Molecular evolution of shark and other vertebrate DNases I
Toshihiro Yasuda
1
, Reiko Iida
2
, Misuzu Ueki
1
, Yoshihiko Kominato
3
, Tamiko Nakajima
3
, Haruo Takeshita
3
,
Takanori Kobayashi
4
and Koichiro Kishi
3
1
Division of Medical Genetics and Biochemistry and
2
Division of Legal Medicine, Faculty of Medical Sciences, University of Fukui,
Japan;
3
Department of Legal Medicine, Gunma University Graduate School of Medicine, Japan;
4
National Research Institute of
Fisheries Science, Japan
We purified pancreatic deoxyribonuclease I (DN ase I) from
the shark Heterodontus japonicus using three-step column
chromatography. Although its enzymatic properties resem-

such as the pancreas and parotid glands, from which it is
secreted into the alimentary tract to hydrolyse exogenous
DNA [1–3]. Recently, it has been demonstrated that
DNase I-deficient mice have an increased incidence of
systemic lupus erythematosus (SLE), with classical findings
including the presence of autoreactive antibodies and
glomerulonephritis occurring in a DNase I-level-dependent
manner; this suggests that DNase I may protect against
autoimmunity by digesting extracellular nucleoprotein [4].
Furthermore, serum DNase I activity levels h ave been
reported to be lower in SLE patients than i n healthy
subjects, resulting in expansion of the autoreactive lympho-
cytes that react with nucleosomal antigens [5,6]. Thus, it is
plausible that DNase I activity must be maintained at a
certain level in the serum to prevent the initiation of SLE.
We have also found that serum DNase I activity levels w ere
transiently reduced by somatostatin through an effect on
gene expression [7], and were elevated at the onset of acute
myocardial infarction [8]. These, together with other
findings suggesting t hat DNase I or D Nase I-like e ndo-
nucleases may be responsible for internucleosomal DNA
degradation during apoptosis [9,10], have focused attention
on the potential physiological r oles of DNase I. In this
context, we have attempted to elucidate the intrinsic intra-
and extracellular f unction(s) of DNase I, as well a s the
phylogenetic origins of the vertebrate DNase I family, by
carrying out comprehensive comparisons of the enzymes
from lower and higher vertebrates: the biochemical
and molecular characterizations of mammalian [ 11–16],
avian [ 17], reptilian [18] and amphibian [19] DNases I

classes. Comprehensive characterization of Chondrichthye
DNases I will thus allow us t o elucidate the m olecular
evolutionary aspect of the vertebrate DNase I family.
In this s tudy, we cloned cDNAs encoding DN ases I
from two Chondrichthyes, Heterodontus japonicus and
Triakis scyllia, species of shark which are widely
distributed in the seas around Japan, and purified the
former’s enzyme. The expression of a series of m utant
constructs was a lso e xamined in mammalian cells,
allowing several common structural and functional char-
acteristics o f shark DNas es I to b e confirmed. The
molecular evolutionary aspect of the vertebrate DNase I
family is also discussed.
Materials and methods
Materials and biological samples
Two different species of shark, H. japonicus and T. scyllia,
weighing approximately 5 .0 kg (1.2 m long) and 4.7 kg
(1.0 m long), respectively, were obtained from T oba
Aquarium, Mie, Japan. LipofectaminPlus, all oligonucle-
otide primers, and the 3¢-and5¢-RACE systems were
obtained from Invitrogen; CM-Sepharose CL-6B, Mono
S 5/50 GL and Superdex 75 were from Amersham
Pharmacia Biotech; the Expanded High Fidelity PCR
system was from Boehringer Mannheim GmbH. Anti-
bodies s pecific to human, hen, Ra na catesbeiana (frog),
Elaphe quadrivirgata (snake) and Cyprin us carpio (carp)
DNases I were prepared using previously described
methods [11,17–19,21]. All other chemicals used were of
reagent grade and commercially available.
Analytical methods

