Báo cáo khoa học: Mechanistic studies on bovine cytosolic 5¢-nucleotidase II, an enzyme belonging to the HAD superfamily doc

Mechanistic studies on bovine cytosolic 5¢-nucleotidase II, an enzyme
belonging to the HAD superfamily
Simone Allegrini
1,
*, Andrea Scaloni
2,
*, Maria Giovanna Careddu
1
, Giovanna Cuccu
3
, Chiara D’Ambrosio
2
,
Rossana Pesi
3,
*, Marcella Camici
3
, Lino Ferrara
2
and Maria Grazia Tozzi
3
1
Dipartimento di Scienze del Farmaco, Universita
`
di Sassari, Italy;
2
Proteomics and Mass Spectrometry Laboratory, ISPAAM,
National Research Council, Naples, Italy;
3
Dipartimento di Fisiologia e Biochimica, Universita
`

di Pisa, Italy
Cytosolic 5¢-nucleotidase/phosphotransferase specific for
6-hydroxypurine monophosphate derivatives (cN-II),
belongs to a class of phosphohydrolases that act through t he
formation o f an enzyme–phosphate intermediate. Sequence
alignment with members of the P-type ATPases/L-2-halo-
acid dehalogenase superfamily identified three highly con-
served motifs in cN-II and other cytosolic nucleotidases.
Mutagenesis studies at specific amino acids occurring in
cN-II conserved motifs were performed. The modification of
the m easured k inetic parameter s, c aused by conservative a nd
nonconservative substitutions, suggested that motif I is
involved in the formation and stabilization of t he covalent
enzyme–phosphate intermediate. Similarly, T249 in motif II
as well as K292 in motif III also contribute to stabilize the
phospho–enzyme adduct. Finally, D351 and D356 in motif
III coordinate magnesium ion, which is required for cata-
lysis. These findings were consistent with data already
determined for P -type ATPases, haloacid dehalogenases and
phosphotransferases, thus suggesting that cN-II and other
mammalian 5¢-nucleotidases are characterized by a 3D
arrangement related to the 2-haloacid dehalogenase super-
fold. Structural determinants involved in differential regu-
lation by nonprotein ligands and redox reagents of the two
naturally occurring cN-II forms generated b y proteolysis
were ascertained by combined biochemical and mass
spectrometric investigations. These experiments indicated
that the C-terminal r egion of cN-II contains a cysteine prone
to form a disulfide bond, thereby inactivating the enzyme.
Proteolysis events that generate the observed cN-II forms,

eliminating this C-terminal portion, may prevent loss of
enzymic activity and can be regarded a s regulatory pheno-
mena.
Keywords: catalytic residues ; HAD; nucleotidase; regulation;
site-directed mutagenesis.
Mammalian 5¢-nucleotidases (eN, cN-Ia, cN-Ib, cN-II,
cN-III, cdN and mdN) make up a family of proteins with
different subcellular locations and remarkably low sequence
similarities [1]. Besides ectosolic 5¢-nucleotidase, one mito-
chondrial and five cytosolic enzymes have been described
to date. According to its substrate specificity and tissue
distribution, each protein seems to play a specific role within
the cell. In fact, cN-Is, which is highly expressed in skeletal
muscle, heart and testis, is specific for AMP and seems to be
involved in adenosine production during hypoxia or
ischemia, because it mediates the cell response to low
energy charges [2]. On the other hand, cN-II is more specific
for i nosine monophosphate (IMP) and GMP, and i s a
ubiquitous enzyme involved in the regulation of intracellular
IMP and GMP concentrations [3]. Furthermore, cN-III,
which is expressed in red blood cells and is specific for
pyrimidines, seems to participate in RNA degradation
during erythrocyte maturation [4]. Likewise, cytosolic and
mitochondrial deoxynucleotidases (cdN and mdN) regulate
nucleotide pools in their respective compartments [1].
cN-II was the first member of the cytosolic 5¢-nucleotid-
ases whose r eaction mechanism was elucidated [5]. During
catalysis, this enzyme was shown to become phosphorylated
on the first aspartate of its DMDYT sequence. A similar
motif DXDX(T/V) (motif I) is present in all members of the

HAD superfamily, where the nucleophilic attack of this
aspartate is essential for the catalytic machinery [6–8].
P-type ATPase/phosphotransferase members of the HAD
superfamily share a similar structural fold and a co mmon
reaction mechanism, which requires the formation o f a
covalent enzyme–phosphate inte rmediate [8]. Furthermore,
crystallographic and site-directed mutagenesis studies on
these p roteins have demonst rated that a series of other
common amino acids always occur in their active site [7–9],
thus confirming the presence of two additio nal sequence
motifs common to all members of the HAD family [8,9].
The first (motif II) is characterized by a threonine/serine
residue included in a hydrophobic region; the second (motif
III) presents a conserved lysine and a pair of aspartic acid
Correspondence to S. Allegrini, Universita
`
di Sassari, Dipartimento di
Scienze del Farmaco, via Muroni 23/A, 07100 Sassari, Italy.
Fax: +39 079 228708, Tel .: +39 079 228715,
E-mail:
Abbreviations: BPG, 2,3-biphosphoglycerate; CAM, carboxyamido-
methylated; cdN, cytosolic deoxynucleotidase; cN, cytosolic nucleo-
tidase; eN, ectosolic nucleotidase; HAD, L-2-haloacid dehalogenase;
IMP, inosine monophosphate; mdN, mitochondrial deoxynucleoti-
dase; PSP, phosphoserine phosphatase.
*Note: Th ese authors c ontributed e qu ally to t he w ork p resent ed in this
article.
(Received 3 August 2004, revised 11 October 2004,
accepted 25 October 2004)
Eur. J. Biochem. 271, 4881–4891 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04457.x

residues. Very recently, the resolution of the crystal structure
of mdN, a dimeric mitochondrial nucleotidase specific for
deoxynucleotides, has been reported, proving that this
enzyme is the first example of a 5¢-nucleotidase belonging to
the HAD superfamily [7]. On this basis, a large number
of proteins differing in catalytic activity against various
substrates, polypeptide length (from 200 to 1400 amino
acids), domain arrangement, oligomerization and conform-
ational change following ligand binding, have been related
to the H AD/P-type ATPases/phosphotransferases s uper-
fold [10]. However, no structural data on cN-II are currently
available.
Unlike other 5¢-nucleotidases, cN-II activity is modula-
ted by various ligands; it is activated by ADP, ATP,
2,3-biphosphoglycerate (BPG) and decavanadate, and is
inhibited by phosphate. On the basis of these regulatory
properties, its physiological role has been hypothesized as
being associated with the hydrolysis o f excess IMP that has
been newly synthesized or salvaged in the presence of a high-
energy charge [11]. The enzyme generates inosine, which, in
turn, can leave the cell and/or be converted into hypoxan-
thine and uric acid. However, when I MP accumulates as a
consequence of ATP hydrolysis, cN-II becomes virtually
inactive, allowing the accumulation of the monophosphate
and p reventing the loss of precious purine molecules
[3,11–14]. T wo enzyme forms of bovine cN-II have been
reported, which can be distinguished in terms of electroph-
oretic, chromatographic a nd regu latory characteristics [ 13].
The physiological relevance of t his observation remains
obscure, together with the nature (either genetic or regula-

