Tài liệu Báo cáo khoa học: Purification and sequence identification of anserinase ppt

Purification and sequence identification of anserinase
Shoji Yamada, Yoshito Tanaka and Seiichi Ando
Faculty of Fisheries, Kagoshima University, Japan
Na-Acetylhistidine is found in high concentration
exclusively in the brain, retina, lens, and occasionally
the heart of poikilothermic vertebrates (bony fishes,
amphibians and reptiles) excluding jawless and cartila-
ginous fishes, but is absent from these tissues in
homothermic vertebrates (birds and mammals) [1–3].
However, little is known about its biological roles in
poikilothermic vertebrates. It is synthesized from l-His
and acetyl-CoA by histidine acetyltransferase (EC
2.3.1.33) in the brain and lens [4], and hydrolyzed to
histidine by anserinase (Xaa-methyl-His dipeptidase,
EC 3.4.13.5) in the brain and eye [5,6]. Baslow &
Lenney [5] isolated the enzyme that deacetylates
Na-acetylhistidine from the brain of skipjack tuna
Katsuwonus pelamis, and thus this enzyme was tempor-
arily classified as ‘Na-acetylhistidine deacetylase’ (EC
Keywords
acetylhistidine; anserinase; carnosinase;
cytosolic nonspecific dipeptidase; MEROPS
M20A metallopeptidase
Correspondence
S. Yamada, Faculty of Fisheries, Kagoshima
University, 4-50-20 Shimoarata, Kagoshima
890-0056, Japan
Fax: +81 99 2864015
Tel: +81 99 2864172
E-mail: yamada@fish.kagoshima-u.ac.jp
Enzyme

EC 3.4.13.5; recommended name: Xaa-
methyl-His dipeptidase; other names:
anserinase, aminoacyl-methylhistidine
dipeptidase, acetylhistidine deacetylase,
N-acetylhistidine deacetylase, a-N-acetyl-
L-histidine aminohydrolase, X-methyl-His
dipeptidase
Note
The nucleotide sequences reported in this
paper have been submitted to
DDBJ ⁄ EMBL ⁄ GenBank databank with
accession numbers AB179777 for anserinase
and AB219566 for CNDP-like protein.
(Received 7 August 2005, revised 2
September 2005, accepted 23 September
2005)
doi:10.1111/j.1742-4658.2005.04991.x
Anserinase (Xaa-methyl-His dipeptidase, EC 3.4.13.5) is a dipeptidase that
mainly catalyzes the hydrolysis of Na-acetylhistidine in the brain, retina
and vitreous body of all poikilothermic vertebrates. The gene encoding
anserinase has not been previously identified. We report the molecular
identification of anserinase, purified from brain of Nile tilapia Oreochromis
niloticus. The determination of the N-terminal sequence of the purified
anserinase allowed the design of primers permitting the corresponding
cDNA to be cloned by PCR. The anserinase cDNA has an ORF of 1485
nucleotides and encodes a signal peptide of 18 amino acids and a mature
protein of 476 amino acids with a predicted molecular mass of 53.3 kDa.
Sequence analysis showed that anserinase is a member of the M20A metal-
lopeptidase subfamily in MEROPS peptidase database, to which ‘serum’
carnosinase (EC 3.4.13.20) and cytosolic nonspecific dipeptidase (EC

3.4.13.18, CNDP) belong. A cDNA encoding CNDP-like protein was also
isolated from tilapia brain. Whereas anserinase mRNA was detected only
in brain, retina, kidney and skeletal muscle, CNDP-like protein mRNA
was detected in all tissues examined.
Abbreviations
CNDP, cytosolic nonspecific dipeptidase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.
FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing 6001
3.5.1.34). Subsequently, Lenney et al. [6] reported that
the substrate specificity of Na-acetylhistidine deacety-
lase purified from brain and eye of skipjack tuna
resembled that of anserinase purified from skeletal
muscle of codling Gadus callarias. In 1981, therefore,
Na-acetylhistidine deacetylase was judged to be identi-
cal with anserinase by NC-IUB, and its EC number
was deleted. Anserinase was discovered by Jones [7,8],
who found that anserine (b-alanyl-1-methylhistidine)
was hydrolyzed by this enzyme in skeletal muscle of
G. callarias. Anserinase is activated by bivalent metal
ions and has broad specificity, with ability to
hydrolyze many kinds of substrates such as Na-ace-
tylhistidine, N-acetylmethionine, anserine, carnosine,
homocarnosine (c-aminobutyrylhistidine), alanylhisti-
dine, glycyl-leucine and leucylglycine [6,8,9]. Previ-
ously, we reported that this enzyme, purified from the
brain of rainbow trout Oncorhynchus mykiss to appar-
ent homogeneity, is a homodimeric protein with a
subunit of 55 kDa [9]. It is commonly believed that
anserinase is universally distributed in poikilothermic
animals containing Na-acetylhistidine in their tissues
[5,6,9–11].

Mammalian tissues contain another peptidase
called carnosinase. Carnosinase resembles anserinase in
hydrolytic ability against carnosine, anserine and homo-
carnosine, which are unusual dipeptides containing
non-a-amino acids (i.e. b-alanine and c-aminobutyric
acid). No other enzymes except anserinase and carno-
sinase can hydrolyze these three dipeptides. Human tis-
sues were aggressively investigated for carnosinase
because its deficiency has been associated with neuro-
logical deficits including intermittent seizures and men-
tal retardation [12,13]. Carnosinase exists as two types:
‘tissue’ carnosinase (Xaa-His dipeptidase, EC 3.4.13.3)
and ‘serum’ carnosinase (b-Ala-His dipeptidase, EC
3.4.13.20) [14–16]. As ‘tissue’ carnosinase with broad
specificity is present in every human tissue, it has been
suggested that this enzyme is identical with cytosolic
nonspecific dipeptidase (CNDP, EC 3.4.13.18) [16]. On
the other hand, ‘serum’ carnosinase with a narrow
specificity is a glycoprotein and found in human
serum, as well as in the brain and spinal fluid [17].
Recently, Teufel et al. [18] reported the molecular iden-
tification of the two types of carnosinase in human,
and the sequence analyses revealed that both CNDP
and ‘serum’ carnosinase belong to the M20A metallo-
peptidase subfamily in MEROPS database [19].
Unlike carnosinase, the gene encoding anserinase
has not been previously identified. We report here for
the first time the molecular identification of anserinase.
Results and Discussion
Purification and characterization of anserinase