pH 6.0, containing 1 m
M
phenylmethanesulfonyl fluoride
(buffer I). After centrifugation, the supernatant (crude
extract) was applied to a CM-Sepharose CL-6B column
(1.6 · 8 cm) pre-equilibrated w ith t he same buffer. The
adsorbed materials were eluted with a 200-mL linear
gradient of 0–1
M
NaCl in buffer I. The DNase I-active
fractions eluted with 0.2
M
NaCl were colle cted and dialysed
against buffer I containing 10 m
M
CaCl
2
. The dialysate was
subjected to cation exchange chromatography using t he
A
¨
KTAFPLC
1
system (Amersham Pharmacia Biotech)
equipped with a Mono S 5/50 GL column ( 0.46 · 10 cm).
The adsorbed materials were eluted with a 100-mL linear
gradient of 0–1
M
NaCl in buffer I. The active fractions
eluted with 0.2

turer’s instructions. The RACE products were subcloned
into the pCR 2.1 TA cloning vector (Invitrogen) and
sequenced. The nucleotide sequences were determined by
the dideoxy chain-termination method using a Dye Termi-
nator Cycle Sequencing kit (Applied Biosystems). The
sequencing run was performed on a Genetic Analyzer
(model 310, Applied Biosystems) and all the DNA
sequences were confirmed by reading both strands.
Construction of expression vectors and transient
expression of the constructs in mammalian cells
A DNA fragment containing the entire coding sequence of
H. japonicus DNase I cDNA, corresponding to both the
signal sequence and mature enzyme regions, was prepared
from the total RNA derived from the pancreas by reverse
transcription/PCR amplification using an Expanded High
Fidelity PCR system with a set of two primers correspond-
ing to the nucleotide sequences of the cDNA at positions
48–69 and 881–901, respectively. The amplified fragment
was ligated into a p cDNA3.1(+) vector (Invitrogen) to
construct the wild-type expression vector. Expression vec-
tors for wild-type human, frog, Anguilla japonica (eel) and
T. scyllia DNases I were constructed in the same manner.
A chimeric mutant, H. japonicus-chi(human:sig), in which
the signal sequence region of H. japonicus DNase I was
replaced by its counterpart from the human enzyme, was
constructed by splicing using the overlap extension method
[27] with each of the wild-type constructs as a template.
Seven other chimeric mutants, human-chi(H. japonicus:sig),
Ó FEBS 2004 Molecular evolution of vertebrate DNase I family (Eur. J. Biochem. 271) 4429
-chi(T. scyllia:sig), -chi(eel:sig), -chi(frog:sig), frog-

lysates were assayed for b-
D
-galactosidase. All transfections
were performed in triplicate w ith a t least two different
plasmid preparations.
Results and Discussion
Purification and characterization of shark pancreatic
DNase I
Among various tissue samples collected from H. japonicus,
the pancreas s howed the h ighest DNase I activity
(1.3 ± 0.31 UÆmg
)1
protein); moderate activity was also
detected in the small intestine (0.010 ± 0.0013 UÆmg
)1
protein). Thus, the pancreas was u sed as the starting
material for the purification of shark DNase I.
The results of purification are summarized in Table 1.
The purification procedure, using three different kinds of
column chromatography, allowed the enzyme to be repro-
ducibly isolated and purified approximately 2000-fold to
electrophoretic homogeneity (Fig. 1). Although anion
exchange chromatography using resins such as DEAE–
Sepharose CL-6B has generally been found useful for the
purification of vertebrate DNases I, including the human
[11], rat [12], rabbit [13], amphibian [19] and reptile [18]
enzymes, shark DNase I was retained on a cation exchange
resin but not on an anion exchange one. As shown below,
H. japonicus DNase I consisted of 262 amino acid residues;
however, it was found to contain more basic amino acids

