Tài liệu Báo cáo khoa học: Destabilization of psychrotrophic RNase HI in a localized fashion as revealed by mutational and X-ray crystallographic analyses pdf

Destabilization of psychrotrophic RNase HI in a localized
fashion as revealed by mutational and X-ray
crystallographic analyses
Muhammad S. Rohman
1
, Takashi Tadokoro
1
, Clement Angkawidjaja
1
, Yumi Abe
1
,
Hiroyoshi Matsumura
2,3
, Yuichi Koga
1
, Kazufumi Takano
1,3
and Shigenori Kanaya
1
1 Department of Material and Life Science, Graduate School of Engineering, Osaka University, Japan
2 Department of Applied Chemistry, Graduate School of Engineering, Osaka University, Japan
3 CREST, JST, Osaka, Japan
Psychrophiles and psychrotrophs are defined as micro-
organisms that can grow even at around 0 °C [1].
Enzymes from these microorganisms are usually less
stable than those from mesophiles and thermophiles
[2–4]. It has been reported that a decreased number of
ion pairs and hydrogen bonds, decreased hydrophobic
interactions and packing at the core, an increased
fraction of nonpolar surface area, a decreased surface

hydrophilicity, decreased helix stability and a
decreased number of proline residues in the loop
regions are responsible for their thermolability [5–8].
However, the destabilization mechanism of these
enzymes remains to be fully understood. One promis-
ing strategy to understand this mechanism is to
Keywords
crystal structure; destabilization mechanism;
RNase HI; Shewanella oneidensis MR-1;
thermostabilizing mutations
Correspondence
S. Kanaya, Department of Material and Life
Science, Graduate School of Engineering,
Osaka University, 2-1, Yamadaoka, Suita,
Osaka 565-0871, Japan
Fax: +81 6 6879 7938
Tel: +81 6 6879 7938
E-mail:
(Received 26 September 2008, revised 11
November 2008, accepted 19 November
2008)
doi:10.1111/j.1742-4658.2008.06811.x
The Arg97 fi Gly and Asp136 fi His mutations stabilized So-RNase HI
from the psychrotrophic bacterium Shewanella oneidensis MR-1 by 5.4
and 9.7 °C, respectively, in T
m
, and 3.5 and 6.1 kJÆmol
)1
, respectively, in
DG(H

2
O). These mutations also stabilized the So-RNase HI derivative
(4·-RNase HI) with quadruple thermostabilizing mutations in an additive
manner. As a result, the resultant sextuple mutant protein (6·-RNase HI)
was more stable than the wild-type protein by 28.8 °CinT
m
and 27.0
kJÆmol
)1
in DG (H
2
O). To analyse the effects of the mutations on the pro-
tein structure, the crystal structure of the 6·-RNase HI protein was deter-
mined at 2.5 A
˚
resolution. The main chain fold and interactions of the
side-chains of the 6·-RNase HI protein were basically identical to those of
the wild-type protein, except for the mutation sites. These results indicate
that all six mutations independently affect the protein structure, and are
consistent with the fact that the thermostabilizing effects of the mutations
are roughly additive. The introduction of favourable interactions and the
elimination of unfavourable interactions by the mutations contribute to the
stabilization of the 6·-RNase HI protein. We propose that So-RNase HI is
destabilized when compared with its mesophilic and thermophilic coun-
terparts in a localized fashion by increasing the number of amino acid
residues unfavourable for protein stability.
Abbreviations
4·-RNase HI, So-RNase HI derivative with Asn29 fi Lys, Asp39 fi Gly, Met76 fi Val and Lys90 fi Asn mutations; 5·-RNase HI, 4·-RNase
HI derivative with additional Arg97 fi Gly mutation; 6·-RNase HI, 5·-RNase HI derivative with additional Asp136 fi His mutation; D136H-
RNase HI, So-RNase HI derivative with Asp136 fi His mutation; Ec-RNase HI, E. coli RNase HI; GdnHCl, guanidine hydrochloride; PDB,

Protein Data Bank; R97G-RNase HI, So-RNase HI derivative with Arg97 fi Gly mutation; So-RNase HI, RNase HI from
Shewanella oneidensis MR-1; Tt-RNase HI, RNase HI from Thermus thermophilus.
FEBS Journal 276 (2009) 603–613 ª 2008 The Authors Journal compilation ª 2008 FEBS 603
construct the thermostabilized mutants of a given psy-
chrophilic or psychrotrophic enzyme and analyse their
stabilization mechanisms.
RNase H (EC 3.1.26.4) is an enzyme that specifically
cleaves the RNA strand of RNA ⁄ DNA hybrids [9].
The enzyme is widely present in bacteria, archaea,
eukaryotes and retroviruses [10]. RNase HI from the
psychrotrophic bacterium Shewanella oneidensis MR-1
(So-RNase HI) is a monomeric protein with 158 amino
acid residues [11]. It shows amino acid sequence iden-
tity of 67% with its mesophilic counterpart Escherichia
coli RNase HI (Ec-RNase HI), for which structure–
stability–function relationships have been extensively
studied [12]. The crystal structure of So-RNase HI has
been determined [11]. This structure strongly resembles
that of Ec-RNase HI. Nevertheless, So-RNase HI is
less stable than Ec-RNase HI by 22.4 °CinT
m
and 12.5 kJÆmol
)1
in D G (H
2
O) [11]. We used So-RNase
HI as a model protein to analyse the destabilization
mechanism of a psychrotrophic protein.
We have recently shown that four single mutations
identified by directed evolution stabilize So-RNase HI

by 3.6–6.7 °CinT
m
and 1.7–5.2 kJÆmol
)1
in DG(H
2
O)
[13]. They include Asn29 fi Lys, Asp39 fi Gly,
Met76 fi Val and Lys90 fi Asn. The effects of these
mutations are roughly additive, and a combination of
these mutations strikingly increases the stability of
So-RNase HI to a level similar to that of Ec-RNase HI.
These results suggest that Asn29, Asp39, Met76 and
Lys90 are not optimal for the stability of So-RNase HI
and their replacement with other residues increases
stability. However, the stabilization mechanisms of the
protein with these mutations remain to be understood.
In addition, it remains to be determined whether the
four residues mentioned above are the only ones that
are not optimal for the stability of So-RNase HI.
It has been reported that Ec-RNase HI is stabilized
by the Lys95 fi Gly [14] or Asp134 fi His [15] muta-
tion by approximately 7 °CinT
m
at pH 5.5. Because
Lys95 and Asp134 are conserved as Arg97 and
Asp136, respectively, in So-RNase HI, and the struc-
tures around these residues are conserved in So-RNase
HI [9], the Arg97 fi Gly and Asp136 fi His mutations
are also expected to increase the stability of So-RNase