tive) of the mechanisms generating these species. Moreover,
cN-II presents both phosphatase and phosphotransferase
activities. Even though the physiological re levance of the
phosphotransferase activity is not clear, the enzyme has
been demo nstrated as being responsible for the phosphory-
lation of nucleoside analogs in use as a ntineoplastic and
antiviral drugs [15,16]. Furthermore, c N-II seems to b e
responsible for t he resistance to several purine derivative
drugs [17,18]. Therefore, it would seem that cN-II plays a
fundamental role in the effectiveness of several purine drugs
and its activity may be predictive of patient survival in acute
myeloid leukaemia [19]. Finally, cN-II overactivity has been
demonstrated in Lesch–Nyhan syndrome, which might be
associated with neurological symptoms related to this
disease [20–22].
For these reasons, biochemical studies, aimed at com-
pletely elucidating the cN-II structure with re spect to its
functional and regulatory properties, are particularly
important. In fact, these investigations will be fundamental
for the design of nucleoside derivatives that could interfere
with enzyme function and stability, thus playing a role both
in the therapy of malignancies and neurological disorders
caused by purine dismetabolisms. In this article, we report
the kinetic characte rization of a series of cN-II mutants,
designed on the basis o f sequence alignment with P-type
ATPases, haloacid dehalogenases and phosphotransferases.
Our results indicate that cN-II presents an active site
strongly resembling those present in other members of the
HAD superfamily. Furthermore, we investigated t he struc-
tural determinants involved in the regulation of cN-II

activity in the presence of ligands or redox reagents by using
a combined biochemical and p roteomic approach.
Experimental procedures
Materials
Talon metal affinity resin was from Clontech Laboratories
(Palo Alto, CA, USA). [8-
14
C]Inosine was purchased from
Sigma Chemical Co. (St Louis, MO, USA). Thrombin was
from Amersham Pharma cia Biotech (Uppsala, Sweden).
Poly(vinylidene d ifluoride) (PVDF) membrane was pur-
chased from Millipore Co. (Billerica, MA, USA). G oose
anticytosolic 5¢-nucleotidase (from pig lung) IgG and rabbit
anti-goose IgG serum were kind gifts from R. I toh ( Tokyo
Kasei Gakuin University, Tokyo, Japan). All other chem-
icals were reagent grade. All s olvents were HPLC grade.
Sequence alignment
Iterated sequence comparisons and position-specific iterated
PSI-BLAST search results, starting from P-type ATPases
and H AD, were used as starting multiple a lignments [9,23].
Several human and bovine 5¢-nucleotidases (cN-Ia, cN-Ib,
cN-II, cN-III, cdN a nd m dN) were aligned by using the
same approach. These proteins were also analysed for
sequence motif by using the
MOST
program [24] with
stringent cut-offs (e.g. r ¼ 0.0085). Protein secondary
structure was predicted by using the
PROF
,

SCRATCH
/
SSPRO
and
PSIPRED
programs [25–27]. Identified sequence motifs
were verified o n the basis of the predicted secondary
structure. All sequences were further aligned by using the
MACAW
program [28], with minor manual adjustments.
Site-directed mutagenesis
Point mutants were obtained as previously described [6],
with minor changes. The protocol adopted included two
successive PCR reactions. In the first, a mutagenic primer
was used together with a primer specific for cN-II to amplify
a dsDNA fragment (megaprimer) that contained the desired
mutation. Each megapr imer, purified from the agarose gel,
was used in a second PCR reaction together with a second
specific primer, to amplify the final nucleotide fragment,
including specific sites for restriction endonucleases at the 5¢
and 3¢ terminus and, in the central part, the mutated t riplet.
Once it had been cleaved, this fragment was used t o replace
the corresponding one present in the expression plasmid
containing bovine w ild-type cN-II. The specific for ward
primers used in the PCR reactions were: NheI_F)
5¢-CCGCTAGCATGACAACC TCCTG-3¢ (from base )8
to base 14 of the pET28c-cNII construct); and AflII_F)
5¢-CAGTTGACTGGGTTCATT-3¢ (f rom base 611 to base
628). The specific reverse primers used were: KpnI_R)
5¢-AGTAGACGATGCCATGCT-3¢ (from base 982–965);

and Csp45I_R) 5¢-GTTCAGCCAAGAAAATATC-3¢
(from base 1205–1186). The mutagenic primers used were
as follows (the mutagenic triplette is shown in bold):
M53_F) 5¢-TGGGTTTGACANHGATTATACACTTGC
TGTGTA-3¢ (from base 147 to base 179) (potentially able
to produce s ix different m utants: M53I, M53T, M 53N,
M53K, M53S, M53R); T56_F) 5¢-TGGGTTTGACATG
GATTATADNCTTGCTGTGTA-3¢ (from base 147 to
base 179) (potentially able to pr oduce six different mutants:
T56I, T56M, T56N, T56K, T56S, T56R); T249S_F)
4882 S. Allegrini et al. (Eur. J. Biochem. 271) Ó FEBS 2004
5¢-TTTCTTGCCTCCAACAGTGA-3¢ (from base 736
to base 755); T249V_F) 5¢-TTTCTTGCCGTCAACAG
TGA-3¢ (from base 736 to base 755); S251(T/A)_F)
5¢-TGC CACCAACRCTGA CTATAA A-3¢ (from base
741 t o base 762); K292(R/M)_F) 5 ¢-GCACGGAKGC
CACTGTTCT-3¢ (from base 868 to base 886); D351E_R)
5¢-C CCAAAAATGTGCTCTCCAATA-3¢ (from b ase
1065 to base 1044); D 351N_R) 5¢-CCC AAAAA
TGTGCTCTCCAATA-3¢ (from base 1065 to base 1044);
D356E_R) 5¢-A ATCTCCCCAAAAATGTGA TCT-3¢
(from base 1071 to base 1050); D356N_R) 5¢-AAT GTTCC
CAAAAATGTGATCT-3¢ (from base 1071 to base 1050).
Table 1 shows t he primer couples used to produce the
mutants described in this article.
The P CR mixtures and cycling c onditions were as
follows. First PCR mixture: 5 0 lL containing 7.5 ng of
pET28c-cNII DNA as template, 2 l
M
of mutagenic primer,

1 l
M
of specific primer, 200 l
M
of dNTP, 1 m
M
MgSO
4
and 1.25 U of Platinum Pfx DNA polymerase in P CR
reaction buffer. The first PCR c ycling c onditions were:
2minat94°C; 15 s at 94 °C; 30 s a t 5 0–60 °C (dependin g
on the couple of primers used); and 30 s at 68 °C. Steps 2–4
were repeated 30 ti mes. The second PCR mixture was:
25 lL containing 10 ng of pET28c-cNII DNA, all the
megaprimers recovered after purification from the agarose
gel (usually 0.5–0.8 l
M
), 3 l
M
specific primer, 200 l
M
dNTP, 1 m
M
MgSO
4
and 0.7 U of Platinum Pfx DNA
polymerase in PCR reaction buffer. The cycling conditions
in the second PCR were the same as those used in the first
PCR, but in the second PCR, the annealing temperature
was always 60 °C.

Expression of the recombinant proteins
Bovine wild-type and recombinant cN-II mutants were
prepared and purified as previously described [29]. At the N
terminus all the recombinant products presented an addi-
tional MGSSHHHHHHSSGLVPRGSHMAS sequence
(whose amino acids were numbered with negative values)
containing the histidine tag and the thrombin cleavage site.
The p ro tein concentration w as determined according to
Bradford [30], using BSA as a standard. The molar
concentration of the enzymes was determined by using the
calculated subunit molecular mass (67 300 Da).
Electrophoresis and immunoblotting
Electrophoresis under denaturing conditions was performed
on 12% polyacrylamide gels, according to L aemmli [31].
After electrophoresis, proteins were blotted onto a PVDF
membrane. Immunostaining with specific antibody was
carried out as previously described [ 5].
Enzyme assays
Unless stated otherwise, the nucleotidase activity of cN-II
and its mutants was measured as the rate of [8-
14
C]inosine
formation from 2 m
M
[8-
14
C]IMP in the presence of 1.4 m
M
inosine, 20 m
M