from Nile tilapia brain
The procedure for the purification of the enzyme from
Nile tilapia is summarized in Table 1. The brains were
collected from  1000 specimens of Nile tilapia (com-
mercial size). Crude extracts from the fish brains were
first subjected to ammonium sulfate fractionation.
Anserinase was recovered from the 50–60% saturated
ammonium sulfate fraction. The active fraction was
then subjected to hydrophobic interaction chromato-
graphy on octyl-Sepharose CL-4B. When the fractions
were screened for hydrolysis against Na-acetylhistidine,
the activity was recovered as an unbound fraction
(data not shown). The unbound fraction containing
anserinase was then subjected to gel filtration using
Superdex 200 HR (Fig. 1A). The molecular mass of
anserinase as determined by gel filtration was
120 kDa. The active fractions were pooled, and subjec-
ted to anion-exchange chromatography using Resource
Q. The bound enzyme was eluted from the column as
a single peak of enzyme activity when the salt concen-
tration was  0.30 m (Fig. 1B). The active fractions
were concentrated, and applied to a preparative native
PAGE (Fig. 2). As shown in Fig. 2A, several protein
bands were detected on the native PAGE. When the
samples extracted from the gel slices were assayed for
hydrolysis against Na-acetylhistidine, the activity was
Table 1. Purification of anserinase from brain of Nile tilapia. One enzyme unit is defined as that activity of enzyme that catalyzes the hydro-
lysis of 1 lmol Na-acetylhistidine in 1 h under the standard conditions.
Step Fraction Total protein (mg) Total activity (U) Specific activity (UÆmg
)1

) Purification (fold) Yield (%)
1 Crude extract 10479 6949 0.7 1 100
2 Ammonium sulfate precipitation 992 2811 2.8 4 40
3 Octyl-Sepharose CL-4B 125 1936 15 21 28
4 Superdex 200 HR 32 1494 47 67 21
5 Resource Q 7.5 651 87 124 9
6 Preparative native PAGE 0.52 289 556 794 4
7 Preparative SDS ⁄ PAGE 0.11 – – – –
Purification and sequence identification of anserinase S. Yamada et al.
6002 FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing
mainly recovered from gel slice numbers 13 and 14
(Fig. 2B). Gel slices 11–15 were subjected to SDS ⁄
PAGE under reducing conditions. Protein bands were
visualized with silver staining (Fig. 2C). The visualized
intensity of the 55-kDa protein band, indicated by an
arrow in Fig. 2C, correlated with enzyme activity levels
shown in Fig. 2B. Moreover, we previously reported
that anserinase purified from brain of rainbow trout
consists of a subunit of 55 kDa [9]. Therefore, the
55-kDa protein was assumed to be a single subunit
of Nile tilapia anserinase. As the molecular mass of
anserinase determined by gel filtration is 120 kDa in
the present study, Nile tilapia anserinase, like the trout
enzyme [9], is apparently a homodimeric protein. The
sample solutions from gel slices 13 and 14 containing
high enzyme activity were pooled and concentrated.
The separation procedures by preparative native
PAGE (step 6) resulted in 794-fold purification and
4% of the enzyme activity. However, the enzyme was
not homogeneous, as there were several protein bands

besides anserinase visualized from gel slices 13 and 14,
as shown in Fig. 2A. Therefore, the final purification
was conducted using preparative SDS ⁄ PAGE (step 7).
The purified enzyme obtained from step 7 showed a
single protein band (55 kDa) on SDS ⁄ PAGE (Fig. 3).
The final recovery of the anserinase protein was
110 lg from 345 g fish brain. The sequence of the
N-terminal 20 amino acids for the purified Nile tilapia
anserinase, determined by automated Edman degrada-
tion, was FXYMDLAQYVDSXQDEYVEM. In the
N-terminal sequence, two amino acids expressed as ‘X’
at the 2nd and 13th residue from the N-terminus failed
to be identified for unknown reasons.
In Table 2 the substrate specificities of the Nile til-
apia anserinase obtained by step 6 were compared with
the previous data obtained from trout anserinase [9],
which was purified from the trout brain to apparent
homogeneity. The Nile tilapia and rainbow trout
enzymes had similar broad specificities. Both enzymes
showed strong hydrolytic activity against Na-chloro-
acetyl-l-Leu and Gly-Gly. Anserine, carnosine and
homocarnosine were also hydrolyzed. The Nile tilapia
enzyme, however, hydrolyzed l-Ala-l-His, l-Leu-Gly
and l-Pro-Gly at a much higher rate than the trout
enzyme. Moreover, the Nile tilapia enzyme hydrolyzed
both l-His-Gly and l-Ala-l-Pro, which were hardly
cleaved at all by the trout enzyme. From these results,
it is likely that the specificity of Nile tilapia anserinase
is broader than that of rainbow trout anserinase.
Molecular cloning of anserinase and CNDP-like

protein
Database search
Initially we searched for a candidate for anserinase in
the GenBank database by blastp using the N-terminal
sequence of Nile tilapia anserinase. An ‘unnamed pro-
tein’ product (DDBJ ⁄ EMBL ⁄ GenBank accession num-
ber CAF95589) of the spotted river puffer Tetraodon
nigroviridis was extracted from the database (Fig. 4).
The N-terminal amino-acid sequence of Nile tilapia
anserinase displayed 13 of 18 residues (excluding two
residues of X) identical with the deduced amino-acid
sequence 19–38 from the N-terminus of the Tetraodon
Fig. 1. Gel filtration and anion-exchange chromatography of Nile til-
apia anserinase. (A) Gel filtration on a Superdex 200 HR column
(step 4). The concentrated sample from octyl-Sepharose CL-4B
chromatography was applied to a Superdex 200 HR 10 ⁄ 30 column
equilibrated with 50 m
M sodium phosphate buffer, pH 7.0, contain-
ing 150 m
M NaCl. Fractions of 200 lL each were collected at a
flow rate of 0.4 mLÆmin
)1
. Standard proteins (molecular mass in
parentheses) were aldolase (158 kDa), BSA (67 kDa), ovalbumin
(43 kDa), chymotrypsinogen A (25 kDa) and ribonuclease A
(13.7 kDa). (B) Chromatography on a Resource Q column (step 5).
The pooled active fraction from Superdex 200 HR chromatography
was applied to a Resource Q column (1 mL size) equilibrated with
20 m
M Tris ⁄ HCl buffer, pH 7.8. The column was washed with