HI. However, these mutations have not been identified
by directed evolution. Therefore, it would be informa-
tive to examine whether these mutations increase the
stability of So-RNase HI and its derivative (4·-RNase
HI) with quadruple thermostabilizing mutations identi-
fied by directed evolution.
In this report, we show that the Arg97 fi Gly and
Asp136 fi His mutations increase the stability of
So-RNase HI and 4·-RNase HI. We determined the
crystal structure of the sextuple mutant protein of
So-RNase HI (6·-RNase HI), in which the Arg97 fi
Gly and Asp136 fi His mutations were combined with
the four thermostabilizing mutations identified by
directed evolution. Based on this structure, which is
basically identical to that of the wild-type protein,
except for the mutation sites, we discuss the destabili-
zation mechanism of So-RNase HI.
Results
Stabilization of So-RNase HI with Arg97
fi
Gly
and Asp136
fi
His mutations
To examine whether the single Arg97 fi Gly and
Asp136 fi His mutations stabilize So-RNase HI, two
mutant proteins, R97G-RNase HI and D136H-RNase
HI, were constructed. These mutant proteins were
overproduced in E. coli in a soluble form and purified
to give a single band on SDS-PAGE (data not shown).

The far-UV CD spectra of these mutant proteins were
similar to that of the wild-type protein (data not
shown), suggesting that these mutations do not
seriously affect the conformation of the protein. The
specific activities of R97G-RNase HI and D136H-
RNase HI were 99% and 65%, respectively, of that of
the wild-type protein (Table 1).
The stabilities of R97G-RNase HI and D136H-
RNase HI against thermal denaturation were analysed
at pH 5.5 in the presence of 1 m guanidine hydrochlo-
ride (GdnHCl) by monitoring the change in the CD
values at 220 nm. Thermal denaturation of these
mutant proteins was fully reversible in this condition.
The thermodynamic parameters characterizing the
thermal denaturation curves of the wild-type and
mutant proteins are summarized in Table 1. The tem-
perature of the midpoint of the transition, T
m
, was
30.4 °C for the wild-type protein, 35.8 °C for R97G-
RNase HI and 40.1 °C for D136H-RNase HI. Thus,
R97G-RNase HI is more stable than the wild-type
protein by 5.4 °CinT
m
and 3.9 kJÆmol
)1
in DDG
m
.
D136H-RNase HI is more stable than the wild-type

protein by 9.7 °CinT
m
and 7.0 kJÆmol
)1
in DDG
m
.
The stabilities of the mutant proteins against
urea-induced denaturation were also analysed by
monitoring the change in the CD values at 220 nm.
Urea-induced denaturation of these proteins was fully
reversible in this condition and showed a two-state
transition. The thermodynamic parameters characteriz-
ing the urea-induced denaturation curves of the wild-
type and mutant proteins are summarized in Table 2.
The apparent free energy changes of unfolding in the
absence of denaturant, DG(H
2
O), and the urea concen-
Destabilization mechanism of psychrotrophic RNase HI M. S. Rohman et al.
604 FEBS Journal 276 (2009) 603–613 ª 2008 The Authors Journal compilation ª 2008 FEBS
trations of the midpoints of the denaturation curves,
C
m
, of the mutant proteins were higher than those of
the wild-type protein by 3.5 kJÆmol
)1
and 0.4 m,
respectively, for R97G-RNase HI, and 6.1 kJÆmol
)1

and 0.7 m, respectively, for D136H-RNase HI. Thus,
the stabilities of the mutant proteins against urea-
induced denaturation show good agreement with those
against thermal denaturation.
Stabilization of 4
·
-RNase HI with Arg97
fi
Gly
and Asp136
fi
His mutations
To examine whether the Arg97 fi Gly and Asp136 fi
His mutations stabilize the quadruple mutant protein
of So-RNase HI (4·-RNase HI), in which the four
thermostabilizing mutations identified by directed evo-
lution are combined, the quintuple (5·-RNase HI) and
sextuple (6·-RNase HI) mutant proteins of So-RNase
HI were constructed. The 5·-RNase HI and 6·-RNase
HI proteins represent the 4·-RNase HI derivatives
with additional Arg97 fi Gly mutation and additional
Arg97 fi Gly and Asp136 fi His mutations, respec-
tively. These mutant proteins were overproduced in
E. coli and purified to give a single band on SDS-
PAGE like the wild-type protein (data not shown).
The far-UV CD spectra of these mutant proteins were
similar to that of the wild-type protein (data not
shown), suggesting that the quintuple and sextuple
mutations do not seriously affect the protein confor-
mation. The specific activities of the 5·-RNase HI and

6·-RNase HI proteins were 65% and 43%, respec-
tively, of that of the wild-type protein, and 93% and
66%, respectively, of that of the 4·-RNase HI protein
(Table 1). These results suggest that the effects of the
Arg97 fi Gly and Asp136 fi His mutations on the
enzymatic activity of the protein are not seriously
Table 1. Activities and thermostabilities of So-RNase HI and its derivatives.
Protein
Specific activity
a
(unitsÆmg
)1
)
Relative
activity
a
(%)
T
m
b
(°C)
DT
m
b
(°C)
DDG
m
b
(kJÆmol
)1

)
DH
m
b
(kJÆmol
)1
)
So-RNase HI 7.8 100 30.4 – – 217
R97G-RNase HI 7.7 99 35.8 5.4 3.9 264
D136H-RNase HI 5.1 65 40.1 9.7 7.0 267
4·-RNase HI 5.5 70 49.1 18.7 13.5 359
5·-RNase HI 5.1 65 52.5 22.1 15.9 366
6·-RNase HI 3.4 43 59.2 28.8 20.7 433
Ec-RNase HI
c
9.1 120 52.8 22.4 – 325
a
The enzymatic activity was determined at 30 °C using M13 DNA ⁄ RNA hybrid as a substrate, as described in Experimental procedures.
Each experiment was carried out at least twice and the average value is shown. Errors are within 15% of the values reported.
b
Parameters
characterizing the thermal denaturation of So-RNase HI and its derivatives. The thermal denaturation curves of these proteins were mea-
sured at pH 5.5 in the presence of 1
M GdnHCl. The thermal denaturation of these proteins was reversible in this condition. The melting
temperature (T
m
) is the temperature of the midpoint of the thermal denaturation transition. The difference in the melting temperature
between the wild-type and mutant proteins (DT
m
) was calculated as T

m
(mutant))T
m
(wild-type). DH
m
is the enthalpy change of unfolding at
T
m
calculated by van’t Hoff analysis. The difference between the free energy change of unfolding of the mutant protein and that of the wild-
type protein at T
m
(DDG
m
) was estimated by the equation, DDG
m
= DT
m
DS
m
(wild-type), where DS
m
(wild-type) is the entropy change of the
wild-type protein at T
m
[44]. The DS
m
(wild-type) value of 0.72 kJÆmol
)1
, which has been determined previously [11], was used to calculate the
DDG