MgCl
2
,4.5m
M
ATP and 5 m
M
dithiothre-
itol, as previously described [11]. Phosphotransferase activ-
ity was measured as the rate of [8-
14
C]IMP formation from
1.4 m
M
[8-
14
C]inosine, in the presence of 2 m
M
IMP, 20 m
M
MgCl
2
,4.5m
M
ATP and 5 m
M
dithiothreitol, as previously
described [11]. For the determination of kinetic p arameters
(K
m
and k

cat
) the concentration of t he labelled substrates
ranged from 0.02 to 4 m
M
. A curve of dependence of the
rate of phosphotransferase activity on MgCl
2
concentration
wasusedtodetermineK
50
for MgCl
2
.
Under these experimental conditions, the accumulation
of radiolabeled inosine (nucleotidase activity) represents the
sum of the phosphatase and the phosphotransferase activ-
ities. It has previously been reported that, at a concentration
close to the K
m
value (1.4 m
M
), inosine reduces phosphatase
activity to 50% without affecting the V
max
for both
reactions [11]. Thus, the expected value of 2 was determined
for the ratio between nucleotidase and phosphotransferase
activities, under the experimental conditions used for the
wild-type recombinant cN-II assay. Accordingly, an alter-
ation o f t his r atio for a mutant was considered as being

caused either by an alteration of the K
m
value for one of the
two s ubstrates or by a variation of the k
cat
value for one of
the two activities.
The oxidative inhibitory effect was measured by incuba-
ting the enzyme w ith CuCl
2
(final concentration 1–250 l
M
)
in 50 m
M
Tris/HCl, pH 7.4, for 10 min. Enzyme was
quickly measured for nucleotidase activity, before and after
the addition of 5 m
M
dithiothreitol to the incubation
mixture. P arallel experiments were also performed by
incubating cN-II with or without 20 l
M
5,5¢-dithiobis-
(2-nitro-benzoic acid), in 50 m
M
Tris/HCl, pH 7.4, at room
temperature. At different time-points, samples were with-
drawn and assayed for nucleotidase activity. After 80 min,
Ellman’s reagen t treated-cN-II was added with 5 m

M
dithiothreitol and assayed f or nucleotidase activity.
Structural characterization of 5¢-nucleotidase samples
Purified wild-type r ecombinant 5¢-nucleotidase s amples
(100 lg), obtained by treatment with or without 5 m
M
dithiothreitol, and with or without thrombin (1 lg), in
50 m
M
Tris/HCl, pH 7.4, were alkylated with 1.1
M
iodo-
acetamide i n 0.25
M
Tris/HCl, 1.25 m
M
EDTA, containing
6
M
guanidinium chloride, pH 7.0, at room temperature for
1 m in in the dark. Proteins were freed from s alt and excess
Table 1. Primers used in PCR reactions for the production of cytosolic
nucleotidase-II (cN-II) point mutants. Mut. Pr., mutant primer; Sp. Pr.,
specific primer; MP, megaprimer.
Mutant
First PCR Second PCR
Mut. Pr. Sp. Pr. MP (bp) Sp. Pr.
M53(I/N) M53_F + KpnI_R fi 836 + NheI_F
T56R T56_F + KpnI_R fi 836 + NheI_F
T249S T249S_F + Csp45I_R fi 470 + NheI_F

T249V T249V_F + Csp45I_R fi 470 + NheI_F
S251(T/A) S251(T/A)_F + Csp45I_R fi 465 + NheI_F
K292(R/M) K292(R/M)_F + Csp45I_R fi 338 + NheI_F
D351E D351E_R + AflII_F fi 455 + NheI_F
D351N D351N_R + AflII_F fi 455 + NheI_F
D356E D356E_R + AflII_F fi 461 + NheI_F
D356N D356N_R + AflII_F fi 461 + NheI_F
Ó FEBS 2004 Cytosolic 5¢-nucleotidase II mechanism (Eur. J. Biochem. 271) 4883
reagents by passing the r eaction mixtures through PD10
columns (Amersham Pharmacia Biotech), as previously
reported [32]. Protein samples were manually collected,
lyophilized and analysed/concentrated by SDS/PAGE
under nonreducing conditions.
Bands from SDS/PAGE were excised from the gel,
triturated and washed with water. Proteins were in-gel
digested with trypsin or 2% (v/v) formic acid, as p reviously
described [33]. Digest aliquots were removed and s ubjected to
a desalting/co ncentration s tep o n ZipTipC
18
devices (Milli-
pore Corp., Bedford, MA, USA) before analysis by
MALDI-TOF-MS. Peptide mixtures were eluted from the
ZipTipC
18
in a stepwise manner, using an i ncreasing
concentration of acetonitrile in the elution solution, and
loaded directly on the MALDI target by using the dried
droplet technique and a-cyano-4-hydroxycinnamic as mat-
rix. Samples were analysed on a Voyager-DE PRO mass
spectrometer (Applied Biosystems, Framingham, MA,

USA). Assignments o f the reco rded mass values to individual
peptides were performed on t he bas is o f t heir molecular mass
and proteolytic agent specificity, as previously described [6].
Peptide mixtures were also fractionated by RP-HPLC on
a Vydac 218TP52 column (250 · 2.1 mm), 5 lm, 300 A
˚
pore size ( The Separation Group, Hesperia, CA, USA) by
using a linear, 5–60% gradient of acetonitrile in 0.1% (v/v)
trifluoroacetic acid over 60 min, at a flow rate of
0.2 m LÆmin
)1
. Individual components were collected manu-
ally. Disulfide-containing peptides were identified on the
basis of their mass value. Sequence analysis was performed
by using a Procise 491 protein sequencer (Applied Biosys-
tems) eq uipped with a 140C microgradient HPLC and a
785 A UV detector (Applied Biosystems) for the identifica-
tion of PTH a mino acids.
Results
cN-II and HAD superfamily
CN-II chemical labelling a nd site-directed mutagenesis
experiments identified D52 as a residue that is essential for
enzyme activity and involved in phosphate-adduct forma-
tion [6]. This amino acid occurs in a sequence region similar
to motif I , which is common to all members of the HAD
superfamily [8]. Iterated sequence comparisons and posi-
tion-specific iterated searches starting from bovine cN-II
or different phosphomonoesterases, phosphotransferase s,
phosphomutases and dehalogenases were used to identify,
in the cN-II primary structure, the remaining two motifs

already reported for these proteins. Similarly to HAD
superfamily members, these motifs should contain amino
acids h ypothetically present in the enzyme active site which
are essential f or metal ion coordination, nucleophilic attack
to substrate and stabilization of an excess of negative charge
in the reaction intermediate. Furthermore, analysis of HAD
superfamily members demonstrated that conserved residues
from each of the motifs appear to occur at specific positions
in the succession of secondary structure elements [9]. For
this reason, the sequence of cN-II was also analysed in order
to predict the secondary structure of the protein. In addition
to the already identifie d motif I, this investigation highligh-
ted two separate regions in the cN-II primary structure as
being a ssociated with motif II and motif III (Fig. 1). As
Fig. 1. Multiple alignment of mammalian 5¢-nucleotidases and members of the P-type ATPase-L-2-haloacid dehalogenase (ATPase-HAD) super-
family. Proteins are listed under their SWISS-PROT codes. Bb, Bos bovis;Ec,Escherichia coli;Eh,Enterococcus hirae;Hs,Homo sapiens;Mg,
Mycoplasma genitalium;Psp,Pseudomonas sp.; Sa, Staphylococcus aureus;Sc,Saccharomyces cerevisiae;andSp,Schizosaccharomyces pombe.
Only the t hree common sequence motifs are reported. The numbers in dicate the distances to the N term inus of each protein and the sizes of the gaps
between aligned s egment s. Th e upper, mid dle and lower block of sequences i nclude mammalian 5¢-nucleotidases, members of the HAD superfamily
and P-type ATPases, respectively. Blue shading indicates conserved amino acid residues required for catalytic activity. Red shading indicates
conserved a mino acids alternatively present in motif I. Yellow shading indicates uncharged amino acid r esidues. Common secondary s tructure
elements are indicated as a-helices, b-strand s and l (loop) regions.
4884 S. Allegrini et al. (Eur. J. Biochem. 271) Ó FEBS 2004
clearly illustrated in the reported multiple alignment,
conserved amino acids in each o f the motifs were inserted
in regions that always presented uncharged residues at
specific positions and w ith a well-defined secondary struc-
ture content. Moreover, on the basis of the recent obser-
vation that md N also belongs to the HAD s uperfamily, the
above reported sequence analysis was extended to all