10 mL of the equilibration buffer. A linear NaCl gradient (0–0.5
M;
20 mL of 20 m
M Tris ⁄ HCl buffer, pH 7.8) was then applied. Frac-
tions of 1 mL each were collected at a flow rate of 0.4 mLÆmin
)1
.
S. Yamada et al. Purification and sequence identification of anserinase
FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing 6003
‘unnamed protein’. The organization principle of the
MEROPS peptidase database is a hierarchical classifi-
cation in which homologous sets of the proteins of
interest are grouped into families, and the homologous
families are grouped into clans [19]. Therefore, the
blastp program of the MEROPS database was used
to search homologous peptidase genes to the Tetrao-
don ‘unnamed protein’. Members of CNDP (MEROPS
ID M20.005) and ‘serum’ carnosinase (MEROPS ID
M20.006) in the M20A metallopeptidase family ⁄ sub-
family of the MH clan were extracted from the data-
base. Multiple sequence alignments of the extracted
genes of vertebrates were performed using the clustal
w program to reveal highly conserved amino-acid
sequences for designing degenerate primers for PCR
amplification (data not shown).
Cloning of anserinase cDNA
PCR was performed using a set of primers (A and B)
and the first-strand cDNA for 3¢ rapid amplification
of cDNA ends (RACE), prepared from total RNA of
Nile tilapia brain, as the template. The first-round

PCR product was then used as a template for nested
PCR amplification using a set of nested primers (A
and C). As a result, the 281-bp PCR product was spe-
cifically amplified. A full-length cDNA sequence of
Nile tilapia anserinase was finally obtained by 3¢ and
5¢ RACE PCR (Fig. 5). The cDNA contained 1840-bp
nucleotides including 56 bp of 5¢ UTR, 1482 bp of
ORF, and 299 bp of 3¢ UTR. The coding region of the
sequence was translated into 494 amino acids, which
included a typical signal peptide of 18 amino acids and
two potential N-glycosylation sites at the 104th and
134th residue (Asn) from the N-terminus. The predic-
ted N-terminal amino-acid sequence of the ORF exclu-
ding signal peptide completely matched the sequence
determined by automated Edman degradation. cDNA
sequence analysis predicted that two unknown amino
acids at the 2nd and 13th residue from the N-terminus
of the purified protein were both histidines. The cal-
culated molecular mass and isoelectric point of the
mature protein, excluding the signal peptide, were
53 311 Da and pH 5.3, respectively.
A
B
C
Fig. 2. Preparative native PAGE (step 6) of
Nile tilapia anserinase. (A) Preparative native
PAGE (7.5% running gel and 4.5% stacking
gel) was performed as described in Experi-
mental procedures. Proteins were stained
with Coomassie Brilliant Blue R-250. (B) Un-

stained gels were cut into 30 gel slices
(numbered 1–30), and each fraction was
assayed for hydrolytic activity against Na-
acetylhistidine. (C) The samples of gel slices
11–15 were subjected to SDS ⁄ PAGE (7.5%
running gel and 4.5% stacking gel) under
reducing conditions. Protein bands were
visualized with silver staining. Standard pro-
teins (STD) were phosphorylase b (94 kDa),
BSA (67 kDa), ovalbumin (43 kDa), and car-
bonic anhydrase (30 kDa). The arrow indi-
cates the position of the 55-kDa protein that
proved to be anserinase.
Purification and sequence identification of anserinase S. Yamada et al.
6004 FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing
Cloning of the CNDP-like protein cDNA
We also cloned a full-length cDNA encoding Nile til-
apia CNDP-like protein using a partial nucleotide
sequence of Mozambique tilapia (Oreochromis mossam-
bicus) CNDP-like protein for primer design. PCR was
performed using a set of primers (D and E) and the
first-strand cDNA for 3¢ RACE as the template. As a
result, the 527-bp PCR product was specifically ampli-
fied. To obtain the 3¢ and 5¢ terminal segments of the
cDNA, 3¢ and 5¢ RACE were then performed. The
ORF of CNDP-like protein coded for a cytoplasmic
protein (no signal peptide) of 474 amino acids with a
calculated molecular mass of 52 807 Da and isoelectric
point of 5.6 (data not shown). The deduced amino-acid
sequence of CNDP-like protein showed 52% identity

with and 66% similarity to that of anserinase from
Nile tilapia (Fig. 6).
Tissue distribution of the mRNAs for anserinase
and CNDP-like protein in Nile tilapia
The possibility of genomic contamination was elimin-
ated by the observation of amplifications spanning
the exon–intron boundaries, which were based on the
information of the two scaffold files (M001527 for
Fig. 3. SDS ⁄ PAGE of purified anserinase (step 7). The purified
enzyme obtained from preparative SDS ⁄ PAGE was subjected to
SDS ⁄ PAGE (12.5% running gel and 4.5% stacking gel) under redu-
cing conditions. Protein bands were visualized with Coomassie Bril-
liant Blue R-250. Standard proteins (STD) were phosphorylase b
(94 kDa), BSA (67 kDa), ovalbumin (43 kDa), carbonic anhydrase
(30 kDa), soybean trypsin inhibitor (20.1 kDa) and a-lactalbumin
(14.4 kDa).
Table 2. Substrate specificity of Nile tilapia anserinase.
Substrate
Relative rates of hydrolysis
Nile tilapia
a
Rainbow trout
b
Na-Acetyl-L-His 100 100
Na-Acetyl-
L-Asp 0 3
Na-Acetyl-
L-Glu 0 23
Na-Acetyl-
L-Met 116 254

Na-Acetyl-
L-Cys 0 0
Na-Acetyl-
L-Trp 3 30
Na-Acetyl-
L-Leu 58 110
Na-Chloroacetyl-
L-Leu 755 1235
L-Leu-b-naphthylamide 0 0
Anserine 39 16
Carnosine 74 83
Homocarnosine 155 67
L-Ala-L-His 447 105
Gly-
L-His 327 166
L-His-Gly 58 6
Gly-
L-Leu 443 823
Gly-
D-Leu 0 0
L-Leu-Gly 360 101
L-His-L-Leu 26 0
Gly-Gly 543 385
L-Cys-Gly 0 0
L-Pro-Gly 301 76
L-Ala-L-Pro 46 0
Gly-Gly-
L-Leu 0 0
Gly-
L-Leu-L-Tyr 0 0

