 4.5 11 4.0 9 3.2 11 3.9
Leu 18 6.2 15 5.2 15 5.2 19 6.8 19 6.8 14 5.0
Lys 5 1.7 5 1.7 9 3.1 4 1.4 2 0.7 8 2.9
Met 3 1.0 3 1.0 3 1.0 3 1.1 2 0.7 5 1.8
Phe 7 2.4 6 2.1 5 1.7 5 1.8 3 1.1 6 2.2
Pro 12 4.1 7 2.4 13 4.5 13 4.7 11 3.9 9 3.2
Ser 38 13.1 38 13.3 26 9.1 22 7.9 29 10.3 37 13.3
Thr 20 6.9 19 6.6 22 7.7 34 12.2 25 8.9 22 7.9
Trp 4 1.4 4 1.4 3 1.0 4 1.4 3 1.1 2 0.7
Tyr 8 2.8 9 3.1 8 2.8 15 5.4 12 4.3 17 6.1
Val 24 8.2 32 11.2 22 7.7 24 8.63 25 8.9 19 6.8
S 291 286 287 278 281 288
Table 2. Comparison of structural parameters based on amino acid composition of Vibrio-proteinase and related enzymes from the proteinase K family.
Parameter Vibrio-proteinase V. alginol V. cholerae Thermus Rt41a AQUAI PRK
% charged
a
15.12 13.99 17.77 12.60 13.53 15.05
% acidic
b
8.93 8.04 9.06 5.04 5.70 6.45
% basic

LMIVWPAF;
f
Calculated at />g
GRAVY, grand
average of hydropathicity.
Ó FEBS 2002 Characterization of a cloned psychrotrophic proteinase (Eur. J. Biochem. 269) 5541
possibly because of detrimental background expression.
The gene was cloned into another vector, pJOE 3075.3
[40]. The gene was overexpressed using the pJOE vector in
E.colistrain JM109, but the protein formed inclusion
bodies. Changing the conditions in the expression culture,
such as lowering the temperature or changing the
concentration of the inducer has not given production of
active enzyme, nor have attempts to refold the protein by
different means in vitro been successful. The gene was
moderately expressed, giving a yield of  2mgÆL
)1
,using
the pBAD TOPO expression system. No activity was
detected after induction and culturing at 37 °C. Culturing
at 32 °C after induction resulted in a detectable produc-
tion of active proteinase, which was further enhanced by
lowering the growth temperature to room temperature
after induction.
Characterization of the recombinant VPR
The purified VPR
rt
showed identical properties to VPR
wt
with respect to migration on SDS/PAGE and the time

plots (Table 3). They also had comparable Michaelis–
Menten kinetic parameters when their amidase activity
against succinyl-AAPF-NH-Np was measured at 25 °C
(Table 3).
DISCUSSION
The extracellular proteinase K-like proteinase from the
psychrotolerant Vibrio strain PA44 is produced as a
55.7-kDa precursor protein, containing an 139 residue
N-terminal prosequence, a 291 residue proteinase domain,
and a 100 residue C-terminal domain. Production of such
relatively large precursor proteins appears to be a common
trait of proteinases belonging to the proteinase K family. In
aqualysin I, from the thermophile Thermus aquaticus, a 127-
residue N-terminal prosequence is cleaved off in an intra-
molecular autocatalytic reaction apparently after assisting in
the correct folding of the proteinase [49,50,59]. The
N-terminal prosequence is necessary for the production of
active aqualysin I and most likely acts as an intramolecular
chaperone, by similar mechanisms as has been described for
such N-prosequences of subtilisins [46–48] and a-lytic
protease, a trypsin-like serine proteinase [60]. Because of
the similarity that exists between the sequences of VPR and
AQUI, we predict that the N-terminal prosequence of the
cold-adapted enzyme has such an intramolecular chaperone-
like activity for that enzyme. A 105 residue C-terminal
prosequence has been found to play an important role in
extracellular secretion of AQUI in an expression system
using Thermus thermophilus cells, and it was suggested that
the sequence is required for the translocation of the
precursor across the outer membrane [61]. VPR is secreted

and k
cat
) against succinyl-AAPF-p-nitroanilide of wild-
type (VPR
wt
) and recombinant (VPR
rt
)formsoftheVibrio-proteinase.
Property VPR
wt
VPR
rt
T
50%
56 °C56°C
K
m
179 ± 12 l
M
164 ± 17 l
M
k
cat
68.9 s
)1
76.5 s
)1
5542 J. Arno
´
rsdo