mammalian 5¢-nucleotidases. I terated s equence c ompari-
sons and position-specific iterated searches identified the
three motifs in all of these proteins, except for eN, (Fig. 1)
suggesting that cytosolic and mitochondrial 5¢-nucleotidases
and deoxynucleotidases present a structural arrangement
related to the HAD sup erfamily fold.
Site-directed mutagenesis
To verify the p redicted role for the conserved residues
reported in Fig. 1, 13 mutated cN-II products were
constructed and expressed. Extracts were prepared 16 h
after the addition of isopropyl thio-b-
D
-galactoside (IPTG)
and recombinant wild-type and cN-II mutants were purified
as previously reported [6]. In all cases, SDS/PAGE analysis
showed a single component migrating with an apparent
molecular mass of 60 kDa (data not shown). Purified
proteins were used in kinetic measurements, and their
parameters are reported in Table 2. The fact that all the
expressed p roteins conserved the same chromatographic
behaviour (data not shown) indicated that observed changes
of activity were not cau sed by gross folding problems [34].
In a previous work [6], we demonstrated that both
conservative and nonconservative substitution of D52 and
D54 (motif I) completely a bolished b oth e nzyme a ctivity
and formation of the cN-II–phosphate intermediate.
Replacement at other positions of this motif strongly
affected how cN-II functioned. In fact, substitution at
position 56 (mutant T56R) resulted in a protein devoid of
nucleotidase and phosphotransferase activities. On the other

hand, the mutagenesis of M53 h ad a l ess severe effect.
However, while nonconservative substitutions (M53N)
resulted in a strong decrease in catalytic efficiency, conser-
vative substitution (M53I) led t o e ffects o n b oth enzyme
activity and affinity towards IMP. This latter mutant
exhibited the lowest n ucleotidase vs. phosphotransferase
activity ratio ; it also showed a sigmoid dependence on
Mg
2+
concentration. These results indicate the essential role
of these residues for proper D52 and D54 orientation and
effective c N-II catalysis, confirming the function of motif I,
as deduced b y previous c hemical labelling and site-directed
mutagenesis experiments.
The conserved amino acid present in motif II, which is
common to all HAD superfamily members, is always a
serine or a t hreonine residue. T his amino acid is important
for a correct orientation of the substrate within the active
site through specific hydrogen bonding. The alignment
reported in Fig. 1 shows that T249 is an essential residue for
cN-II activity. However, another amino acid (S251), with
similar properties, occurs closely i n m otif II. In order to
unambiguously identify the residue present in motif II ,
conservative and nonconservative mutants of both amino
acids were p repared. Mutant T249V showed a strongly
reduced enzyme activity and an alteration of K
m
values for
both substrates. On the other hand, a conservative substi-
tution (mutant T249S) yielded a kinetic behaviour more

similar to that of wild-type e nzyme. In addition, the effect
on enzyme ac tivity exerted by a nonconservative mutation
of residue 251 (mutant S251A) was far less pronounced than
that produced by the conservative S251T mutation. These
results demonstrated the essential role of the T249 hydroxyl
group for cN-II catalysis, thus confirming the nature of
motif II deduced by multiple sequence alignments.
Motif III in the HAD superfamily is characterized by the
presence of a conserved lysine residue and two negatively
charged residues which are involved in stabilizing the
negatively charged reaction intermediate and metal ion
Table 2. Effect of point mutations on various kinetic parameters of bovine recombinant cytosolic nucleotidase-II (cN-II). Nucleotidase and phos-
photransferase activities were measured as described in the Experimental procedures. The results reported are the average of at least three
independent assays. IMP, inosine monophosphate; nd, not detectable; ND, not determined; WT, wild type.
Mutant
k
cat
Phosphotransferase
(s
)1
)
Nucleotidase/
phosphotransferase
K
m
IMP
(m
M
)
K

m
inosine
(m
M
) k
cat
/K
m
inosine
K
50
MgCl
2
(m
M
)
WT 40.0 ± 13 1.8 ± 0.1 0.1 ± 0.04 1.1 ± 0.3 36.0 ± 2.4 1.8 ± 0.6
Motif I
M53I 3.7 ± 2.1 0.9 ± 0.05 1.0 ± 0.4 0.8 ± 0.05 4.5 ± 2.3 3.0 ± 1.0 sigmoid
M53N 0.6 ± 0.1 2.4 ± 0.1 0.1 ± 0.01 1.1 ± 0.4 0.6 ± 0.1 ND
T56R nd nd nd nd nd nd
Motif II
T249S 19.8 ± 1.9 3.4 ± 0.2 0.2 ± 0.06 0.9 ± 0.1 22.7 ± 1.2 4.1 ± 1.1
T249V 0.3 ± 0.15 1.5 ± 0.1 0.4 ± 0.25 0.25 ± 0.07 1.1 ± 0.3 4.4 ± 1.2
S251T 3.4 ± 0.6 6.7 ± 0.3 0.3 ± 0.15 1.5 ± 0.5 2.3 ± 0.4 3.5 ± 1.8
S251A 12.3 ± 1.3 3.7 ± 0.2 0.1 ± 0.01 1.0 ± 0.15 11.8 ± 0.5 3.5 ± 1.3
Motif III
K292R nd nd nd nd nd nd
K292M < 0.1 1.8 ± 0.1 ND 0.5 ± 0.1 < 0.2 ND
D351E < 0.1 1.5 ± 0.1 ND 0.9 ± 0.2 < 0.1 > 30

D351N nd nd nd nd nd nd
D356E 2.2 ± 1.0 2.5 ± 0.1 0.2 ± 0.1 1.0 ± 0.25 2.2 ± 0.8 15.0 ± 4
D356N 1.4 ± 0.1 3.2 ± 0.2 0.2 ± 0.06 10.2 ± 0.8 0.1 ± 0.02 > 30
Ó FEBS 2004 Cytosolic 5¢-nucleotidase II mechanism (Eur. J. Biochem. 271) 4885
coordination, respectively. As expected, on the basis of the
proposed alignment, the mutation of K292 (mutant K292R
and K292M) strongly affected cN-II activity. Similarly, t he
mutation of two aspartate residues (D351 and D356)
resulted in very poor nucleotidase and phosphotransferase
activities and a significantly reduced affinity towards Mg
2+
.
Conservative mutations (mutant D351E and D3 56E) had a
less pronounced effect than nonconservative mutations.
Furthermore, D356N show ed a 10-fold increase in the K
m
value for inosine, suggesting a possible role for this amino
acid in the interaction with the second substrate. All these
data confirmed the nature of the residues p resent in motif
III, as deduced by sequen ce alignment.
A general comparison of all the kinetically characterized
mutants showed that the nucleotidase vs. phosphotrans-
ferase activity ratio was significantly altered in two cases:
mutant S251T and mutant M53I . In the first case, this
phenomenon might be caused by a decrease of phospho-
transfer efficiency, as alterations of the K
m
value measured
for both s ubstrates were very slight. In the latter case, the
value observed for this parameter was in line with a 10-fold