(%)
Crude extract 42 65 22 1.5 1.0 100
CM-Sepharose CL-6B 4.7 57 65 12 8.0 88
Mono S 5 /50 GL 0.88 48 5.0 55 36 74
Superdex75 0.014 40 3.5 2800 1900 62
Fig. 1. Electrophoretic patterns of purified shark DNase I and the
recombinant enzyme revealed by silver-staining and activity-staining.
The final DNase I preparation recover ed from the gel- filtrat ion step
was concentrated and used for SDS/PAGE analysis. The purified
enzyme (approx. 0.5 lg) from H. japonicus (lane 1) was dissolved
in 10 m
M
Tris/HCl pH 6.8, containing glycerol (10%, v/v), SDS ( 2%,
w/v) and 25 m
M
dithiothreitol, heated at 100 °Cfor5minandsub-
jected to SDS/PAGE using a 12.5% gel, followed by silver-staining.
An expression ve ctor, H . japonicus-chi(human:sig), co ntaining an
H. japonicus DNase I cDNA insert (lane 2) was transfected into
COS-7 cells by the lipofection method. The recombinant DNase I
secreted into the medium was subjected to DNA-casting PAGE, fol-
lowed by activity-staining [25]. The purified enzyme (lane 3) was
analysed in the same m anner. The cathode is at the top. The positions
of the molecular mass markers are indicated on the left.
4430 T. Yasuda et al. (Eur. J. Biochem. 271) Ó FEBS 2004
enzyme remained almost unchanged. Therefore, shark
DNase I is more unstable than the mammalian enzymes.
The catalytic properties such as the effects of pH, ionic
strength and m etal ions on the activity, o f the purified shark
DNase I resembled those of the other vertebrate DNases I.

detected in either the cell l ysate o r the medium of the
transfected cells. When we constructed a chimeric mutant,
H. japonicus-chi(human:sig), in which the signal sequence
region of the shark enzyme was substituted by its human
counterpart, and transfected this into th e COS-7 cells,
unambiguous activity levels were expressed. This activity
was completely abolished by 1 m
M
EGTA. Furthermore,
the enzyme activity expressed in the cells migrated to the
position corresponding to the purified D Nase I on the
DNA-casting PAGE gels [25] (Fig. 1), confirming that
the cloned cDNA did indeed encode the expected
H. japonicus DNas e I.
In order to elucidate any common features unique to
shark DNases I, we also cloned a nd sequenced cDNA
encoding the DNase I of another shark, T. scyllia (acces-
sion number AB126700), and found it to contain 998 bp.
This cDNA was composed of a 48 bp 5¢-UTR, an 855 bp
ORF and a 95 bp 3¢-UTR. Thus, shark DNase I cDNAs
appear to be characterized by a shorter 3¢-UTR (average of
96 bp) than those cloned from most o ther vertebrates
(200 ± 89 bp), including the human [29], rabbit [13], m ouse
[14], rat [30], cow [31], hen [17], pig [15] and snake [18]
enzymes. In this respect, shark DNases I resemble the
amphibian (89 ± 22 bp) [19] and Osteichtye (112 ± 20 bp)
[20,21] enzymes. It could t herefore be postulated that the 3¢-
UTR of vertebrate DNase I cDNA lengthened about
twofold during the evolutionary stage between amphibians
and reptiles.