L-Leu-Gly-Gly 0 0
a
The data from this study. The semipurified Nile tilapia enzyme
obtained from step 6 (Table 1) was incubated at 30 °C for 1 h with
1m
M substrate in 150 mM N-ethylmorpholine ⁄ HCl buffer, pH 6.5,
containing 1 m
M CoSO
4
.
b
The data from the previous study [9]. The
anserinase of rainbow trout was purified to apparent homogeneity,
and incubated under the same conditions as in the present study.
Fig. 4. Alignment of amino-acid sequences of Nile tilapia anserinase
N-terminus and ‘unnamed protein’ product (DDBJ ⁄ EMBL ⁄ GenBank
accession number CAF95589) of spotted river puffer Tetraodon
nigroviridis. Vertical lines indicate amino-acid identities, whereas
colons indicate conservative substitutions. In the N-terminal
sequence of Nile tilapia anserinase, two amino acids expressed as
‘X’ at the 2nd and 13th residue from the N-terminus were not able
to be analyzed for unknown reasons.
S. Yamada et al. Purification and sequence identification of anserinase
FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing 6005
anserinase-like protein and M001163 for CNDP-like
protein) extracted from the Fugu genome database.
The RT-PCR products of the expected size for CNDP-
like protein were observed in all tissues examined
(Fig. 7). These distributions of CNDP-like protein in
the fish are exactly consistent with those of human

CNDP. In mouse, Otani et al. [20] recently reported
that Western blotting analysis using the antibody
against the recombinant carnosine-hydrolyzing protein,
which is identical with CNDP-like protein, revealed
the presence of the protein in kidney, liver, brain and
spleen, and weakly in heart muscle and skeletal muscle.
Although no enzymological information for CNDP in
fish is at present available, it is likely that fish CNDP-
like protein plays the same role as mammalian CNDP.
On the other hand, anserinase mRNA was expressed
strongly in brain, retina, skeletal muscle and kidney,
and slightly in spleen (Fig. 7). We could not detect the
RT-PCR products of the expected size for anserinase
mRNA in any other tissues. According to our previous
works [11,21], the enzymatic activity of anserinase was
detected strongly in kidney, brain, liver and ocular
fluid, and weakly in skeletal muscle and spleen of Nile
tilapia. The mRNA expression in the brain and kidney
obtained in this study are therefore consistent with the
distribution of the enzyme activity. It seems likely that
the enzyme activity in the ocular fluid originates from
anserinase secretion from the retina, in which anseri-
nase mRNA was strongly expressed. Interestingly,
anserinase mRNA was not expressed in the liver,
which contained the enzyme activity. Therefore, the
question arises which tissue is the origin of liver anseri-
nase. Whereas human ‘serum’ carnosinase is expressed
only in central nervous system, rat and mouse ortho-
logues are found exclusively in the kidney and are not
expressed in the central nervous system [18]. Margolis

et al. [22,23] revealed that tissues with carnosinase
activity can be divided into two groups: kidney, uterus
and olfactory mucosa represent one group, and central
nervous system, muscle, spleen, etc. represent the sec-
ond. Therefore, it can be suggested that mouse kidney
carnosinase is translated from a gene orthologous to
Fig. 5. Nucleotide sequence and predicted amino-acid sequence of cDNA encoding Nile tilapia anserinase. The putative signal sequence is
shown in italics, and the N-terminal amino acids as determined by protein sequencing are underlined. Potential N-glycosylation sites are
surrounded by rectangles. The active site residues are surrounded by ovals, and the metal binding sites are also highlighted in black. The
terminal stop codon is marked with an asterisk.
Purification and sequence identification of anserinase S. Yamada et al.
6006 FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing
human ‘serum’ carnosinase. Interestingly, the mouse
protein predicted from ‘serum’ carnosinase cDNA,
unlike the human ‘serum’ carnosinase, does not have a
typical N-terminal signal peptide. The gene expression
and protein distribution of either anserinase or ‘serum’
carnosinase are therefore very complicated in verteb-
rate animals.
Molecular phylogenetic analysis
We aligned the deduced amino-acid sequences of ver-
tebrate M20A genes including the gene of Nile tilapia
anserinase and constructed the unrooted phylogenetic
tree shown in Fig. 8. Gene sequences from ascidian
and vertebrate animals cluster within three distinct
groups, CNDP-like, ‘serum’ carnosinase-like, and anse-
rinase-like types. Only one gene was extracted from
the databases for the ascidian Ciona intestinalis.It
is likely that the ascidian gene is grouped as a
CNDP-like type. The primary structures of all

Fig. 6. Amino-acid alignment of Nile tilapia anserinase and CNDP-like genes. The putative signal sequence is underlined. Identical amino
acids are indicated by an asterisk, and chemically similar amino acids are indicated by dots. Gaps inserted into the sequences are indicated
by dashed lines. The active-site and metal-binding-site residues are highlighted in gray and black, respectively. The deduced amino-acid
sequence of the ORF-encoded anserinase was aligned with the encoded CNDP-like protein using
CLUSTAL W, showing 52% sequence identity
and 66% similarity.
Fig. 7. Tissue distribution of the mRNAs for anserinase and CNDP-
like protein. Total RNA was prepared from Nile tilapia tissues, and
RT-PCR was performed using specific primers. The expected sizes
of the amplified bands of anserinase, CNDP-like protein and
GAPDH were 530, 395 and 517 bp, respectively.
S. Yamada et al. Purification and sequence identification of anserinase
FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing 6007
CNDP-like proteins are relatively conserved; however,
the function of these proteins is unknown except in
humans [18]. In both African clawed frog Xenopus
laevis and Atlantic salmon Salmo salar, three genes,
which are separately grouped into CNDP-like, ‘serum’
carnosinase-like, and anserinase-like types, were extrac-
ted from the databases. Darmin, the function of which
is unknown, is a secreted protein expressed during
endoderm development in African clawed frog [24].
From our phylogenetic analysis, Darmin protein is
grouped as an anserinase-like type. However, as the
phylogenetic divergence of Xenopus Darmin protein is
a relatively long way from fish anserinase, as shown
in Fig. 8, the enzymatic properties of Darmin protein
need to be investigated and compared with those
of fish anserinase. Another hypothetical MGC68563
protein of African clawed frog is closely related to a