increase of the K
m
for IMP.
Sensitivity to oxidizing conditions
It has been reported that freshly purified calf thymus cN-II
displays full activity only i n the presence of dithiothreitol
[5,11,12], suggesting that ox idation may modulate enzyme
properties. To confirm this observation, recombinant wild-
type cN-II was incubated with different concentrations of
CuCl
2
, and its remaining activity was measured before a nd
after the addition of dithiothreitol to the reaction mixture.
The results reported in Fig. 2 A demonstrate that CuCl
2
treatment inhibited the enzyme in a concentration-depend-
ent manner. The loss of activity was reverted by the presence
of the reducing agent. Recombinant wild-type cN-II was
also treated with a different oxidizing agent, 5,5¢-dithiobis-
(2-nitro-benzoic acid), and its remaining activity was
measured. The results shown in Fig. 2B clearly demonstrate
that cN-II was str ongly sensitive t o this r eagent. A fter
80 min of incubation, dithiothreitol was added to the
reaction mixture, resulting i n a complete recovery of
activity, t hus demonstrating that 5,5¢-dithiobis-(2-nitro-
benzoic acid) oxidation can be reverted by reducing agents.
Proteolytic generation of two cN-II forms
In a previous work, w e observed t he simultaneous presence
of two cN-II forms in preparations from calf thymus,
distinguishable f or electrophoretic, c hromatographic and

regulatory properties [35]. In fact, the cN-II species (form B)
with faster electrophoretic mobility (54 kDa) was activated
by ADP and BPG, and a synergistic stimulatory effect of
these compounds was also observed. On the other hand, the
slower migrating species (form A) (59 kDa), was activated
to a greater extent by ADP and BPG, and the synergistic
effect was absent [35]. T o a scertain whether form B was
arising from an intracellular proteolytic event or from a
degradative process during preparation, freshly isolated
tissues were solubilized directly in hot sample buffer for
SDS/PAGE, and the extracted proteins were analysed by
Western blotting following SDS/PAGE (Fig. 3A). Two
major immunoreactive polypeptides migrating at 54 and
59 kDa were d entified, thus demonstrating that the 54 kDa
species is not generated b y a preparation artefact.
We also noted that highly purified recomb inant cN-II
preparations, although stable for activity, degraded very
slowly to enzyme forms with a lower apparent molecular
mass (results not shown). Furthermore, w hen a freshly
prepared recombinant product (60 kDa) was incubated
with thrombin to remove the His-tag-containing sequence,
different polypeptide species were obtained, depending on
the experimental conditions. In addition to the expected
59 kDa polypeptide, overnight incubation with thrombin at
25 °C induced production of a cN-II form with an apparent
molecular mass o f 54 kDa (Fig. 3B), while overnight
incubation with thrombin at 4 °C only gave rise to the
expected 59 kDa protein. Parallel experiments with native
cN-II from calf thymus, containing both form A and form
B, demonstrated that thrombin treatment induced increased

Fig. 2. Effect of CuCl
2
and 5,5¢-dithiobis-(2-nitro-benzoic acid) treat-
ment on the activity of cytosolic nuc leotidase-II (cN-II). Purified
wild-type recombinant cN-II (1 l
M
) was incubated with different
concentration s of CuCl
2
, for 10 min, at room temperature (A). The
rate of ino sine mon ophosphate (IMP) hydrolysis was measured, as
described in the Experimental procedures, b efore (square) and after
(circle) the addition of 5 m
M
dithiothreitol to the incubation mixture.
Similarly, wild-type recombinant c ytoso lic nucleotidase-II (cN-II)
(1 l
M
) was incubated with (square) or without (circle) 20 l
M
5,5¢-dithiobis-(2-nitro-benzoic acid ), at room temperature, and cN-II
activity was measured as described in the Exp erimental p rocedures (B).
After 80 min, the Ellman’s reagen t treated-en zyme was added with
5m
M
dithiothreitol and IMP hydrolysis was measured.
4886 S. Allegrini et al. (Eur. J. Biochem. 271) Ó FEBS 2004
amounts of form B, thus yielding a polypeptide with an
apparent molecular m ass of 54 kDa co-migrating with that
obtained from the recombinant enzyme (Fig. 3B). These

results suggest that an unpredicted site of protease cleavage
is present in the primary structure of cN-II, in addition to
the predicted site present at t he N terminus of the
recombinant product. To demonstrate definitively that the
59 kDa and 54 kDa species obtained from the recombinant
enzyme were similar to those occurring in calf thymus cN-II,
their sensitivity towards different regulatory ligands was
investigated. A s a lread y reported for form B from calf
thymus, the 54 kDa form purified from thrombin-treated
recombinant cN-II w as a ctivated by ADP a nd BPG
(Fig. 4A). Also in this case, the reported synergistic
stimulatory effect of thes e compounds was observed; in
fact, a K
50
value for ADP of 4 m
M
and 1 m
M
was measured
in the absence and in the presence o f 200 l
M
BPG,
respectively. Similarly t o form A from calf thymus, t he
slower migrating recombinant species (59 kDa) was activa-
ted to a greater extent by A DP and BPG, and the reported
synergistic effect was absent (Fig. 4B). In fact, a K
50
value
for A DP of 2.5 m
M

was measured e ither in the presence or
absence of 200 l
M
BPG. Sequencing analysis of all species
reported in Fig. 3B yielded the same N-terminal sequence,
thus indicating that this proteolytic event occurred at the
protein C terminus. Hence, a still-unknown intracellular
proteolytic process, which removed a C-terminal polyp ep-
tide, wo uld seem to control the relative abundance of these
two cN-II forms, thus modulating the different response to
enzyme ligands.
Interestingly, in addition to the above-mentioned differ-
ences in regulatory properties, the two cN-II forms gener-
ated by thrombin treatment presented significant differences
also in their activity a s a function of dithiothreitol concen-
trations (results not shown). In fact, the 59 kDa species
obtained from recombinant wild-type cN-II, similarly to the
unprocessed product, was highly sensitive to dithiothreitol.
On the contrary, the 54 kDa species was fully active, even in
the absence of reducing agents.
Structural characterization of 5¢-nucleotidase forms
In order t o determine the unexpected site for t hrombin
hydrolysis and to identify possible r esidues involved in
cN-II redox regulation, various protein samples were
subjected or not subjected to the reductive conditions used
in the enzymatic assay, and digested or not digested with
thrombin, and then alkylated with iodoacetamide under
denaturing nonreducing c onditions. Following the r eac-
tion, the samples were desalted and analysed by SDS/
PAGE, and in each case yielded a single component. The

excised bands were digested either w ith trypsin o r 2%
(v/v) formic acid a nd the c orresponding peptide m ixtures
were analysed by MALDI-TOF-MS or resolved by
RP-HPLC and further characterized. Table 3 summarizes
the results obtained; in all cases, mass spectrometry
experiments allowed a complete structural characterization
of the analysed species.
As expected, in th e sample prepared under reductive
conditions and in the absence of thrombin, signals at m/z
AB
Fig. 3. Immunob lotting characterization o f different cyto solic nucleotidase-II (cN-II) forms. (A) I mmunoblotting analysis of fresh bovine calf thymus
tissues d irectly homogenized and extract ed in a hot sample buffe r for SDS/PAGE. The analysis was performed on 10 lg of total pro teins. (B)
Immunoblotting analysis of partially purified calf thymus and purified wild-type recombinant cN-II samples following incubation with thrombin.
Lanes 1–3: 3 lgofcalfthymuscN-IItreatedasfollows:lane1,keptat4°C; lane 2, incubated overnight at 25 °C; lane 3, incubated overnight at
25 °C in the presence o f thrombin. Lanes 4–6: 3 lg of wild-type recombinant e nzyme treated as follows: lane 4, k ept at 4 °C; lane 5, incubated
overnight at 25 °C; lane 6, incubated overnight at 25 °C in the presence of thrombin.
AB
Fig. 4. Regulatory effect of ADP and 2,3-biphosphoglycerate (BPG) on
the different cytosolic nucleotidase-II (cN-II) forms g enerated following
thrombin treatment of the wild-type recombinant enzyme. Enzyme
assays were performed in the absence of ATP, as described in the
Experimental procedures. (A) Twenty-five nanograms of the 54 kDa
form were assayed with a variable amount of ADP, in the presence
(open symbols) or absence (closed symbols) of 200 l
M
BPG.
(B) Twenty-five nanograms of the 59 kDa form were assayed with
variable amounts of ADP, in the presence (open symbols) or absence
(closed s ymbols) of 200 l
M