m
values. Each experiment was carried out at least twice and the average value is shown. Errors are within ± 0.3 °C for T
m
, ± 26 kJÆmol
)1
for DH
m
, ± 0.12 kJÆmol
)1
ÆK
)1
for DS
m
and ± 0.3 kJÆmol
)1
for DDG
m
.
c
Data from Tadokoro et al. [11].
Table 2. Parameters characterizing the urea-induced denaturation
of So-RNase HI and its derivatives
a
.
Protein
C
m
a
(M)
M

a
(kJÆmol
)1
ÆM
)1
)
DG
(H
2
O)
a
(kJÆmol
)1
)
DDG
(H
2
O)
a
(kJÆmol
)1
)
So-RNase HI 2.6 8.5 22.3 –
R97G-RNase HI 3.0 8.9 26.3 3.5
D136H-RNase HI 3.3 9.0 30.1 6.1
4·-RNase HI 4.0 9.3 37.3 12.2
5·-RNase HI 4.9 8.5 41.3 20.0
6·-RNase HI 5.7 8.1 45.9 27.0
Ec-RNase HI
b

4.3 8.2 34.8 12.5
a
The urea-induced denaturation curves of these proteins were
measured at pH 5.5 and 20 °C. Urea-induced denaturation of these
proteins was reversible in this condition. The urea concentration of
the midpoint of the urea-induced denaturation curve (C
m
), the mea-
surement of the dependence of DG on the urea concentration (m),
and the free energy change of unfolding in H
2
O[DG(H
2
O)] were
calculated from the urea-induced denaturation curves. The differ-
ence in DG(H
2
O) [DDG(H
2
O)] between the wild-type and mutant
proteins was calculated using the equation: DDG(H
2
O) = m
av
DC
m
,
where m
av
represents the average m value (8.7 kJÆmol

)1
ÆM
)1
) and
DC
m
= C
m
(mutant))C
m
(wild-type). Each experiment was carried out
at least twice and the average value is shown. Errors are within
± 0.1
M for C
m
, ± 0.8 kJÆmol
)1
ÆM
)1
for m and ± 1.0 kJÆmol
)1
for
DG(H
2
O).
b
Data from Tadokoro et al. [11].
M. S. Rohman et al. Destabilization mechanism of psychrotrophic RNase HI
FEBS Journal 276 (2009) 603–613 ª 2008 The Authors Journal compilation ª 2008 FEBS 605
changed regardless of whether they are introduced into

So-RNase HI or 4·-RNase HI.
The stabilities of the 4·-RNase HI, 5·-RNase HI
and 6·-RNase HI proteins against thermal denatur-
ation were analysed as described for R97G-RNase HI
and D136H-RNase HI. Thermal denaturation of these
proteins was fully reversible in this condition. The
thermodynamic parameters characterizing the thermal
denaturation curves of these proteins are summarized
in Table 1. The temperature of the midpoint of the
transition, T
m
, was 49.1 °C for 4·-RNase HI, 52.5 °C
for 5·-RNase HI and 59.2 °C for 6·-RNase HI. Thus,
the 5·-RNase HI protein is more stable than the wild-
type and 4·-RNase HI proteins by 22.1 and 3.4 °C,
respectively, in T
m
, and 15.9 and 2.4 kJÆmol
)1
, respec-
tively, in DDG
m
. The 6·-RNase HI protein is more
stable than the wild-type, 4·-RNase HI and 5·-RNase
HI proteins by 28.8, 10.1 and 6.7 °C, respectively, in
T
m
, and 20.7, 7.2 and 4.8 kJÆmol
)1
, respectively, in

DDG
m
.
The stabilities of the 4·-RNase HI, 5·-RNase HI
and 6·-RNase HI proteins against urea-induced dena-
turation were also analysed by monitoring the change
in the CD values at 220 nm. Urea-induced denatur-
ation of these proteins was fully reversible and showed
a two-state transition. The thermodynamic parameters
characterizing the urea-induced denaturation curves of
the wild-type and mutant proteins are summarized in
Table 2. The DG(H
2
O) and C
m
values of the 5·-RNase
HI protein were higher than those of the wild-type and
4·-RNase HI proteins by 20.0 and 7.8 kJÆmol
)1
and
2.3 and 0.9 m, respectively. The DG(H
2
O) and C
m
values of the 6·-RNase HI protein were higher than
those of the wild-type, 4·-RNase HI and 5·-RNase
HI proteins by 27.0, 14.8 and 7.0 kJÆmol
)1
and 3.1, 1.7
and 0.8 m, respectively. Thus, the stabilities of the

5·-RNase HI and 6·-RNase HI proteins against urea-
induced denaturation show good agreement with those
against thermal denaturation, although the DDG(H
2
O)
and DDG
m
values are significantly different from each
other for these proteins.
Overall structure of 6
·
-RNase HI
The crystal structure of the 6·-RNase HI protein with
the sextuple thermostabilizing mutations was deter-
mined at 2.5 A
˚
resolution. The asymmetric unit of the
crystal structure consists of four protein molecules
(A–D). The structures of these four protein molecules
are virtually identical with one another with rmsd val-
ues of 0.73 A
˚
between molecules D and A, 0.59 A
˚
between molecules D and B, and 0.61 A
˚
between mole-
cules D and C for 148 Ca atoms. In the structures of
these protein molecules, however, three N-terminal
(Met1–Glu3) and four C-terminal (Gln155–Ser158) res-

idues are disordered. In the structures of molecules A
and B, a part of the loop between the bE strand and
aV helix (Ala127–His129) is also disordered. We used
the structure of molecule D in this study.
The overall structure of 6·-RNase HI is essentially
the same as that of the wild-type protein (Fig. 1A).
The rmsd value between the wild-type and 6·-RNase
HI proteins is 0.85 A
˚
for 148 Ca atoms. The shifts of
the Ca coordinates of 6·-RNase HI relative to those
of the wild-type protein are shown in Fig. 2. The dif-
ferences between the Ca coordinates of molecules C
and D are also shown in this figure as a reference.
Relatively large shifts were observed around Gly17
and Asn18 in a turn between the bA and bC strands,
around Ser95 in a loop between the aIII and aIV
Fig. 1. Stereoview of the three-dimensional structure of 6·-RNase HI. The structure of molecule D of 6·-RNase HI (gold) is superimposed
on the structure of the wild-type protein (green). The entire structure (A), and the structures around residue 29 (B), residue 39 (C), residue
76 (D), residues 90 and 97 (E) and residue 136 (F) are shown. The side-chains of the amino acid residues are shown as stick models, in
which the oxygen, nitrogen and sulfur atoms are coloured red, blue and yellow, respectively. The PDB code for the wild-type protein is
2E4L. For the entire structure (A), N and C represent the N- and C-termini of the protein, and a and b represent the a helix and b strand,
respectively. The side-chains of the mutated and parent amino acid residues at the six mutation sites are shown. D39 ⁄ G39 and M76 ⁄ V76
are simply labelled as 39 and 76, respectively. The side-chains of the five active site residues, D12, E50, D72, H126 and D136, are also
shown. For the structure around residue 29 (B), the side-chains of residues 29, T34 and E131 are shown. The hydrogen bonds between
N29 and Oc and T34 and Oc, and between N29 and Nd and E131 and Oe2, in the wild-type protein are shown as green broken lines, and
the ion pair between the e-amino group of K29 and the carboxyl group of E131 in 6·-RNase HI is shown as a gold broken line, together with
the distances. For the structure around residue 39 (C), the side-chains of Y24, residue 39, F41 and Q149 are shown. The hydrogen bonds
between D39 and Od1 and Q149 and Ne, and between D39 and Od2 and Q149 and Ne, in the wild-type protein are shown as green broken
lines together with the distances. For the structure around residue 76 (D), the side-chains of L51, P54, residue 76, W106, L109, W120 and