human ‘serum’ carnosinase. A homologous gene to
‘serum’ carnosinase in Atlantic salmon was also
obtained by assembling three EST clones (TC33189,
CX353277 and TC29012) extracted from the databases.
We therefore suggest that before the vertebrate–ascidian
divergence, an ancestral CNDP gene was first duplica-
ted to form the original CNDP and the copied
CNDP genes. The copied CNDP gene was further dupli-
cated to form ‘serum’ carnosinase and anserinase genes.
This second gene duplication event occurred before the
divergence of ray-finned fish and tetrapod lineages.
In conclusion, we report the molecular identification
of anserinase, and demonstrate that the enzyme is a
member of the M20A metallopeptidase subfamily, as
well as ‘serum’ carnosinase and CNDP. The anserinase
Fig. 8. Phylogenetic tree of anserinase and M20A genes of vertebrate animals. The tree was constructed by neighbor-joining distance analy-
sis. Bootstrap values of 1000 resampling are indicated for all nodes on the tree. The scale bar corresponds to the estimated evolutionary dis-
tance units. In Fugu Takifugu rubripes, the homologous genes to anserinase-like and CNDP-like proteins were located on scaffolds M001527
and M001163, respectively. The accession numbers of the homologues extracted from the DDBJ ⁄ EMBL ⁄ GenBank or the TIGR (http://
www.tigr.org/tdb/tgi/) databases are as follows: human Homo sapiens, ‘serum’ carnosinase (NM_032649) and CNDP (BC003176); mouse
Mus musculus, ‘serum’ carnosinase-like (NM_177450) and CNDP-like (NM_023149); chicken Gallus gallus, ‘serum’ carnosinase-like
(BX931960) and CNDP-like (TC188297); African clawed frog Xenopus laevis, MGC68563 protein (BC060450), Darmin protein (AY166869) and
CNDP-like (BC056069); zebrafish Danio rerio, CNDP-like (AY391414); medaka Oryzias latipes, CNDP-like (TC30957); salmon Salmo salar,
‘serum’ carnosinase-like (assembled using TC33189, CX353277 and TC29012), anserinase-like (assembled using CK873786 and TC31285)
and CNDP-like (assembled using CK884742, CX352802 and TC22931); ascidian Ciona intestinalis, CNDP-like (TC64855).
Purification and sequence identification of anserinase S. Yamada et al.
6008 FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing
mRNA was expressed strongly in brain, retina, skeletal
muscle and kidney of Nile tilapia, whereas the CNDP
mRNA was expressed in all tissues. It is also expected

that a set of three genes, CNDP-like, anserinase-like
and ‘serum’ carnosinase-like genes, exists in tetrapods
(African clawed frog) and fish (Atlantic salmon). Fur-
ther studies are therefore required to extensively
investigate the existence of anserinase-like and ‘serum’
carnosinase-like genes in vertebrates.
Experimental procedures
Enzyme assay
As Na-acetylhistidine is a major physiological substrate for
anserinase in brain and eye of fish, we used it instead of
anserine as a substrate for the anserinase assay throughout
this study. Enzyme activity was assayed as follows: sample
containing enzyme was incubated at 30 °C for 1 h with
1mm Na-acetylhistidine in 150 mm N-ethylmorpholine ⁄ HCl
buffer, pH 6.5, containing 1 mm CoSO
4
[9]. The reaction
was terminated by the addition of HClO
4
at a final concen-
tration of 5% (w ⁄ v). The sample was then centrifuged for
15 min at 8000 g to precipitate the protein. Released histi-
dine in the supernatant was quantified by HPLC using the
o-phthalaldehyde post-column labeling method [21].
Protein determination
Protein concentration was calculated as the sum of amino-
acid contents after acid hydrolysis (6 m HCl, 24 h). Amino
acid content was determined by HPLC as above.
Analytical SDS/PAGE
Electrophoresis in the presence of SDS and 2-mercaptoeth-

anol was performed by the method of Laemmli [25], with a
7.5% or a 12.5% polyacrylamide running gel and a 4.5%
polyacrylamide stacking gel. Proteins were stained with
0.25% Coomassie Brilliant Blue R-250 in 50% methanol
containing 10% acetic acid or Silver Stain II Kit (Wako
Pure Chemical Industries, Osaka, Japan).
Purification of brain anserinase
All operations were conducted at 0–4 °C unless otherwise
mentioned. Fresh brains (345 g) of Nile tilapia O. niloticus
were stored at )20 ° C.
Step 1: extraction
The frozen brains were homogenized with a Polytron homo-
genizer in 10 vol. 10 mm sodium phosphate buffer, pH 7.8.
The crude homogenate was centrifuged at 20 000 g for 1 h.
Step 2: ammonium sulfate precipitation
The supernatant was brought to 50% saturation with
solid ammonium sulfate and left overnight. The precipi-
tate was removed by centrifugation (20 000 g, 1 h) and
discarded. The supernatant was precipitated by increasing
ammonium sulfate to 60% saturation and left for 1 h.
The precipitate was collected by centrifugation (20 000 g,
1 h), dissolved in 67 mL 10 mm N-ethylmorpholine ⁄ HCl
buffer, pH 7.2, containing 0.1 mm CoSO
4
and brought to
30% saturation with solid ammonium sulfate. Insoluble
material was removed by centrifugation (20 000 g,1h)
and discarded.
Step 3: octyl-Sepharose CL-4B chromatography
The supernatant was applied to a column (2.6 · 40 cm) of

octyl-Sepharose CL-4B (Amersham Pharmacia Biotech)
previously equilibrated with 10 mm N-ethylmorpholine ⁄ HCl
buffer, pH 7.2, containing 30% saturated ammonium sul-
fate and 0.1 mm CoSO
4
at 7 °C. The column was washed
with 500 mL of the equilibration buffer at a flow rate of
1.5 mLÆmin
)1
, and a linear ammonium sulfate gradient
(30–0% saturation; 2 L) was applied. The effluent was
fractionated into 15-mL portions. The active fractions were
pooled and concentrated to 1.2 mL by ultrafiltration
through a PM-10 membrane (Amicon, Inc.).
Step 4: Superdex 200 HR gel filtration
The sample (200 lL) was injected at room temperature
into a Superdex 200 HR 10 ⁄ 30 column (10 · 300 mm;
Amersham Pharmacia Biotech) equilibrated with 50 mm
sodium phosphate buffer, pH 7.0, containing 150 mm
NaCl at a flow rate of 0.4 mLÆmin
)1
. Fractions of
200 lL each were collected. The active fractions were
combined and concentrated to 800 lL using a Centricon
10 (Amicon, Inc.). This separation step was separately
performed six times (200 lL · 6).
Step 5: Resource Q chromatography
The sample was applied at room temperature to a
Resource Q column (1 mL; Amersham Pharmacia
Biotech) equilibrated with 20 mm Tris ⁄ HCl buffer,