contributes to the structural stability of the enzyme protein,
in addition to another disulfide bond formed by Cys173
and Cys209 [36,37]. The substitution mutant human-sub
Fig. 2. Heat stability of shark (A) and human (B) DNases I and their
mutant constructs. The medium from COS-7 cells transfected with
(A) H. japonicus-chi(human:sig) (d) and its substitution mu tant
H. japonicus-sub(Ala100Cys/Thr103Cys) (j), and (B) human wild-
type DNase I (d) and it s substitution m utants hum an-sub (Cys101Ala/
Cys104Thr) (s) and hu man-ins (Thr204) ( j), was incubated at 45 °C
for the durations indicated, and residual DNase I activities were
determined by the SRED method. The same amounts ( 1 · 10
)7
unit) of each enzyme were u sed for the inc ubatio n.
Ó FEBS 2004 Molecular evolution of vertebrate DNase I family (Eur. J. Biochem. 271) 4431
(Cys101Ala/Cys104Thr), in which the two Cys residues of
human DNase I were substituted by their counterparts
from H. japonicus, e xhibited lower t hermal stability than the
corresponding wil d-type, whereas double substitution o f
Ala100 and Thr103 by Cys residues in the H. japonicus
DNase I, H. japonicus-sub(Ala100Cys/Thr103Cys), made
the enzyme more thermally stable compared with the wild-
type (Fig. 2 ). Deletion of these two Cys r esidues is also seen
in some species of the Osteichthye class, such as tilapia and
eel [21]. T aken together, t hese findings s uggest that the
formation of t he disulfide bo nd between Cys101 and Cys104
may have been acquired during the evolutionary stage in
Osteichthyes, resulting in the production of a more stable
form of the enzyme. Recently, Chen et al. have reported
that the corresponding disulfide bond is important for the
structural integrity of bovine DNase I [38].

two cysteine r esidues and the insertion of an additional Thr/
Asn residue. W e have previously reported that a single
Leu130Ile substitution in reptilian DNases I may produce
the thermally stable form of the higher vertebrates [18].
Therefore, with regard to the genesis of the DNase I e nzyme
present in higher vertebrates such as humans during the
course of evolution, it could be postulated that the DN ase I
protein has acquired incre asing structural stability through
the introduction of the two Cys residues and deletion of the
additional residue, followed by Leu130Ile substitution.
Effect of the signal sequence regions of vertebrate
DNases I on their expression in transfected cells
Although no a ctivity w as detectable in the medium o r
lysates of cells that had been transfected with expression
vectors containing the entire coding regions of the wild-type
shark DNases I, substitution of the signal sequence regions
of each of the s hark enzymes w ith that d erived from human
DNase I gave rise to expression of activity. In order to
examine the possible effect of the signal sequence region on
the expression o f activity, we constructed a series of chimeric
Fig. 3. Alignment of the amino acid sequences of the t wo shark DNase I molecules wi th those of the human, amphibian, re ptilian and p iscine enzymes.
The amino ac id seq ue nces o f t he shark D N ases I were deduced f rom t heir respective cDN As a nd co mpared with published s eque nces for the hum an
[29], snake [18], frog [19] and eel [21]. The amino acids of each mature prot ein are nu mbe red fro m the N terminus. The dots in dicate residue s that a re
the same as those in H. japonicus, while the horizontal bars indicate deleted amino acid residues.
4432 T. Yasuda et al. (Eur. J. Biochem. 271) Ó FEBS 2004
mutant enzymes, and compared the activity secreted into
the medium from the transfected COS-7 cells (Table 2).
When the signal sequence region of human DNase I was
replaced by the c orresponding r egions of the frog, eel,
T. scyllia and H. japonicus enzymes, the activities detected

of the central core and t he presence of an a-helical stru cture
in the signal s equence r egion of the protein [40–42]. A nalysis
using
DNASIS PRO
software revealed no distinct differences i n
the hydrophobicity profiles of the signal sequence regions
of human, eel, frog, T. scyllia and H. japonicus DNases I.
However, prediction of the secondary structure of the
corresponding part of the enzyme using the
SSTHREAD
2
Program ( />according to the method of Ito et al. [43] revealed that
the a-helical structure contents of the T. scyllia, H. japon-
icus, eel and frog DNases I were significantly lower than
that o f the human enzyme (Table 3). The lower the a-helic al
structure c ontent i n t he signal sequence r egion o f each
DNase I, the lower the expression level each enzyme
exhibits; replacement of the human DNase I signal
sequence by the counterpart of the frog enzyme having
the lowest a-helical structure content had the greatest effect
on reducing the expression levels. It seems reasonable to
assume that the low a-helical structure contents of the signal
sequence regions of the T. scyllia, H. japonicus, eel and frog
DNases I may reduce their ability to function as cotrans-
lational targeting signals compared with the latter, resulting
in the observed d iscrepancy in the efficiency of enz yme
expression by the cells transfected with each of the
expression v ector s. Base d on t he D Nase I cDNA data
available from databases, the average a-helical structure
contents of the signal sequence regions of DNase I proteins