BPG. Inserts show the SDS/PAGE ana-
lysis of the respective enzyme species.
Ó FEBS 2004 Cytosolic 5¢-nucleotidase II mechanism (Eur. J. Biochem. 271) 4887
1093.2, 1731 .2, 2052.1, 20 66.7, 2178.4, 2180.3, 2968.5,
3457.2, 3502.4, 4023.6 (trypsin) and 2131.4, 2920.2,
3035.4, 3242 .8, 3252.6, 37 28.5, 3813.3, 3928.7, 3944.8,
4001.1, 4334.9 (2% formic acid), corresponding to the
carboxyamidomethylated peptide species, clearly demon-
strated that all eight cysteine residues occurring in the
polypeptide chain were present in a reduced state. On the
other hand, the cN-II sample prepared in the presence of
reducing agents and thrombin treatment, in addition to the
mass signals reported above, showed the occurrence of clear
MH
+
signals at m/z 7363.4 (trypsin) and 5742.8, 5959.1,
6349.4 (2% formic acid) that were tentatively assigned to
S-S-containing peptides (Table 3). T hese peaks suggested
the occurrence in this sample of an oxidized cN-II f orm
containing the d isulfide bridge C175-C547, in addition to a
fully reduced enzyme species. This hypothes is w as con-
firmed by Edman degradation analysis of the purified
disulfide-containing peptides. I n fact, the t ryptic peptide
(MH
+
at m/z 7363.4) revealed the occurrence of PTH-Cys-
carboxyamidomethylated (PTH-Cys-CAM) at position 167
and P TH-cystine at pos ition 175, together with the absence
of any PTH amino acids at position 547 [36]. Similarly,
sequencing analysis of the acid-generated peptides (MH

+
at
m/z 5742.8, 5959.1, 6349.4) demonstrated the presence of
PTH-Cys-CAM at position 181 and PTH-cystine at posi-
tion 547, together with the absence of any PTH amino acids
at position 175 [36]. These results definitively d emonstrate
the nature of the oxidized enzyme and the extreme
sensitivity of C175 and C547 to changes in redox conditions.
Different data were generated from the MALDI-TOF-
MS analysis of a cN-II sample obtained f ollowing pro-
longed thrombin treatment. This species migrated with an
apparent mass of 54 kDa. The occurrence in t he spectr a of
new signals at m/z 1101.8, 1318.5, 1398.7, 1709.2, 2167.8,
3623.7 (2% formic acid), as well as the disappearance of the
signals at m/z 1901.3, 4023.6 (trypsin) and 3280.6, 3728.5,
3944.8, 4334.9 (2% formic acid), corresponding to the
N- and C -terminal r egion o f t he intact protein, clearly
demonstrated that, in addition to the expected site (R-6)
present at t he N t erminus, cN-II was hydro lysed by
thrombin also at R526 (Table 3). C onsequently, in t he
spectrum there were no mass signals corresponding to
peptides containing the d isulfide bridge C175–C547. These
data were in perfect agreement with the above-mentioned
insensitivity o f the thromb in-generated 54 kDa species to
the o xidative conditions. In f act, as thrombin-treated cN-II
is devoid of the C-terminal p eptide containing C547,
involved in the S -S bridge , i t i s no l onger sensitive t o
changes in redox conditions.
Discussion
The HAD fold defines a versatile hydrolase/mutase/trans-

ferase superfamily which a ppears to f unction on the
Table 3. MALDI-TOF-MS analysis of air-exposed, reduced and thrombin-treated cytosolic nucleotidase-II (cN-II) samples. Protein s ample s su b-
jected or not subjected to the reductive conditions used in enzymatic assay, and digested or not digested with thrombin, w ere alkylated with
iodoacetamide in denaturing, nonreducing conditio ns and separated by S DS/PAGE under no nreduc ing conditions. B ands were digeste d in situ with
trypsin or 2% (v/v) formic acid and peptide extracts were analysed by MALDI-TOF-MS. For simplicity, the table only reports the carboxy-
amidomethylated (CAM), disulfide-linked (S-S) and N-/C-terminal peptides. The mass is reported as average values.
Trypsin
Peptide
2% (v/v) Formic acid
Peptide
Native
cN-II
MH
+
(m/z)
Reduced
cN-II
MH
+
(m/z)
Thrombin-
treated cN-II
MH
+
(m/z)
Native
cN-II
MH
+
(m/z)

Reduced
cN-II
MH
+
(m/z)
Thrombin-
treated cN-II
MH
+
(m/z)
634.6 634.7 634.2 (35–39) 1101.8 (519–526)
654.8 654.9 654.7 (522–526) 1318.5 (517–526)
675.7 675.4 675.9 (30–34) 1398.7 ()5–7)
1093.5 1093.2 1093.0 (178–186)CAM 1709.2 (514–526)
1433.3 1433.6 1433.5 (516–526) 1934.3 1934.1 1934.6 (497–513)
1475.4 1475.4 1475.8 (510–521) 2131.2 2131.4 2131.7 (171–187)CAM
2
1554.6 1554.7 1554.1 ()5–8) 2324.9 2324.2 2324.5 (497–516)
1730.9 1731.2 1730.6 (48–61)CAM 2920.7 2920.2 2920.7 (147–170)CAM
1901.0 1901.3 ()23–6) 3035.3 3035.4 3035.7 (146–170)CAM
2052.2 2052.1 2052.6 (426–442)CAM 3242.5 3242.8 3242.9 (307–337)CAM
2067.0 2066.7 2066.9 (112–129)CAM 3252.9 3252.6 3252.4 (432–459)CAM
2178.2 2178.4 2178.2 (178–195)CAM 3280.7 3280.6 ()23–7)
2180.8 2180.3 2180.9 (425–442)CAM 3623.7 (497–526)
2968.3 2968.5 2968.1 (314–342)CAM 3728.3 3728.5 (519–549)CAM
3253.5 3253.6 3253.9 (479–507) 3813.1 3813.3 3813.6 (114–145)CAM
3456.7 3457.2 3457.0 (150–177)CAM
2
3928.5 3928.7 3928.3 (114–146)CAM
3502.1 3502.4 3501.9 (99–129)CAM 3944.5 3944.8 (517–549)CAM

4023.4 4023.6 (527–560)CAM 4000.8 4001.1 4000.5 (21–52)CAM
7363.4 (150–177)CAM +
(527–560)S-S
4335.2 4334.9 (514–549)CAM
5742.8 (171–187)CAM + (519–549)S-S
5959.1 (171–187)CAM + (517–549)S-S
6349.4 (171–187)CAM + (514–549)S-S
4888 S. Allegrini et al. (Eur. J. Biochem. 271) Ó FEBS 2004
common scheme of an e nzyme-aspartate ester formation
followed by the transfer of the ester group to a nucleophile
molecule, i.e. water for hydrolases and dehalogenases, a
different chemical m oiety o n the initial substrate for
mutases, or a second substrate for transferases. With the
sole exception of dehalogenases, all family members require
Mg
2+
ion in the active site to promote covalent intermediate
formation. Enzyme functionalities involved in catalysis are
conserved within members. They appear to be juxtaposed in
space, as determined on the basis of the crystallographic
structures solved to date [7,9,37,38]. This determines the
occurrence of three specific sequence motifs conserved
among all superfamily members [9,23]. On the other hand,
the utilization of the HAD fold for diverse functions is
demonstrated by its rearrangement in a variety of topolo-
gical variants [38]. In all cases, catalysis always involves
residues present in an a/b Rossmann-fold domain c onsisting
of a centrally located p arallel b-sheet surrounded by
a-helices. Depending on the nature of t he enzyme, t his core
presents large helical insertions which lead to additional