W122 are shown. For the structure around residues 90 and 97 (E), the side-chains of K89, residue 90 and residue 97 are shown. The dis-
tances between the e-amino groups of K89 and K90, and between the e-amino group of K90 and the guanidino group of R97, in the wild-
type protein are shown. For the structure around residue 136 (F), the side-chains of D12, E50, D72, H126, E133 and residue 136 are shown.
A p-stacking interaction between His126 and His136 in 6·-RNase HI is shown as a gold broken line, together with the distance.
Destabilization mechanism of psychrotrophic RNase HI M. S. Rohman et al.
606 FEBS Journal 276 (2009) 603–613 ª 2008 The Authors Journal compilation ª 2008 FEBS
A
BC
D
F
E
M. S. Rohman et al. Destabilization mechanism of psychrotrophic RNase HI
FEBS Journal 276 (2009) 603–613 ª 2008 The Authors Journal compilation ª 2008 FEBS 607
helices, and around Gly128 in a loop between the aV
helix and bE strand. The shifts around Gly17 and
Asn18 are probably a result of fluctuations rather than
perturbations caused by the mutations, because any
mutation site is located close to this region. The shifts
around Ser95 are probably a result of the
Lys90 fi Asn and ⁄ or Arg97 fi Gly mutations, and
those around Gly128 are probably caused by the
Asp136 fi His mutation. The details of these shifts are
described in the Discussion section.
The solvent accessibilities of the amino acid residues
that are located around the mutation sites, including
the parent and mutated residues at these sites, were
calculated on the basis of their accessible surface areas
in a native and extended structure. Comparison of
these values for the wild-type and 6 ·-RNase HI
proteins indicated that the solvent accessibilities of all

residues, except for residues 29, 39, 126 and 133, were
not seriously changed by the sextuple mutations. The
solvent accessibilities of residues 29 and 39 were signifi-
cantly increased from 20 to 39 A
˚
2
and decreased from
44 to 17 A
˚
2
by the Asn29 fi Lys and Asp39 fi Gly
mutations, respectively. The solvent accessibilities of
His126 and Glu133 were significantly decreased from
68 to 35 and 60 to 44 A
˚
2
, respectively, as a result in
the shift of a loop between the aV helix and bE strand.
Discussion
In this study, we have shown that the simultaneous
introduction of six thermostabilizing mutations causes
conformational changes predominantly around the
respective mutation sites, and has only a slight effect
on the backbone conformation of the protein. This
result indicates that the mutations affect the protein
structure independently, but not cooperatively, and is
consistent with the fact that the thermostabilizing
effects of the mutations are roughly additive. The pos-
sible stabilization mechanism of the protein by each
mutation is described below, based on a local confor-

mational change caused by each mutation.
Asn29
fi
Lys
The 6·-RNase HI structure around residue 29 is com-
pared with that of the wild-type protein in Fig. 1B.
Asn29 and Lys29 are located in the bB strand and are
partially exposed to the solvent by 20 and 39%, respec-
tively. In the wild-type protein, Asn29 forms hydrogen
bonds with Thr34 and Glu131, which are located in the
bC strand and aV helix, respectively. In 6·-RNase HI,
Lys29 forms an ion pair with Glu131. The distances
between the Nf atom of Lys29 and the Oe2 atom of
Glu131 are 2.7, 4.1, 3.3 and 3.3 A
˚
for molecules A, B,
C and D, respectively. Thus, by the Asn29 fi Lys
mutation, one ion pair is introduced and two hydrogen
bonds are eliminated at the mutation site. Both the
hydrogen bond and ion pair have been reported to
contribute to protein stabilization [16,17]. However, the
finding that So-RNase HI is stabilized by the
Asn29 fi Lys mutation by 3.6 °CinT
m
and 3.5
kJÆmol
)1
in DG(H
2
O) [13] suggests that the stabilization

effect caused by the introduction of an ion pair at the
mutation site is stronger than the destabilization effect
caused by the elimination of two hydrogen bonds at the
same site. Several proteins have also been reported to
be stabilized by the introduction of ion pairs [18–21].
Asp39
fi
Gly
So-RNase HI is stabilized by the Asp39 fi Gly muta-
tion by 5.8 °CinT
m
and 3.5 kJÆmol
)1
in DG(H
2
O)
[13]. The 6·-RNase HI structure around residue 39 is
compared with that of the wild-type protein in
Fig. 1C. The deviation in the shifts of this residue in
molecules A–D is less than 0.4 A
˚
. Asp39 and Gly39
are located in the bC strand and exposed to the sol-
vent by 44 and 17%, respectively. In the vicinity of
residue 39, Tyr24, Phe41 and Gln149 are located.
Tyr24 and Phe41 are almost fully buried inside the
protein molecule, whereas Gln149 is relatively well
exposed to the solvent. We have shown previously that
the Asp39 fi Ala mutation also stabilizes the protein
to a similar level as that of D39G-RNase HI [13].