their homologs from several mesophilic Methanococcus
species, the strongest correlation with thermophily of the
observed amino acid exchanges was the decrease in the
content of uncharged polar residues (Ser, Thr, Asn and Gln)
[64]. Of the specific amino acid replacements, the Ser fi Ala
exchange showed the highest correlation with thermophily
[64]. The Ser fi Ala replacement was also observed as one
of the most frequent ÔthermostabilizingÕ amino acid ex-
change observed by Argos and coworkers [65,66] in their
sequence comparisons of meso- and thermophilic proteins.
We are now carrying out mutagenic studies on VPR to
examine whether some of the Ala fi Ser substitution we
have observed in this comparison may contribute to its cold
adaptation.
The cold-adapted VPR also seems to be less hydro-
phobic than the thermophilic Thermus proteinases, as
calculated in Table 2. The lower hydrophobic content
and aliphatic index of VPR can largely be accounted for
by the lower Ala content. Four Pro residues located on
surface loops in AQUI, that are not present in VPR, may
contribute to the higher stability of the thermophilic as
compared with the cold-adapted enzyme. As a result of
inherent constraints on rotations around the N–Ca bond
of proline residues, their introduction into mobile regions
of proteins, such as loops, is expected to restrict available
backbone configurations and thus increase rigidity and
hence stability in such regions [67,68]. Stabilization of
loops may also take place by shortening or deletion, or
by substitutions that reduce the number of ÔflexibleÕ
glycines in such regions [67–69]. In addition to restricting

least one of the two well-defined calcium binding sites (the
stronger Ca1-site) in the structure of PRK is also present
in these enzymes, but the weaker, Ca2-site, cannot be
present due to the lack of the carboxylate ligand
corresponding to Asp260 in PRK. For VPR it is
interesting to observe that a highly similar amino acid
sequence as the one that makes up one of the three
calcium binding sites (the Ômedium strengthÕ Ca2-site)
identified in thermitase is present in the enzyme (Fig. 5).
This Ca-site is positioned in a loop that leads into an
a-helix containing the active site His residue of the
catalytic triad. Except for one substitution (Ala66 fi Ser)
the sequences of VPR and the V. alginolyticus proteinase
are identical in this region, but are different from AQUI.
In addition to the ligands indicated in Fig. 5, an Arg residue
(Arg102) stabilizes the Ca-site by making a salt bridge to the
ligands Asp57 and Asp60 [54,55]. Both of the Vibrio sp.
enzymes, as well as AQUI and the Thermus Rt41 proteinase,
also contain a conserved Arg residue at equivalent sites in
their sequences (Fig. 3). It still remains to be seen if a
difference in calcium binding affinity in any way affects the
temperature adaptation among these proteinases.
In common with many cold-adapted enzymes character-
ized so far, VPR is anionic, e.g. it has an acidic isoelectric
point. It apparently shares this property with its mesophilic
counterparts (Table 3), but is different from the basic pIs for
AQUI (pI >9–10) [70] and the Thermus Rt41A proteinase
(pI  10.25–10.5) [63]. The lower pI of VPR compared with
its thermophilic counterparts results from higher content of
Fig. 5. Comparison of the amino acid sequence of the main ligands

1. Hazel, J.R. & Prosser, C.L. (1974) Molecular mechanisms of
temperature compensation in poikilotherms. Physiol. Rev. 54,
620–677.
2. Somero, G.N. (1995) Proteins and temperature. Annu. Rev.
Physiol. 57, 43–68.
3. Feller, G., Arpigny, J.L., Narinx, E. & Gerday, C. (1997) Mole-
cular adaptations of enzymes from psychrophilic organisms.
Comp.Biochem.Physiol.118A, 495–499.
4. Franks, F. (1995) Protein destabilization at low temperatures.
Adv. Prot. Chem. 46, 105–139.
5. Jaenicke, R. & Bo
¨
hm, G. (1998) The stability of proteins in
extreme environments. Curr. Opin. Struct. Biol. 8, 738–748.
6. Feller, G., Narinx, E., Arpigny, J.L., Aittaleb, M., Baise, E.,
Genicot, S. & Gerday, C. (1996) Enzymes from psychrophilic
organisms. FEMS Microbiol. Lett. 18, 189–202.
7. Gerday, C., Aittaleb, M., Arpigny, J.L., Baise, E., Chessa, J.P.,
Garsoux, G., Petrescu, I. & Feller, G. (1997) Psychrophilic
enzymes: a thermodynamic challenge. Biochem. Biophys. Acta
1342, 119–131.
8. Kristja
´
nsson, M.M., A
´
sgeirsson, B. & Bjarnason, J.B. (1997)
Serine proteinases from cold-adapted organisms. Adv. Exp. Med.
Biol. 415, 27–46.
9.Za
´