noncatalytic domain(s ). E xcluding the three conserved
motifs, the level of sequence divergence a mong proteins
belonging t o the HAD superfamily is so elevated (sequence
identity < 10%) that the relationship between the family
members would go unrecognized. This means that some
members of t his family may not yet have been identified, as
a result of the limitation in the approaches used for sequence
database search and/or various molecular dimensions of the
analysed polypeptides. In our opinion, this is true for the six
mammalian 5¢-nucleotidases reported in F ig. 1. This hypo-
thesis, originally proposed by our laboratory [6], has been
recently c onfirmed by the resolution of the crystallographic
structure of mdN [7].
The characterization of the reaction mechanism [5], the
absolute requirement of a Mg
2+
ion [11,12], the detection of
a phosphorylated intermediate involving the first aspartate
of its DMDYT motif [6] and the identification of the three
conserved motifs reported in this study, all strongly support
the idea t hat cN-II also belongs to the HAD superfamily.
Mutagenesis studies at specific amino acid positions
predicted b y the reported alignment allowed us to identify
a series of residues essential for cN-II c atalysis. In fact, the
modification of m easured kinetic parameters caused by
conservative and nonconservative substitutions suggested a
specific role of these amino acids in the cN-II active site. Our
results are perfectly in line with t hose already reported for
phosphoserine phosphatase [8,39] and Ca
2+

-ATPase [40–
42], thus demonstrating that cN-II presents a catalytic
machinery which very much resembles those of the other
members of the HAD superfamily. Table 4 summarizes the
results obtained for these three enzymes with mutants at
equivalent positions.
Similarly to phosphoserine phosphatase (PSP) and Ca
2+
-
ATPase, mutations of the two aspartates in motif I (D52
and D54) totally abolished cN-II activity. The first residue is
directly responsible for the formation of the enzyme–
phosphate intermediate [6], and the second would seem to
be involved in adduct stabilization and Mg
2+
ion coordi-
nation, as already observed for the other two enzymes [9,38].
Conservative mutations at cN-I I D54 had a more pro-
nounced effect than those observed a t t he cor responding
residues in phosphoserine phosphatase and Ca
2+
-ATPase,
thus emphasizing the important role of this amino acid for
effective c N-II c atalysis. In addition, our e xperiments
demonstrated, for the first time, that mutations at other
conserved positions of motif I also affect enzyme function-
ing ( Fig. 1) (Table 2). In fact, substitution at position 56
(mutant T56R) resulted in a protein totally devoid of
nucleotidase and ph osphotransferase activity. Even though
this effect might be caused by differences in relative steric

hindrance, this conserved threonine residue has been
reported t o establish a specific h ydrogen bond e ssential
for a correct positioning of the active site nucleophile in the
structure of 2 -haloacid dehalogenases and phosphonoacet-
aldehyde hydrolases [43]. S imilarly, kinetic analyses o f M53
mutants d emonstrate that replacements in this position can
also affect cN-II catalysis, probably by influencing the
correct orientation of the two aspartate residues.
On the other hand, the clear similarity observed in the
effect of conservative and nonconservative m utations at
T249, with respect to S109 in phosphoserine phosphatase
and T625 in Ca
2+
-ATPase, identified this amino acid as the
cN-II r esidue of motif I I e ssential f or enzyme catalysis
(Table 4). These results are in line with t he hypothesis that
the h ydroxyl group of T249 is implicated in stabilization of
the covalent intermediate, as already demonstrated for the
corresponding residues of motif II in mdN, HAD and PSP
[7,37,38]. Finally, the effect of mutations at the conserved
lysine and the two negatively c harged residues putatively
present i n cN-II motif III paralleled well with those
observed for phosphoserine phosphatase and Ca
2+
-ATPase
(Table 4). A comparison of the kinetic parameters observed
for these proteins suggests t hat, in cN-II, K292 is the basic
amino acid essential for the stabilization of the negatively
Table 4. Comparison of the effect of mutations on the activity o f three
different enzymes of the L-2-haloacid dehalogenase (HAD) superfamily.

The values reported indicate the relative activity with respect to the
wild-type enzyme. Residues in equ iva len t positio ns in Fig. 1 are in the
same row. Results a re taken from the following references: cytosolic
nucleotidase-II (cN-II), this article and [6] phosphoserine phosphatase
[8], and [29] Ca
2+
ATPase [30–32]. PSP, phosphoserine phosphatase;
WT, wild type; low exp, low e xpression.
cN-II PSP Ca
2+
-ATPase
Mutation Activity Mutation Activity Mutation Ca
2+
transport
WT 100 WT 100 WT 100
Motif I
D52E 0 D20E 0 D351E 0
D52A 0 D20N 0 D351N 0
D54E 0 D22E 50 T353S 20
D54A 0 D22N 0 T353A 0
Motif II
T249S 20 S109T 115 T625S 79
T249V 1.6 S109A 6 T625A low exp
Motif III
K292R < 0.1 K158R 1
K292M 0.1 K158A < 0.4
D351E 0.7 D179E 78 D703E 31
D351N < 0.1 D179N 0.6 D703N < 5
D356E 2.3 D183E 63 D707E < 5
D356N 0.6 D183N < 0.4 D707N < 5

Ó FEBS 2004 Cytosolic 5¢-nucleotidase II mechanism (Eur. J. Biochem. 271) 4889
charged reaction i ntermediate. On the other hand, D351
and D356 seem to be involved in metal ion coordination, as
substitutions at these positions completely abolished
enzyme activity and caused a n increase i n the Mg
2+
K
50
values. However, i f conservative mutations at the corres-
ponding residues in PSP and Ca
2+
-ATPase generated
enzyme species still presented a certain residual activity,
these mutations in cN-II strongly affecte d enzyme catalysis
suggesting that, a s already observed for motif I, Mg
2+
coordination in cN-II strictly requires aspartate residues.
Therefore, according to the mutagenesis and chemical
labelling experiments reported above, the active site of cN-II
should be similar to that schematically represented in Fig. 5.
A careful comparison of the sequence a nd secondary
structure elements with other 5¢nucleotidases and members
of the HAD superfamily revealed that cN-II, in addition to
the canonical a/b core domain responsible for enzyme
catalytic activity, presents a large noncatalytic extension at
its C terminus. A differential processing of this domain
identifies the two naturally occurring cN-II forms that
present a distinct response to the regulatory properties
exerted b y v arious ligands and o xidizing reagents . T his
phenomenon was originally hypothesized by us [35] and is

now demonstrated in the work presented above by a
biochemical and structural characterization of the 54 and
59 kDa species observed in calf thymus or generated by
thrombin treatment of the wild-type recombinant product.
These results would seem to show that the cN-II C-terminal
domain is probably involved in the modulation of enzyme
activity, although t he fine structural details associated
with this regulation have not yet been elucidated. A s till-
unknown protease present in calf thymus and other cells
should cleave intact cN-II closely to R526, thus generating
the two differently regulated enzyme forms. Their s imulta-
neous presence should make the enzyme less susceptible not
only to physiological variations in cell energy charge, but
also to fluctuations of cellular redox status, as demonstrated
by the results reported in this work.
In conclusion, the essential residues involved in the
catalysis and regulation of the cN-II-assisted hydrolysis/
phosphate transfer of purine monophosphates were ascer-
tained by a combined mutagenesis/proteomic i nvestigation.
An extended structure-based sequence alignment of
5¢-nucleotidases provided support for a common structural
and m echanistic origin of these enzymes, revealing a strong
relationship to the HAD superfamily. W e are currently
studying the crystallographic structure of the two naturally
occurring forms of cN-II, in the presence or absence of
various nucleotide analogs. These studies should be able to
ascertain the 3D details responsible for the enzyme prop-
erties reported in this work. They will also be useful for
therapeutic d evelopments aimed a t improving nucleotide-
based drugs against malignancies and neurological dis-