Fig. 2. Displacement of the Ca coordinates between the 6·-RNase
HI and wild-type proteins (full line) and between molecules C and D
(broken line). a Helices and b strands are indicated by bars.
Destabilization mechanism of psychrotrophic RNase HI M. S. Rohman et al.
608 FEBS Journal 276 (2009) 603–613 ª 2008 The Authors Journal compilation ª 2008 FEBS
Asp39 is changed to Ala (Ala37) in Ec-RNase HI,
which is buried inside the protein molecule by 83%.
Therefore, the Asp39 fi Gly mutation stabilizes the
protein, probably because hydrophobic interactions
around the mutation site increase. In the wild-type
protein, Asp39 forms hydrogen bonds with Gln149.
However, these hydrogen bonds may not seriously
contribute to the stabilization of the protein, because
the hydrogen bond partner, Gln149, can form hydr-
ogen bonds with water molecules.
Met76
fi
Val
So-RNase HI is stabilized by the Met76 fi Val muta-
tion by 6.7 °CinT
m
and 5.2 kJÆmol
)1
in DG(H
2
O)
[13]. The 6·-RNase HI structure around residue 76 is
compared with that of the wild-type protein in
Fig. 1D. Met76 and Val76 are located in the aII helix
within a hydrophobic core and almost fully buried

inside the protein molecule by more than 98%. The
structures of the 6·-RNase HI and wild-type proteins
have a cavity around residue 76 within a hydrophobic
core. The volume of this cavity is 92 A
˚
3
for the wild-
type protein, and 110, 113, 112 and 111 A
˚
3
for mole-
cules A, B, C and D, respectively, of 6·-RNase HI,
indicating that the cavity volume increases with the
Met76 fi Val mutation by roughly 20 A
˚
3
, which is
comparable with the volume of a methylene group.
The side-chain of Met is larger than that of Val, and
the difference between them is equivalent to one meth-
ylene group in size. Therefore, the decrease in the size
of the side-chain of residue 76 accounts for the
increase in the cavity volume by the mutation. We
have shown previously that Ec-RNase HI is stabilized
by filling a cavity with methyl or methylene groups
[22,23]. However, one of the mutant proteins of
Ec-RNase HI, in which a cavity is filled by the
Ala52 fi Met mutation, is less stable than another, in
which a cavity is filled by the Ala52 fi Val mutation,
by 3.9 °CinT

m
[23]. These results suggest that the fill-
ing of a cavity with Met is not as effective as the filling
of a cavity with Val with respect to protein stabiliza-
tion. The Met residue at the hydrophobic core is less
preferable than Val for protein stability, probably
because its solvation free energy is higher than that of
Val [24], and its linear side-chain is rotated more freely
than the branched one of Val [25].
Lys90
fi
Asn and Arg97
fi
Gly
The 6·-RNase HI structure around residues 90 and 97
is compared with that of the wild-type protein in
Fig. 1E. Lys90 and Arg97 are located in the C-terminal
region of the aIII helix and a long loop between the aIII
and aIV helices, respectively. In the vicinity of Lys90,
Lys89 is located. Lys89, Lys90 and Arg97 are well
exposed to the solvent by 80%, 69% and 95%, respec-
tively. So-RNase HI is stabilized by the Lys90 fi Asn
mutation by 4.1 °CinT
m
and 1.7 kJÆmol
)1
in DG(H
2
O)
[13]. It has been reported that the avoidance of unfa-

vourable electrostatic repulsions is more effective in
increasing protein stability than is the creation of
stabilizing surface ion pairs [26]. Therefore, the
Lys90 fi Asn mutation stabilizes the protein, probably
because positive charge repulsions between Lys90 and
Lys89 and ⁄ or between Lys90 and Arg97 are eliminated.
The Arg97 fi Gly mutation stabilizes the wild-type
and 4·-RNase HI proteins by 5.4 and 3.4 °C, respec-
tively, in T
m
, and 3.5 and 8.0 kJÆmol
)1
, respectively, in
DG(H
2
O). It has been reported that the Lys95 fi Gly
mutation stabilizes Ec-RNase HI by 6.8 °CinT
m
,
because the strain caused by the left-handed backbone
structure in the typical 3 : 5-type loop is eliminated
[14,27]. Non-Gly residues are energetically unfavour-
able for the left-handed helical conformation because of
the steric hindrance between the backbone oxygen atom
and side-chain Cb atom. The Arg97 fi Gly mutation
probably stabilizes the protein with a similar mecha-
nism. In fact, Arg97 in the structure of the wild-type
protein assumes a left-handed helical conformation with
the (/, w) values of (68.4°, 35.2°). This conformation is
not seriously changed by the Arg97 fi Gly mutation,

because the (/, w) values of Gly97 in the 6·-RNase HI
structure are (54.0°, 66.3°). The reason why the effects
of this mutation on the thermal stabilities of the wild-
type and 4·-RNase HI proteins are not consistent with
those on the conformational stabilities (stabilities
against urea denaturation) remains to be clarified.
It should be noted that a loop region (residues
94–97) is shifted towards the aIII helix at most by 0.5,
3.4, 1.5 and 3.0 A
˚
in the structures of molecules A, B,
C and D, respectively, of 6·-RNase HI when
compared with that in the structure of the wild-type
protein. As shown in Fig. 2, the largest shift is
observed for the Ca atom of Ser95. Elimination of the
positive charge repulsions among Lys89, Lys90 and
Arg97 may be responsible for this shift. However, the
mutation sites at residues 90 and 97 are close to the
protein–protein contacts in the crystal packing, which
may account for the large deviation in the loop shift
among the molecules A–D.
Asp136
fi
His
The Asp136 fi His mutation stabilizes the wild-type
and 5·-RNase HI by 9.7 and 6.7 °C, respectively in
M. S. Rohman et al. Destabilization mechanism of psychrotrophic RNase HI
FEBS Journal 276 (2009) 603–613 ª 2008 The Authors Journal compilation ª 2008 FEBS 609
T
m

, and 6.1 and 7.0 kJÆmol
)1
, respectively in DG(H
2
O),
indicating that the stabilizing effect of this mutation is
independent of those of the other five mutations. The
6·-RNase HI structure around residue 136 is com-
pared with that of the wild-type protein in Fig. 1F.
Asp136 and His136 are located in the aV helix and
exposed to the solvent by 42% and 39%, respectively.
In the structure of the wild-type protein, many acidic
residues, such as Asp12, Glu50, Asp72 and Glu133,
are clustered in the vicinity of Asp136. It has
been reported that the corresponding mutation
(Asp134 fi His) stabilizes Ec-RNase HI by 7.0 °Cin
T
m
as a result of elimination of negative charge repul-
sions [15]. The Asp136 fi His mutation probably
stabilizes the protein with a similar mechanism.
A loop containing His126 is greatly shifted by, at
most, 4.0 and 4.1 A
˚
in the structures of molecules C
and D, respectively, of 6· -RNase HI when compared
with that in the structure of the wild-type protein. This
shift is largest amongst those observed in the
6·-RNase HI structure (Fig. 2). With this shift, two His
residues, His126 and His136, make a p-stacking inter-

action (Fig. 1F). The distances of this interaction are
3.9 and 3.7 A
˚
for molecules C and D, respectively. A
p-stacking interaction has been reported to contribute
to protein stabilization [28]. However, this interaction
may not be a major stabilization factor of the mutant
protein with the Asp136 fi His mutation, because this
interaction is not observed in the structures of the
Ec-RNase HI variants with the corresponding mutation
[31,32]. According to the crystal structures of these
Ec-RNase HI variants, the position of His124, which
corresponds to His126 of So-RNase HI, varies for dif-
ferent proteins, because of the intrinsic flexibility of the
loop containing His124 and the crystal packing effect.
Destabilization mechanism of So-RNase HI
A combination of the six thermostabilizing mutations
increases the stability of So-RNase HI by 28.8 °Cin
T
m
and 27.0 kJÆmol
)1
in DG(H
2
O). Five of the six
substituted residues in the resultant sextuple mutant
protein (6·-RNase HI) are found in the corresponding
positions of at least one of the amino acid sequences
of its mesophilic and thermophilic counterparts. Lys29
is conserved as Arg27 in Ec-RNase HI and Arg31 in