Eel 9.9 ± 0.24 · 10
-4
0.67
T. scyllia 4.5 ± 0.41 · 10
-4
0.30
H. japonicus 1.0 ± 0.24 · 10
-3
0.67
Frog Frog 3.9 ± 0.39 · 10
-4
–
Human 9.9 ± 0.86 · 10
-4
2.5
Eel Eel 1.6 ± 0.40 · 10
-4
Human 2.6 ± 0.23 · 10
-4
1.6
T. scyllia T. scyllia n.d. –
Human 1.0 ± 0.51 · 10
-5
–
H. japonicus H. japonicus n.d. –
Human 3.7 ± 0.82 · 10
-6
–
Table 3. a-Helical structure contents of the signal sequence regions of the vertebrate DNases I used in expression analysis. a-Helical structure contents
were estimated by t he m ethod o f Ito et al. [43]. The portions of the signal sequence regions of each vertebrate DNase I with an a-helical s t ructu re a re

Acknowledgements
This work was supported in part by Grants-in-Aid from the Japan
Society for the Prom otion o f S cience (1520 9023 to T Y, 162 09023 to KK
and 15590575 to YK).
References
1. Moore, S. (1981) Chapter 15. Pancreatic Dnase. In The Enzymes
(Boyer, P.D., ed.), 3 rd edn Vol. 14, pp. 281–296. Academic Pres s,
New York, USA.
2.Nadano,D.,Yasuda,T.&Kishi,K.(1993)Measurementof
deoxyribonuclease I activity in h uman tissues and body fluids by a
single radial enzyme-diffusion method. Clin. Chem. 39, 448–452.
3. Takeshita, H., Mogi, K., Yasuda, T., Nakajima, T., Nakashima,
Y., M ori, S., H oshino, T. & Kishi, K . (2000) Mammalian
deoxyribonucleases I are classified into three types: pancreas,
parotid, and pancreas-parotid (mixed), based on differences in
their tissue concentrations. Biochem. Biophys. Res. Commun. 269,
481–484.
4. Napirei, M., Karsunky, H., Zevnik, B., Stephan, H., Mannherz,
H.G. & Mo
¨
ro
¨
y, T. (2000) Features of systemic lupus erythe-
matosus in DNase1-deficien t mice. Nat. Genet. 25, 177–181.
5.Yasutomo,K.,Horiuchi,T.,Kagami,S.,Tsukamoto,H.,
Hashimura, C., Urushihara, M. & Kuroda, Y. (2001) Mutation of
DNASE1 in people with systemic lupus erythematosus. Nat.
Genet. 28, 313–314.
6. Chitrabamrung, S., Rubin, R.L. & Tan, E.M. (1981) Serum
deoxyribonuclease I a nd clin ical activity in systemic lupus