Mainfroid, V., Goraj, K., Lion, M., Hol, W.G.J. & Martial, J.A.
(1993) Cloning and overexpression of the triose-phosphate iso-
merase genes from psychrophilic and thermophilic bacteria.
Structural comparison of the predicted protein sequences. J. Mol.
Biol. 229, 85–93.
17. Davail, S., Feller, G., Narinx, E. & Gerday, C. (1994) Cold
adaptation of proteins. Purification, characterization, and
sequence of the heat-labile subtilisin from the antarctic psychro-
phile Bacillus TA41. J. Biol. Chem. 269, 17448–17453.
18. Feller, G., Payan, F., Theys, F., Qian, M., Haser, R. & Gerday, C.
(1994) Stability and structural analysis of a-amylase from the
antarctic psychrophile Alteromonas haloplanctis. Eur. J. Biochem.
222, 441–447.
19. Feller, G., Zekhnini, Z., Lamotte-Brasseur, J. & Gerday, C. (1997)
Enzymes from cold-adapted microorganisms. The class C b-lac-
tamase from the antarctic psychrophile Psychrobacter immobilis
A5. Eur. J. Biochem. 244, 186–191.
20. Narinx, E., Baise, E. & Gerday, C. (1997) Subtilisin from psy-
chrophilic antarctic bacteria: characterization and site-directed
mutagenesis of residues possibly involved in the adaptation to
cold. Protein Eng. 10, 1271–1279.
21. Wallon, G., Lovett, S.T., Magyar, C., Swingor, A., Szilagyi, A.,
Zavodsky, P., Ringe, D. & Petsko, G.A. (1997) Sequence and
homology model of 3-isopropylmalate dehydrogenase from the
psychrotrophic bacterium Vibrio sp. I5 suggest reasons for thermal
instability. Protein Eng. 10, 665–672.
22. Gerike, U., Danson, M.J., Russel, N.J. & Hough, D.W. (1997)
Sequencing and expression of the gene encoding a cold-active
citrate synthase from an Antarctic bacterium, strain DS2–3R.
Eur.J.Biochem.248, 49–57.

28. Aghajari, N., Feller, G., Gerday, C. & Haser, R. (1998) Structures
of the psychrophilic Alteromonas haloplanctis a-amylase give
insights into cold adaptation at a molecular level. Structure 6,
1503–1516.
29. Feller, G., d’Amico, D. & Gerday, C. (1999) Thermodynamic
stability of a cold-active a-amylase from the Antarctic bacterium
Alteromonas haloplanktis. Biochemistry 38, 4613–4619.
30. D’Amico, S., Gerday, C. & Feller, G. (2001) Structural determi-
nants of cold adaptation and stability in a large protein. J. Biol.
Chem. 276, 25791–25796.
31. Kim, S Y., Hwang, K.Y., Kim, S H., Sung, H C., Han, Y.S. &
Cho, Y. (1999) Structural basis for cold adaptation. Sequence,
biochemical properties, and crystal structure of malate
5544 J. Arno
´
rsdo
´
ttir et al. (Eur. J. Biochem. 269) Ó FEBS 2002
dehydrogenase from a psychrophile Aquaspirillum arcticum.
J. Biol. Chem. 274, 11761–11767.
32. Georlette,D.,Jonsson,Z.O.,VanPetegem,F.,Chessa,J.,Van
Beeumen, J., Hubscher, U. & Gerday, C. (2000) A DNA ligase
from the psychrophile Pseudoalteromonas haloplanktis gives
insight into the adaptation of proteins to low temperatures. Eur. J.
Biochem. 267, 3502–3512.
33. Bentahir, M., Feller, G., Aittaleb, M., Lamotte-Brasseur, J.,
Himri, T., Chessa, J.P. & Gerday, C. (2000) Structural, kinetic,
and calorimetric characterization of the cold-active phosphogly-
cerate kinase from the antarctic Pseudomonas sp. TACII18. J. Biol.
Chem. 275, 11147–11153.