pH 7.8, at a flow rate of 1.0 mLÆmin
)1
. The column was
washed with 10 mL of the equilibration buffer for
10 min. A linear NaCl gradient (0–0.5 m; 20mL 20mm
Tris ⁄ HCl buffer, pH 7.8) was applied, and 1-mL fractions
were collected. The active fractions containing the enzyme
were concentrated, and the buffer was replaced with
10 mm N-ethylmorpholine ⁄ HCl buffer, pH 7.2, containing
30% glycerol, using a centrifugal concentrator (Centricon-
10).
S. Yamada et al. Purification and sequence identification of anserinase
FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing 6009
Step 6: preparative native PAGE
Preparative native PAGE (7.5% running gel and 4.5%
stacking gel) was conducted according to a modification of
the method of Lenney et al. [26]. The sample was applied
to a 11 · 14 · 0.2-cm gel (five slabs). After electrophoresis
for 2.5 h at 30 mA, the gel was cut into 3-mm slices. Each
slice was crushed in 5 mL 10 mm N-ethylmorpholine ⁄ HCl
buffer, pH 7.2, containing 0.1 mm CoSO
4
using a Potter-
Elvehjem homogenizer. After centrifugation at 20 000 g for
30 min, the supernatant was concentrated, and the buffer
was completely replaced with 10 mm N-ethylmorpho-
line ⁄ HCl buffer, pH 7.2, using a Centricon-10. An aliquot
of the sample obtained from each gel slice was assayed for
enzyme activity. The active fractions of gel slices were con-
centrated to 400 lL using a Centricon-10.

Step 7: preparative SDS ⁄ PAGE
The concentrate was applied over the width of a gel
slab (11 · 14 · 0.1 cm, three slabs) and subjected to
SDS ⁄ PAGE (7.5% running gel and 4.5% stacking gel) as
described by Laemmli [25]. After electrophoresis for 2.5 h
at 30 mA, 1-cm vertical strips were cut from the right
and left sides of the slab using a cheese knife with a zig-
zag shaped blade; these were immediately stained with
Quick-CBB (Wako Pure Chemical Industries). The stained
gel strips were replaced to each original position on the
slab joining along the zigzag edge. The horizontal strip
containing the anserinase band was excised from the
unstained gel slab. Elution of the protein from the gel
strips was performed electrophoretically using Electro-
Eluter model 422 (Bio-Rad Laboratories), according to
the manufacturer’s instructions. The sample solution was
concentrated, and the buffer was completely replaced with
distilled water, using a Centricon-10 for N-terminal
sequence analysis.
N-Terminal sequence analysis and BLAST search
Edman degradation was performed on an automated pro-
tein sequencer (model 491; Applied Biosystems). Protein
Sequence Databases were searched for homologies with
N-terminal sequence of anserinase using the world wide
web-based blastp search engine of GenBank (http://
www.ncbi.nlm.nih.gov/BLAST/). A further blastp search
was conducted by an engine of the MEROPS database
() [19] using ‘unnamed protein’
sequence (DDBJ ⁄ EMBL ⁄ GenBank accession number
CAF95589) of spotted river puffer Tetraodon nigroviridis,

which was extracted from blastp for N-terminal sequence
of anserinase. Multiple sequence alignments were performed
using the clustal w program ( to
find highly conserved amino-acid sequences.
Molecular cloning of Nile tilapia anserinase and
CNDP-like protein cDNAs
Isolation of a partial cDNA encoding anserinase
Brain (193 mg) was dissected from Nile tilapia weighing
 100 g and immediately stored in liquid nitrogen. Total
RNA was isolated from the frozen brain using 2 mL TRI-
zol Reagent (Invitrogen Corp., Carlsbad, CA, USA) follow-
ing the manufacturer’s instructions. The first-strand cDNA
for 3¢ RACE was synthesized from 1 lg total RNA using
200 U SuperScript II reverse transcriptase, the 35-mer oligo
(dT)-adaptor (10 pmol) 5¢-GGCCACGCGTCGACTAG
TACTTTTTTTTTTTTTTT-3¢, and reverse transcriptase
buffer (20 lL) of the first-strand cDNA synthesis kit (Invi-
trogen Corp.). The synthesis reaction was performed at
42 °C for 50 min. The forward primer (A) was designed
for PCR; 21-mer degenerate oligonucleotide 5¢-CAR
GAYGARTAYGTNGARATG-3¢ corresponding to N-ter-
minal amino-acid sequences 14–20 (QDEYVEM) of Nile
tilapia anserinase. Multiple sequence alignment of the genes
belonging to CNDP (MEROPS ID M20.005) and ‘serum’
carnosinase (MEROPS ID M20.006), which were extracted
from the MEROPS blastp search using the sequence of
Tetraodon ‘unnamed protein’, revealed two highly con-
served regions suitable for designing degenerate oligonucleo-
tides for amplification of anserinase gene fragments
(data not shown). Therefore, two reverse primers were

designed for PCR; the 23-mer oligonucleotide 5¢-GAG
CCNGWYTCYTCCATBCCYTC-3¢ corresponding to one
consensus sequence (EGMEES ⁄ TGS) as the outer pri-
mer (B), and the 23-mer oligonucleotide 5¢-TCCAG
GYTDGCNGGCTGVACRTC-3¢ corresponding to another
consensus sequence (DVQPAN ⁄ SLD ⁄ E) as the inner pri-
mer (C). The first-round PCR was performed using a set of
the primers (A and B), and the second-round nested PCR
was primed with first-round PCR product and as a tem-
plate a set of the primers (A and C). PCR amplification
was carried out in a total volume of 50 lL containing
0.75 lL of a template, 150 pmol of a forward primer (A),
150 pmol of a reverse primer (B or C), 1 · G-Taq buffer,
10 nmol each of dATP, dGTP, dCTP and dTTP, and
0.5 U G-Taq DNA Polymerase (Cosmo Genetech Co.,
Seoul, Korea). For PCR the following conditions were
used: initial denaturation at 95 °C for 2 min, followed by
40 cycles of denaturation at 95 °C for 20 s, annealing at
50 °C for 30 s, and extension at 72 °C for 1 min, final
extension step at 72 ° C for 7 min.
Isolation of a partial cDNA encoding CNDP-like
protein
A partial sequence of CNDP-like protein (DDBJ ⁄ EMBL ⁄
GenBank accession number AY260749) of Mozambique
Purification and sequence identification of anserinase S. Yamada et al.
6010 FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing
tilapia Oreochromis mossambicus, which is classified in the
same genus as Nile tilapia, was extracted from the MER-
OPS database. For PCR amplification of a partial cDNA
for CNDP-like protein of Nile tilapia, the following primers