immunological and proteochemical chara cteriza tion, nucleotide
sequence, expression in tissues, relationships with other mamma-
lian DNases I. and phylogenetic analysis. Biochem. J. 325, 465–
473.
14. Takeshita, H., Yasuda, T., Nakajima, T., Hosomi, O., Nakashi-
ma, Y. & Kishi, K. (1997) Mouse d eoxyribonuclease I (D Nase I):
biochemical a nd immunological characterizations, cD NA s truc-
ture and tissue distribution. Biochem. Mol. Biol. Int. 42, 65–75.
15. Mori, S., Yasuda, T., Takeshita, H., Nakajima, T., Nakazato, E.,
Mogi, K ., Kaneko, Y. & Kishi, K. (2001) Molecular, biochemical
and immunological analyses of porcine pancreatic DNase I.
Biochim. Biophys. Acta 1547, 275–287.
16. Kaneko, Y., Takeshita, H., Mogi, K., Nakajima, T., Yasuda, T.,
Itoi, M., Kuwan o, H. & Kishi, K. (2003) Molecular, biochemical
and immunological analyses of canine pancreatic DNase I.
J. Biochem. 134, 711–718.
17. Nakashima, Y., Yasuda, T., Takeshita, H., Nakajima, T., Hos-
omi, O., Mori, S. & K ishi, K. (1999) Molecular, biochemical a nd
immunological studies o f hen pancreatic deoxyribonu clease I. In t.
J. Biochem. Cell Biol. 31, 1315–1326.
18. Takeshita, H., Yasuda, T., N akajima, T., Mogi, K., Kaneko, Y.,
Iida, R . & Kishi, K. (2003) A single amino acid substitution of
Leu130Ile in snake DNases I contributes to the acquisition of
thermal stability: a clue t o t he molecular e volutionary m echanism
from cold-blooded to warm-blooded vertebrates. Eur. J. Biochem.
270, 307–314.
19. Takeshita, H ., Yasuda, T., Iida, R., N akajima, T., Mori, S., Mogi,
K.,Kaneko,Y.&Kishi,K.(2001)AmphibianDNasesIare
characterized by a C-terminal end with a unique, cysteine-rich
stretch and by the insertion of a serine residue into the Ca

enzyme activity. Immunol. Invest. 27, 145–152.
26. Sambrook, J. & Russell, D.W. (200 1) Analy sis of gene expression
in cultured mammalian cells. Molecular Cloning: a Laboratory
Manual, Vol. 3, 3rd edn, pp. 17.1–17.112. Cold Spring Harbor
Laboratory, Cold Spring Harbor, N ew York.
27. Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K. & Pease, L.R.
(1989) Site-directed mutagenesis by overlap extension using the
polymerase chain reaction. Gene 77, 51–59.
28. Yasuda, T., Takeshita, H., Iida, R., Nakajima, T., Hosomi, O.,
Nakashima, Y., Mori, S. & Kishi, K. ( 1999) Structural require-
ment of a human deoxyribonuclease II for the development of
theactiveenzymeform,revealedbysite-directedmutagenesis.
Biochem. Biophys. Res. Commun. 256, 591–594.
29. Shak, S., Capon, D.J., Hellmiss, R., Marsters, S.A. & Baker,
C.L. (1990) Recombinant human DNase I reduces the viscosity
of cystic fibrosis sputum. Proc. Natl Acad. Sci. USA 87, 9188–
9192.
30. Polzar, B. & Mannherz, H.G. (1990) Nucleotide sequence of a full
length cDNA clone encoding the deoxyribonuclease I from the rat
parotid gland. Nucleic Acids Res. 18, 7151.
31. Chen,C Y.,Lu,S C.&Liao.T H. (1998) Cloning, sequencing
and expression of a cDNA encoding bovine pancreatic deoxy-
ribonuclease I in Escerichia coli:purificationandcharacterization
of the recombinant enzym e. Gene 206, 181–184.
32. Paudel, H.K. & Liao, T H. (1986) Comparison of the three
primary structures of deoxyribonuclease isolated from bovine,
ovine, and porcine pancreas. D erivation of t he am ino a cid
sequence of ovine DNase and revision of the previously published
amino acid sequence of bovine DNase. J. Biol. Chem. 261,
16012–16017.

Luirink, J. (1995) Early e vents in p reprotein recognition in E.coli:
interaction of SRP and trigger factor with nascent polypeptides.
EMBO J. 14, 5494–5505.
42. Rothe, C. & Lehle, L. (1998) Sorting of invertase signal peptide
mutants in yeast dependent and independent o n the s ignal-
recognition particle. Eur. J. Biochem. 252, 16–24.
43. Ito, M., Matsuo, Y. & N ishikawa, K . (1997) Prediction of protein
secondary structure using the 3D-1D compatibility algorithm.
Comput. Appl. Biosc i. 13, 415–423.
Ó FEBS 2004 Molecular evolution of vertebrate DNase I family (Eur. J. Biochem. 271) 4435