Thermus thermophilus RNase HI (Tt-RNase HI).
Gly39 and Gly97 are conserved as Gly41 and Gly100
in Tt-RNase HI, respectively. Val76 is conserved as
Val74 in Ec-RNase HI. Asn90 is conserved as Gln96
in RNase HI from a cyanobacterium. Arg and Gln are
similar to Lys and Asn, respectively, in size and
charge ⁄ polarity. Another substituted residue His136 is
not found in other RNase H sequences, because the
original residue (Asp136) is one of the active site resi-
dues. However, a possibility that RNase H with His at
this position exists in nature cannot be ruled out,
because the mutation of this Asp residue to His greatly
stabilizes both So-RNase HI and Ec-RNase HI with-
out seriously affecting the activity. These results sug-
gest that So-RNase HI is destabilized when compared
with its mesophilic and thermophilic counterparts by
increasing the number of amino acid residues unfa-
vourable for protein stability in a localized fashion, in
which these residues independently contribute to the
destabilization of the protein.
Experimental procedures
Cells and plasmids
E. coli MIC2067 [F
)
,k
)
, IN(rrnD, rrnE)1, rnhA339::cat,
rnhB716::kam] was kindly donated by M. Itaya [31]. kDE3
lysogen of this strain, E. coli MIC2067(DE3), was con-
structed previously in our laboratory [32]. Plasmids

pET500M [11] and pET500M4x [13] for the overproduction
of So-RNase HI and 4·-RNase HI, respectively, were also
previously constructed in our laboratory.
Mutagenesis
The genes encoding R97G-RNase HI, D136H-RNase HI,
5·-RNase HI and 6·-RNase HI were constructed by site-
directed mutagenesis using PCR as described previously
[33]. Plasmid pET500M or pET500M4x was used as tem-
plate. The mutagenic primers were designed such that the
codons for Arg97 (CGT) and Asp136 (GAT) were changed
to those for Gly (GGT) and His (CAT), respectively. The
nucleotide sequences of the genes encoding the mutant pro-
teins were confirmed using a Prism 310 DNA sequencer
(Applied Biosystems, Tokyo, Japan). Overproduction and
purification of the wild-type and mutant proteins were
carried out as described previously [11]. The protein concen-
tration was determined from the UV absorption at 280 nm,
assuming that the absorption coefficient at this wavelength
(2.1 for 0.1% solution) was not changed by the mutation.
Enzymatic activity
The RNase H activity was determined at 30 °C and pH 8.0
by measuring the radioactivity of the acid-soluble digestion
product from
3
H-labelled M13 DNA ⁄ RNA hybrid, as
described previously [34]. The reaction mixture contained
10 pmol of the substrate and an appropriate amount of
enzyme in 20 lLof10mm Tris ⁄ HCl (pH 8.0) containing
10 mm MgCl
2

,50mm NaCl, 1 mm 2-mercaptoethanol and
50 lgÆmL
)1
BSA. One unit is defined as the amount of
Destabilization mechanism of psychrotrophic RNase HI M. S. Rohman et al.
610 FEBS Journal 276 (2009) 603–613 ª 2008 The Authors Journal compilation ª 2008 FEBS
enzyme producing 1 lmol of acid-soluble material per
minute. The specific activity was defined as the enzymatic
activity per milligram of protein.
CD spectra
The far-UV CD spectra were measured on a J-725 spectropo-
larimeter (Japan Spectroscopic, Tokyo, Japan) at 4 °C. The
protein was dissolved in 10 mm sodium acetate (pH 5.5).
The protein concentration and optical path length were 0.1–
0.2 mgÆmL
)1
and 2 mm, respectively. The mean residue ellip-
ticity h, which has units of degÆcm
2
Ædmol
)1
, was calculated
using an average amino acid molecular weight of 110.
Thermal denaturation
Thermal denaturation curves of So-RNase HI and its deriv-
atives were measured as described previously [11]. The pro-
teins were dissolved in 10 mm sodium acetate (pH 5.5)
containing 1 m GdnHCl. The protein concentration and
optical path length were 0.1–0.2 mgÆmL
)1

and 2 mm,
respectively. The temperature of the protein solution was
increased linearly by approximately 1.0 °CÆmin
)1
. Thermal
denaturation of these proteins was reversible in the presence
of 1 m GdnHCl. The temperature of the midpoint of the
transition, T
m
, was calculated from curve fitting of the
resultant CD values versus temperature data on the basis of
a least-squares analysis. The enthalpy (DH
m
) and entropy
(DS
m
) changes for thermal denaturation at T
m
were calcu-
lated by van’t Hoff analysis.
Urea-induced denaturation
Urea-induced denaturation curves of So-RNase HI and its
derivatives were measured at 20 °C as described previously
[11]. The proteins (0.1–0.2 mgÆmL
)1
) were dissolved in
10 mm sodium acetate (pH 5.5) containing 100 mm NaCl
and the appropriate concentrations of urea. The protein
solution was incubated for at least 2 h at 20 °C before the
measurement. The urea-induced denaturation of these pro-

teins was fully reversible. On the assumption that the
unfolding equilibria of these proteins follow a two-state
mechanism, the pre- and post-transition baselines were
extrapolated linearly, and the difference in free energy
between the folded and unfolded states, DG, and the free
energy change of unfolding in H
2
O, DG(H
2
O), were calcu-
lated by the equations given by Pace [35].
Crystallization and data collection
The 6·-RNase HI protein was concentrated using a Centr-
icon ultrafiltration system (Millipore, Billerica, MA, USA)
to approximately 10 mgÆmL
)1
. The crystallization condi-
tions were initially screened using crystallization kits from
Hampton Research (Alise Viejo, CA, USA) (Crystal
Screens I and II and Crystal Screen Cryo I) and Emerald
Biostructures (Bainbridge Island, WA, USA) (Wizard I and
II). The conditions were surveyed using a sitting-drops
vapour diffusion method at 4 °C. Drops were prepared by
mixing 1 lL each of the protein and reservoir solutions,
and were vapour equilibrated against 100 lL of reservoir
solution. Native crystals suitable for X-ray diffraction anal-
ysis appeared after 2 weeks using Crystal Screen II solution
No. 26 [30% poly(ethylene glycol), MME 5000, 0.1 m Mes,
pH 6.5, 0.2 m ammonium sulfate]. The crystal was cryopro-
tected in mother liquor containing 20% sucrose prior to