were synthesized; the 22-mer forward oligonucleotide (D)
5¢-CTGTAAAGATGGTGGAGTTGGC-3¢ and the 22-mer
reverse oligonucleotide (E) 5¢-GCAGCTTGACTCCCTGA
ATGTA-3¢ corresponding to cDNA sequences 2–23 and
507–528, respectively, of Mozambique tilapia CNDP-like
protein.
3¢ and 5¢ RACE
To obtain the 3¢ end, the first-strand cDNA for 3¢ RACE
was used as a template. The 3¢ end of the cDNA was
obtained by PCR with a gene-specific primer as a forward
primer and the 20-mer oligonucleotide 5¢-GGCCACGCG
TCGACTAGTAC-3¢ as a reverse primer using G-Taq
DNA Polymerase. To obtain the 5¢ end, ligation-mediated
5¢ RACE was accomplished using SMART RACE cDNA
Amplification Kit (BD Bioscience Clontech, Palo Alto, CA,
USA). The first-strand cDNA for 5¢ RACE was synthesized
in a total volume of 10 lL containing 0.68 lg total RNA,
200 U SuperScript II reverse transcriptase, 5¢ RACE cDNA
synthesis primer (10 pmol), and SMART II oligonucleotide
(10 pmol). The synthesis reaction was performed at 42 °C
for 90 min, and the 30-mer oligonucleotide 5¢-CCCGCGT
ACTCTGCGTTGTTACCACTGCTT-3¢ was finally ligated
to the 3¢ end of the first-strand cDNA. Aliquots (0.25 l L)
of the reaction mixture (first-strand cDNA for 5¢ RACE) as
a template were then subjected to PCR using the Advant-
age 2 PCR kit (BD Bioscience Clontech, San Jose, CA,
USA) and a set of the forward Universal Primer Mix and
a reverse gene-specific primer. The PCR profile used was
94 °C for 1 min, and then 30 cycles with denaturation at
94 °C for 5 s, annealing at 68 °C for 10 s and extension at

72 °C for 2 min. This was followed by a final extension step
at 72 °C for 7 min.
DNA sequencing
PCR products were subcloned into pGEM-T easy vector
(Promega, Madison, WI, USA). cDNA clones obtained
were sequenced by the dideoxy-chain termination method
using BigDye Terminator v3.1 Cycle Sequencing kit and an
ABI 3100 DNA sequencer (Applied Biosystems).
Computer assisted analysis
The Fugu (puffer) Takifugu rubripes genome database
( was screened for
potential homologues to Nile tilapia anserinase or CNDP-
like protein by tblastn search. The genscan program
( was then used to
predict Fugu anserinase-like protein and CNDP-like protein
genes in scaffold files extracted from the genome database.
The homologous genes to anserinase-like and CNDP-like
proteins were located on scaffolds M001527 and M001163,
respectively. The accession numbers of the homologues
extracted from the DDBJ ⁄ EMBL ⁄ GenBank or the TIGR
( databases are as follows:
human Homo sapiens, ‘serum’ carnosinase (NM_032649)
and CNDP (BC003176); mouse Mus musculus, ‘serum’ car-
nosinase-like (NM_177450) and CNDP-like (NM_023149);
chicken Gallus gallus, ‘serum’ carnosinase-like (BX931960)
and CNDP-like (TC188297); African clawed frog Xenopus
laevis, MGC68563 protein (BC060450), Darmin protein
(AY166869) and CNDP-like (BC056069); zebrafish Danio
rerio, CNDP-like (AY391414); medaka Oryzias latipes,
CNDP-like (TC30957); salmon Salmo salar, ‘serum’ carno-

sinase-like (assembled using TC33189, CX353277 and
TC29012), anserinase-like (assembled using CK873786 and
TC31285) and CNDP-like (assembled using CK884742,
CX352802 and TC22931); ascidian Ciona intestinalis,
CNDP-like (TC64855). The signal sequence was predicted
by using the prediction tool SIGNALP 3.0 Server (http://
www.cbs.dtu.dk/services/SignalP/) [27]. The N-linked glyco-
sylation sites were also predicted by using the prediction
tool NETNGLYC 1.0 Server ( />services/NetNGlyc/). A phylogenetic tree based on the
amino-acid sequences was constructed by the neighbor-join-
ing method of the clustal w program as described above.
The robustness of internal branches was estimated by 1000
bootstrap resamplings.
Tissue distribution of the mRNA for anserinase
and CNDP-like protein
The ethical guidelines from the Animal Ethics Committee of
Kagoshima University on animal care were followed. Nile
tilapia (body weight 79 g) was fed a commercial diet twice a
day for 1 month at a water temperature of 25 °C. Fish was
killed by decapitation under anesthesia (50 mgÆL
)1
of benzo-
caine) and tissues were then collected. Total RNA was pre-
pared from tissues of the fish using the TRIzol Reagent, and
the first-strand cDNA was synthesized using the 35-mer oli-
go(dT)-adaptor as described above. Anserinase, CNDP-like
protein and glyceraldehyde-3-phosphate dehydrogenase
(GAPDH) transcripts were detected by PCR using the
following primer sets; 5¢-GGTGTCAACCAGCTCATG
TAC-3¢ (anserinase, forward) and 5¢-GGACACTTCATTCA

GGTACGCG-3¢ (anserinase, reverse); 5¢-TATTCCTCG
CAAGGTCATCGGC-3¢ (CNDP-like protein, forward) and
5¢-GCAGCTTGACTCCCTGAATGTA-3¢ (CNDP-like pro-
tein, reverse); 5¢-GATGCTCCCATGTTCGTCATGGG-3¢
(GAPDH, forward) and 5¢-CAGCATCAAAGATGGA
GGAGTG-3¢ (GAPDH, reverse). The sizes expected were
530 bp, 395 bp, and 517 bp for anserinase, CNDP-like
S. Yamada et al. Purification and sequence identification of anserinase
FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing 6011
protein, and GAPDH, respectively. PCR amplification was
performed in a total volume of 50 lL containing 0.5 lLofa
template, 15 pmol of a forward primer, 15 pmol of a reverse
primer, 1· PCR Gold buffer containing 2.5 mm MgCl
2
,
10 nmol each of dATP, dGTP, dCTP and dTTP, and 1 U
AmpliTaq Gold (Applied Biosystems Japan Ltd). For PCR
the following conditions were used: initial denaturation at
95 °C for 9 min, followed by 45 cycles of denaturation at
94 °C for 20 s, annealing at 56 °C for 30 s, and extension at
72 °C for 1 min, final extension step at 72 °C for 7 min.
Acknowledgements
We thank Mr S. Yamamoto, Ms H. Deguchi and Ms
A. Teshima for technical assistance. We also thank Dr
Steven M. Plakas (FDA, Dauphin Island, AL, USA)
for reviewing the manuscript.
References
1 Baslow MH (1965) Neurosine, its identification with
N-acetyl-l-histidine and distribution in aquatic
vertebrates. Zoologica 50, 63–66.