Smith, J.A. & Struhl, K. (1989) Current Protocols in Molecular
Biology. Green Publishing Associates and Wiley-Interscience,
New York.
42. Deane, S.M., Robb, F.T., Robb, S.M. & Woods, D.R. (1989)
Nucleotide sequence of the Vibrio alginolyticus calcium dependent,
detergent-resistant alkaline serine exoprotease A. Gene 76, 281–
288.
43. Pace,C.N.,Vajdos,F.,Fee,L.,Grimsley,G.&Gary,T.(1995)
How to measure and predict the molar absorption coefficient of a
protein. Protein Sci. 4, 2411–2423.
44. Genov, N., Filippi, B., Dolashka, P., Wilson, K.S. & Betzel, C.
(1995) Stability of subtilisins and related proteinases (subtilases).
Int. J. Peptide Protein Res. 45, 391–400.
45. Terada,I.,Kwon,S.T.,Miyata,Y.,Matsuzawa,H.&Ohta,T.
(1990) Unique precursor structure of an extracellular protease,
aqualysin I, NH
2
- and COOH-terminal pro-sequence and its
processing in Escherichia coli. J. Biol. Chem. 265, 6576–6581.
46. Shinde, U. & Inouye, M. (1993) Intramolecular chaperon and
protein folding. Trends Biochem. Sci. 18, 442–446.
47. Eder, J., Rheinecker, M. & Fersht, A.R. (1993) Folding of
subtilisin BPN¢: role of the pro-sequence. J. Mol. Biol. 233,
293–304.
48. Bryan, P., Wang, L., Hoskins, J., Ruvinov, S., Strausberg, S.,
Alexander, P., Almog, O., Gilliland, G. & Gallagher, T. (1995)
Catalysis of a protein folding reaction: Mechanistic implications
of the 2.0 A
˚
structure of the subtilisin-prodomain complex.

hne, W.E. & Wilson, K.S. (1990) Crystal
structure of thermitase at 1.4 A
˚
resolution. J. Mol. Biol. 214,
261–279.
55. Gros, P., Kalk, K.H. & Hol, W.G.J. (1991) Calcium binding of
thermitase. Crystallographic studies of thermitase at 0, 5, and 100
mM calcium. J. Biol. Chem. 266, 2953–2961.
56. Lin, S.J., Yoshimura, E., Sakai, H., Wakagi, T. & Matsuzawa, H.
(1999) Weakly bound calcium ions involved in the thermostability
of aqualysin I, a heat-stable subtilisin-type protease of Thermus
aquaticus YT-1. Biochim. Biophys. Acta 1433, 132–138.
57. Mu
¨
ller, A., Hinrich, W., Wolf, W.M. & Saenger, W. (1994)
Crystal structure of calcium-free proteinase K at 1.5 A
˚
resolution.
J. Biol. Chem. 269, 23108–23111.
58. Betzel, C., Gourinath, S., Kumar, P., Kaur, P., Perbandt, M.,
Eschenburg, S. & Singh, T.P. (2001) Structure of a serine protei-
nase proteinase K from Tritirachium album limber at 0.98 A
˚
resolution. Biochemistry 40, 3080–3088.
59. Kurosaka, K., Ohta, T. & Matsuzawa, H. (1996) 38 kDa pre-
cursor protein of aqualysin I (a thermophilic subtilisin-type pro-
tease) with a C-terminal extended sequence: its purification and
in vitro processing. Mol. Microbiol. 20, 385–389.
60. Cunningham, E.L., Jaswal, S.S., Sohl, J.L. & Agard, D.A. (1999)
Kinetic stability as a mechanism for protease longevity. Proc. Natl

protein thermostability from site-directed mutations that decrease
the entropy of unfolding. Proc. Natl Acad. Sci. USA 84, 6663–
6677.
68. Vielle, C., Burdette, D.S. & Zeikus, J.G. (1996) Thermozymes.
Biotechnol. Ann. Rev. 2, 1–83.
69. Thompson, M.J. & Eisenberg, D. (1999) Transproteomic evidence
of a loop-deletion mechanism for enhancing protein thermo-
stability. J. Mol. Biol. 290, 595–604.
70. Matsuzawa, H., Tokugawa, K., Hamaoki, M., Mizoguchi, M.,
Taguchi, H., Terada, I., Kwon, S.T. & Ohta, T. (1988) Purification
and characterization of aqualysin I (a thermophilic alkaline serine
protease) produced by Thermus aquaticus YT-1. Eur.J.Biochem.
171, 441–447.
5546 J. Arno
´
rsdo
´
ttir et al. (Eur. J. Biochem. 269) Ó FEBS 2002