mounting for X-ray diffraction.
Structure determination and refinement
X-Ray diffraction data sets of the 6·-RNase HI crystal
were collected at 100 K using synchrotron radiation at the
BL44XU station in SPring-8, using a DIP6040 multiple
Table 3. Data collection and refinement statistics for 6·-RNase HI.
Beamline BL44XU
Wavelength (A
˚
) 1.0
Resolution (A
˚
) 50.0–2.49 (2.59)2.49)
Observations 329 997
Unique reflections 23 782
Completeness (%) 100 (100)
R
merge
(%)
a
14.6 (52.5)
Average I ⁄ r(I) 28.4 (6.39)
Refinement
Resolution limit (A
˚
) 47.52–2.49
Space group P4
1
2
1

2
Cell unit (A
˚
) a = b = 68.24, c = 272.82
a = b = c =90°
No. of molecules 4
No. of protein atoms 4721
No. of water molecules 255
R-factor (%) 19.3
R
free
(%)
b
24.7
rmsd
Bond length (A
˚
) 0.025
Bond angles (deg) 2.316
Mean B factors (A
˚
2
)
Main chain 26.99
Side-chain 29.30
Ramachandran plot statistics (%)
Most favoured regions 88.3
Additionally allowed regions 10.9
Generously allowed regions 0.8
a

R
merge
=
P
I
hkl
) <I
hkl
> ⁄
P
I
hkl
, where I
hkl
is the intensity measure-
ment for reflections with indices hkl and <I
hkl
> is the mean inten-
sity for multiply recorded reflections.
b
R
free
was calculated using
5% of the total reflections chosen randomly and omitted from
refinement.
M. S. Rohman et al. Destabilization mechanism of psychrotrophic RNase HI
FEBS Journal 276 (2009) 603–613 ª 2008 The Authors Journal compilation ª 2008 FEBS 611
imaging plate diffractometer (Bruker AXS Inc., Madison,
WI, USA). These data sets were indexed, integrated and
scaled using the hkl2000 program [36]. The crystal struc-

ture was solved by the molecular replacement method using
molrep [37] in the ccp4 program suite [38]. There were four
molecules per asymmetric unit, with a solvent content of
47% and Matthews coefficient of 2.34 [39]. The refined 2 A
˚
structure of So-RNase HI [Protein Data Bank (PDB) code
2E4L] was used as a starting model. Refinement of the
structure was performed using the programs cns [40] and
refmac [41]. The final model was built using coot [42],
with R-factor and R
free
values of 19.3% and 24.7%, respec-
tively. The Ramachandran plot produced by procheck [43]
shows that 100% of the residues in the structure fall in the
most favoured and allowed regions. The statistics for data
collection and refinement are summarized in Table 3. The
figures were prepared using pymol ().
The accessible surface areas of the protein in native and
extended structures were calculated using ACCESS_Surf of
MSI InsightII Ver. 2000 module (Molecular Simulation
Inc. ⁄ Accelrys Inc., San Diego, CA, USA). The extended
structure was built using Biopolymer of the same module.
The volume of the cavity within the hydrophobic core was
calculated using voidoo software [44].
PDB accession number
The coordinates and structure factors of 6·-RNase HI have
been deposited in PDB under accession code 2ZQB.
Acknowledgements
The synchrotron radiation experiments were performed
at the beam line BL44XU in SPring-8 with the

approval of the Institute for Protein Research, Osaka
University, Osaka, Japan (2008A6909). This work was
supported in part by a Grant-in-Aid for Scientific
Research on Priority Areas ‘Systems Genomics’ from
the Ministry of Education, Culture, Sports, Science,
and Technology of Japan, and by an Industrial Tech-
nology Research Grant Program from the New Energy
and Industrial Technology Development Organization
(NEDO) of Japan.
References
1 Morita RY & Moyer CL (2000) Origin of psychro-
philes. In Encyclopedia of Biodiversity, Vol. 4 (Levin
SA, Colwell R, Dailey G, Lubchenco J, Mooney HA,
Schulze ED & Tilman GD, eds), pp. 917–924. Academic
Press, San Diego, CA.
2 Feller G, Narinx E, Arpigny JL, Aittaleb M, Baise E,
Genicot S & Gerday C (1996) Enzymes from psychro-
philic organisms. FEMS Microbiol Rev 18, 189–202.
3 Collins T, Meuwis M-A, Gerday C & Feller G (2003)
Activity, stability and flexibility in glycosidases adapted
to extreme thermal environments. J Mol Biol 328, 419–
428.
4 D’Amico S, Marx JC, Gerday C & Feller G (2003)
Activity–stability relationships in extremophilic
enzymes. J Biol Chem 278, 7891–7896.
5 Smalas AO, Leiros HK, Os V & Willassen NP (2000)
Cold adapted enzymes. Biotechnol Annu Rev 6, 1–57.
6 Gianese G, Bossa F & Pascarella S (2002) Comparative
structural analysis of psychrophilic and meso- and ther-
mophilic enzymes. Proteins 47, 236–249.

7 Feller G & Gerday C (2003) Psychrophilic enzymes: hot
topics in cold adaptation. Nat Rev Microbiol 1, 200–
208.
8 Siddiqui KS & Cavicchioli R (2006) Cold-adapted
enzymes. Annu Rev Biochem 75, 403–433.
9 Crouch RJ & Dirksen ML (1982) Ribonuclease H. In
Nuclease (Linn SM & Robert RJ, eds), pp. 211–241.
Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, NY.
10 Ohtani N, Haruki M, Morikawa M & Kanaya S (1999)
Molecular diversities of Rnases H. J Biosci Bioeng 88,
12–19.
11 Tadokoro T, You D-J, Abe Y, Chon H, Matsumura H,
Koga Y, Takano K & Kanaya S (2007) Structural, ther-
modynamic, and mutational analyses of a psychro-
trophic RNase HI. Biochemistry 46, 7460–7468.
12 Kanaya S (1998) Enzymatic activity and protein stabil-
ity of E. coli ribonuclease HI. In Ribonucleases H
(Crouch RJ & Toulme JJ, eds), pp. 1–38, INSERM,
Paris.
13 Tadokoro T, Matsushita K, Abe Y, Rohman MS,
Koga Y, Takano K & Kanaya S (2008) Remarkable
stabilization of a psychrotrophic RNase HI by combi-
nation of thermostabilizing mutations identified by sup-
pressor mutation method. Biochemistry 47, 8040–8047.
14 Kimura S, Kanaya S & Nakamura H (1992) Ther-
mostabilization of Escherichia coli ribonuclease HI by
replacing left-handed helical Lys95 with Gly or Asn.
J Biol Chem 267, 22014–22017.
15 Haruki M, Noguchi E, Nakai C, Liu Y-Y, Oobatake