2 Erspamer V, Roseghini M & Anastasi A (1965) Occur-
rence and distribution of N-acetylhistidine in brain and
extracerebral tissues of poikilothermal vertebrates.
J Neurochem 12, 123–130.
3 Yamada S & Furuichi M (1990) Na-Acetylhistidine
metabolism in fish. 1. Identification of Na-acetylhisti-
dine in the heart of rainbow trout Salmo gairdneri.
Comp Biochem Physiol 97B, 539–541.
4 Yamada S, Tanaka Y & Furuichi M (1995) Partial puri-
fication and characterization of histidine acetyltransfer-
ase in brain of Nile tilapia (Oreochromis niloticus).
Biochim Biophys Acta 1245, 239–247.
5 Baslow MH & Lenney JF (1967) Na-Acetyl-L-histidine
amidohydrolase activity from the brain of the skipjack
tuna Katsuwonus pelamis. Can J Biochem 45 , 337–340.
6 Lenney JF, Baslow MH & Sugiyama GH (1978) Simi-
larity of tuna N-acetylhistidine deacetylase and cod fish
anserinase. Comp Biochem Physiol 61B, 253–258.
7 Jones NR (1955) The free amino acids of fish; 1-methyl-
histidine and b-alanine liberation by skeletal muscle
anserinase of codling (Gadus callarias). Biochem J 60,
81–87.
8 Jones NR (1956) Anserinase and other dipeptidase
activity in skeletal muscle of codling (Gadus callarias).
Biochem J 64, 20.
9 Yamada S, Tanaka Y, Sameshima M & Furuichi M
(1993) Properties of Na-acetylhistidine deacetylase in
brain of rainbow trout Oncorhynchus mykiss. Comp
Biochem Physiol 106B, 309–315.
10 Yamada S, Tanaka Y, Sameshima M & Furuichi M

(1991) Distribution of Na-acetylhistidine-deacetylating
enzyme in tissues of rainbow trout. Nippon Suisan
Gakkaishi 57, 1601.
11 Yamada S, Tanaka Y, Sameshima M & Furuichi M
(1994) Effects of starvation and feeding on tissue
Na-acetylhistidine levels in Nile tilapia Oreochromis nil-
oticus. Comp Biochem Physiol 109A, 277–283.
12 Perry TL, Hansen S, Tischler B, Bunting R & Berry K
(1967) Carnosinaemia: a new metabolic disorder asso-
ciated with neurological disease and mental defect.
N Engl J Med 277, 1219–1227.
13 Murphey WH, Lindmark DG & Mosovich L (1973)
Serum carnosinase deficiency concomitant with mental
retardation. Pediatr Res 7, 601–606.
14 Lenney JF, George RP, Weiss AM, Kucera CM, Chan
PW & Rinzler GS (1982) Human serum carnosinase:
characterization, distinction from cellular carnosinase,
and activation by cadmium. Clin Chim Acta 123,
221–231.
15 Lenney JF, Peppers SC, Kucera-Orallo CM & George
RP (1985) Characterization of human tissue carnosi-
nase. Biochem J 228, 653–660.
16 Lenney JF (1990) Human cytosolic carnosinase: evi-
dence of identity with prolinase, a non-specific dipepti-
dase. Biol Chem Hoppe Seyler 371, 167–171.
17 Jackson MC, Kucera CM & Lenney JF (1991) Purifica-
tion and properties of human serum carnosinase. Clin
Chim Acta 196, 193–205.
18 Teufel M, Saudek V, Ledig JP, Bernhardt A, Boularand
S, Carreau A, Cairns NJ, Carter C, Cowley DJ,

Duverger D, et al. (2003) Sequence identification and
characterization of human carnosinase and a closely
related non-specific dipeptidase. J Biol Chem 278,
6521–6531.
19 Rawlings ND, Tolle DP & Barrett AJ (2004) MEROPS:
the peptidase database. Nucleic Acids Res 32, 160–164.
20 Otani H, Okumura N, Hashida-Okumura A & Nagai K
(2005) Identification and characterization of a mouse
dipeptidase that hydrolyzes 1-carnosine. J Biochem
(Tokyo) 137, 167–175.
21 Yamada S, Tanaka Y, Sameshima M & Furuichi M
(1992) Occurrence of Na-acetylhistidine in the muscle
and deacetylation by several tissues of Nile tilapia
(Oreochromis niloticus). Comp Biochem Physiol 103B,
579–583.
22 Margolis FL, Grillo M, Brown CE, Williams TH,
Pitcher RG & Elgar GJ (1979) Enzymatic and immuno-
logical evidence for two forms of carnosinase in the
mouse. Biochim Biophys Acta 570, 311–323.
23 Margolis FL, Grillo M, Grannot-Reisfeld N &
Farbman AI (1983) Purification, characterization and
immunocytochemical localization of mouse kidney
carnosinase. Biochim Biophys Acta 744, 237–248.
24 Pera EM, Martinez SL, Flanagan JJ, Brechner M,
Wessely O & De Robertis EM (2003) Darmin is a novel
Purification and sequence identification of anserinase S. Yamada et al.
6012 FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing
secreted protein expressed during endoderm develop-
ment in Xenopus. Gene Expr Patt 3, 147–152.
25 Laemmli UK (1970) Cleavage of structural proteins dur-

ing assembly of the head of bacteriophage T4. Nature
227, 680–685.
26 Lenney JF, Kan SC, Siu K & Sugiyama GH (1977)
Homocarnosinase: a hog kidney dipeptidase with a
broader specificity than carnosinase. Arch Biochem
Biophys 184, 257–266.
27 Nielsen H, Engelbrecht J, Brunak S & Heijne G (1997)
Identification of prokaryotic and eukaryotic signal pep-
tides and prediction of their cleavage sites. Protein Eng
10, 1–6.
S. Yamada et al. Purification and sequence identification of anserinase
FEBS Journal 272 (2005) 6001–6013 ª 2005 The Authors Journal compilation copyright 2005 FEBS ⁄ Blackwell publishing 6013