M, Itaya M & Kanaya S (1994) Investigating the role
of conserved residue Asp134 in Escherichia coli ribonu-
clease HI by site-directed random mutagenesis. Eur J
Biochem 220, 623–631.
16 Alber T (1989) Mutational effects on protein stability.
Annu Rev Biochem 58, 765–798.
17 Dill KA (1990) Dominant forces in protein folding.
Biochemistry 29, 7133–7155.
18 Spek EJ, Bui AH, Lu M & Kallenbach NR (1998)
Surface salt bridges stabilize the GCN4 leucine zipper.
Protein Sci 7, 2431–2437.
19 Sanz-Aparicio J, Hermoso JA, Martı
´
nez-Ripoll M,
Gonza
´
lez B, Lo
´
pez-Camacho C & Polaina J (1998)
Destabilization mechanism of psychrotrophic RNase HI M. S. Rohman et al.
612 FEBS Journal 276 (2009) 603–613 ª 2008 The Authors Journal compilation ª 2008 FEBS
Structural basis of increased resistance to thermal dena-
turation induced by single amino acid substitution in
the sequence of beta-glucosidase A from Bacillus
polymyxa. Proteins 33, 567–576.
20 Bae E & Phillips GN Jr (2005) Identifying and engi-
neering ion pairs in adenylate kinases. Insights from
molecular dynamics simulations of thermophilic and
mesophilic homologues. J Biol Chem 280 , 30943–
30948.

21 Koga Y, Katsumi R, You D-J, Matsumura H, Takano
K, Kanaya S (2008) Crystal structure of highly thermo-
stable glycerol kinase from a hyperthermophilic archa-
eon in a dimeric form. FEBS J 275, 2632–2643.
22 Ishikawa K, Nakamura H, Morikawa K & Kanaya S
(1993) Stabilization of Escherichia coli ribonuclease HI
by cavity-filling mutations within a hydrophobic core.
Biochemistry 32, 6171–6178.
23 Akasako A, Haruki M, Oobatake M & Kanaya S
(1997) Conformational stabilities of E. coli RNase HI
variants with a series of amino acid substitutions at a
cavity within the hydrophobic core. J Biol Chem 272,
18686–18693.
24 Shirts MR & Pande VS (2005) Solvation free energies
of amino acid side chain analogs for common molecular
mechanics water models. J Chem Phys 122, 134508.
25 Doig AJ & Sternberg MJ (1995) Side-chain conforma-
tional entropy in protein folding. Protein Sci 4, 2247–
2251.
26 Perl D & Schmid FX (2001) Electrostatic stabilization
of a thermophilic cold shock protein. J Mol Biol 313,
343–357.
27 Ishikawa K, Kimura S, Kanaya S, Morikawa K &
Nakamura H (1993) Structural study of mutants of
Escherichia coli ribonuclease HI with enhanced thermo-
stability. Protein Eng 6, 85–91.
28 McGaughey GB, Gagne
´
M & Rappe
´

AK (1998)
p-Stacking interactions. Alive and well in proteins.
J Biol Chem 273, 15458–15463.
29 Kashiwagi T, Jeanteur D, Haruki M, Katayanagi M,
Kanaya S & Morikawa K (1996) Proposal of new cata-
lytic roles for two invariant residues in Escherichia coli
ribonuclease HI. Protein Eng 9, 857–867.
30 Haruki M, Tanaka M, Motegi T, Tadokoro T, Koga
Y, Takano K & Kanaya S (2007) Structural and ther-
modynamic analyses of Escherichia coli ribonuclease HI
variant with quintuple thermostabilizing mutations.
FEBS J 274, 5815–5825.
31 Itaya M, Omori A, Kanaya S, Crouch RJ, Tanaka T &
Kondo K (1999) Isolation of RNase H genes that are
essential for growth of Bacillus subtilis 168. J Bacteriol
181, 2118–2123.
32 Ohtani N, Haruki M, Muroya A, Morikawa M &
Kanaya S (2000) Characterization of ribonuclease HII
from Escherichia coli overproduced in a soluble form.
J Biochem 127, 895–899.
33 Horton RM, Cai ZL, Ho SN & Pease LR (1990) Gene
splicing by overlap extension: tailor-made genes using the
polymerase chain reaction. Biotechniques 8, 528–535.
34 Kanaya S, Katsuda C, Kimura S, Nakai T, Kitakuni E,
Nakamura H, Katayanagi K, Morikawa K & Ikehara
M (1991) Stabilization of Escherichia coli ribonuclease
H by introduction of an artificial disulfide bond. J Biol
Chem 266, 6038–6044.
35 Pace CN (1990) Measuring and increasing protein
stability. Trends Biotechnol 8, 93–98.

36 Otwinowski Z & Minor W (1997) Processing of X-ray
diffraction data collected in oscillation mode. Methods
Enzymol 276, 307–326.
37 Collaborative Computational Project, Number 4 (1994)
The CCP4 suite: programs for protein crystallography.
Acta Crystallogr Sect D 50, 760–763.
38 Matthews BW (1968) Solvent content of protein
crystals. J Mol Biol 33, 491–497.
39 Brunger AT, Adams PD, Clore GM, DeLano WL, Gros
P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges
M, Pannu NS et al. (1998) Crystallography and NMR
system: a new software suite for macromolecular struc-
ture determination. Acta Crystallogr Sect D 54, 905–921.
40 Murshudov GN, Vagin AA & Dodson EJ (1997) Refine-
ment of macromolecular structures by the maximum-like-
lihood method. Acta Crystallogr Sect D 53, 240–255.
41 Emsley P & Cowtan K (2004) Coot: model-building
tools for molecular graphics. Acta Crystallogr Sect D
60, 2126–2132.
42 Laskowski RA, MacArthur MW, Moss DS & Thornton
JM (1993) PROCHECK: a program to check the
stereochemical quality of protein structures. J Appl
Crystallogr 26, 283–291.
43 Kleywegt GJ & Jones TA (1994) Detection, delineation,
measurement and display of cavities in macromolecular
structures. Acta Crystallogr D Biol Crystallogr 50, 178–
185.
44 Becktel WJ & Schellman JA (1987) Protein stability
curves. Biopolymers 26, 1859–1877.
M. S. Rohman et al. Destabilization mechanism of psychrotrophic RNase HI

FEBS Journal 276 (2009) 603–613 ª 2008 The Authors Journal compilation ª 2008 FEBS 613