orders caused by purine dismetabolisms.
Acknowledgements
This work was financially supported by grants f rom the Italian
National Research Council and from the Italian MURST.
References
1. Bianchi, V. & Spychala, J. (2003) M ammalian 5¢-nucleotidases.
J. Biol. C hem. 278, 46195–46198.
2. Sala-Newby, G.B., Freeman, N.V.E., Skladanowsky, A.C. &
Newby, A.C. (2000) Distinct roles for r ecombinant cyto-
solic 5¢-nucleotidaseIandIIinAMPandIMPcatabolismin
COS-7 and H9c2 rat myoblast cell lines. J . Biol. Chem. 275, 1 1666–
11671.
3. Zimmerman, H. (1992) 5¢-Nucleotidase: mole cular structure and
functional aspects. Biochem. J. 285, 345–365.
4. Rees, D.C., Duley, J.A. & Marinaki, A.M. (2003) Pyrimidine 5¢
nucleotidase deficiency. Br. J. Haematol. 120, 3 75–383.
5. Baiocchi, C., Pesi, R., Camici, M., Itoh, R. & Tozzi, M.G. (1996)
Mechanism of the reaction catalysed by 5¢-nucleotid ase/ph os-
photransferase: formation of a phosphorylated intermediate.
Biochem. J. 317, 797–801.
6. Allegrini, S., Scaloni, A ., Ferrara, L ., Pesi, R., Pinna, P., Sgarrella,
F.,Camici,M.,Eriksson,S.&Tozzi,M.G.(2001)Bovinecyto-
solic 5¢-nucleotidase acts through the formation of an aspartate
52-phosphoenzyme intermediate. J. Bi ol. C hem. 276, 3 3526–
33532.
7. Rinaldo-Matthis, A., Rampazzo , C., Reichard, P., Bianchi, V. &
Nordlund, P. (2002) Crystal structure of a human mitochondrial
deoxyribonucleotidase. Nat. Struct. Biol. 9, 779–787.
8. Collet, J.F., Stroobant, V., Pirard, M., Delpierre, G. & Van
Shaftingen, E. (1998) A new class of p hosphotransferase phos-

phorylated on an aspartate residue in an amino-terminal DXDX
(T/V) motif. J. Biol. C hem. 273, 14107–14112.
9. Aravind, L., Galperin, M.Y. & Koonin, E.V. (1998) The catalytic
domain of the P-type ATPase has the haloacid dehalogenase fold.
Trends Biol. Sci. 23, 127 –129.
10.Wang,W.,Cho,H.S.,Kim,R.,Jancarik, J.J., Yokota, H.,
Nguyen, H.H., Grigoriev, I.V.,Wemmer,D.E.&Kim,S.(2002)
Structural characterization of th e reaction pathway in pho spho-
serine phosphatase: crystallographic Ôsnap sho tsÕ of intermediate
states. J. Mol. Biol. 319, 421–431.
11.Pesi,R.,Turriani,M.,Allegrini,S.,Scolozzi,C.,Camici,M.,
Ipata, P.L. & Tozzi, M.G. (1994) The bifunctional cytosolic
5¢-nucleotid ase: regu lation of the phosphotransferase an d
nucleotidase activities. Arch. Biochem. Biophys. 31 2, 75–80.
12. Tozzi, M.G., Camici, M., Pesi, R., Allegrini, S., Sgarrella, F. &
Ipata, P.L. (1991 ) Nucleoside phosphotransferase a ctivity o f
human colon carcinoma cytosolic 5¢-n ucleotidase. Arch. Biochem.
Biophys. 291, 212–217.
13. Pesi,R.,Baiocchi,C.,Tozzi,M.G. & Camici, M. (1996) Identifi-
cation, separation and characte risation of two form s of cytosolic
Fig. 5. Schematic representation of the pro posed active site of cytosolic
nucleotidase-II (cN-II).
4890 S. Allegrini et al. (Eur. J. Biochem. 271) Ó FEBS 2004
5¢-nucleotid ase/nucleo side phosph otransferase in calf thymus.
Biochim. Biophys. Acta 1294, 191–194.
14. Itoh, R. (1993) IMP/GMP 5¢-nucleotidase. Comp. Biochem .
Physiol. 105B, 1 3–19.
15. Johnson, M.A. & Fridland, A. (1989) Phosphorylation of
2,3-dideoxyinosine by cytosolic 5¢-nucleotid ase of hu man lym-
phoid cells. Mol. Pharmacol. 36, 291–295.

16. Banditelli, S., Baiocchi, C., Pesi, R., Allegrini, S., Turriani, M.,
Ipata, P.L., Camici, M. & Tozzi, M.G. (1996) The phos-
photransferase activity o f cytosolic 5¢-nucleotidase: a purine ana-
log phosphorylating enzyme. Int. J. Bioc hem. Cell. Biol. 28,
711–720.
17. Mazzon, C., Rampazzo, C., Scraini, M.C., Gallinaro, L., K arls-
son, A., Meier, C., Balzarini, J., Reichard, P. & Bianchi, V. (2003)
Cytosolic and mitochondrial deoxyribonucleotidases: a ctivity with
substrate analogs, inhibitors and implications for therapy. Bio-
chem. Pharmacol. 66, 471–479.
18. Galmarini, C.M., Jordheim, L. & Dumontet, C. (2003) Role of
IMP-selective 5¢-nucle otidase (cN-II) in hematological malignan-
cies. Leuk. Lymphoma 44 , 1105–1111.
19. Galmarini, C.M., Thomas, X., Graham, K., El Jafaari, A., Cros,
E., Jordheim, L., Mackey, J.R. & Dumontet, C. (2003) Deoxy-
cytidine kinase and cN-II nucleotidase expression in blast cells
predict survival in acute myeloid leukaemia patients treated w ith
cytarabine. Br. J. Haematol. 122, 53–60.
20. Page, T., Yu A., F ontanesi, J. & Nyhan, W.L. (1997) Develop-
ment disorder associated with increased cellular nucleotidase
activity. Proc.NatlAcad.Sci.USA94 , 11601–11606.
21. Pesi, R., Micheli, V., Jacomelli, G., Peruzzi, L., Camici, M.,
Garcia-Gil, M., Allegrini, S. & Tozzi, M.G. (2000) Development
disorder associated with increased cellular nucleotidase activity.
Neuroreport 11, 1827–1831.
22. Garcia-Gil, M., Pesi, R., Perna, S., Allegrini, S., G iannecchini, M.,
Camici, M. & Tozzi, M.G. (2003) 5¢-Aminoimidazole-4-carboxa-
mide rib oside induces apoptosis in human neuro blastoma cells.
Neuroscience 117, 811–820.
23. Koonin, E.V. & Tatusov, R.L. (1994) Computer analysis of bac-

terial haloacid dehalogenases defines a large superfamily of
hydrolases with diverse specificity. Application of an iterative
approach to database search. J. Mol. Biol. 244, 125–132.
24. Tatusov, R.L., Altschul, S.F. & Koonin, E.V. (1994) Detection of
conserved segments in proteins: iterative scanning o f sequence
databases with alignment blocks. Proc. Natl Acad. Sci. USA 91 ,
12091–12095.
25. Ouali, M. & King, R.D. (2000) Cascaded m ultiple c lassifiers for
secondary structure prediction. Prot. Sci. 9, 1162–1176.
26. Pollastri, G., Przybylski, D ., Ro st, B. & Baldi, P. (2002) Improving
the prediction of protein secondary structure in three and eight
classes using recurrent neural ne tworks and p rofiles. Proteins 47 ,
228–235.
27. McGuffin, L.J., Bryson, K. & Jones, D.T. (2000) The PSIPRED
protein structure prediction server. Bioinformatics 16 , 404–405.
28. Schuler, G .D., Alt schul, S.F. & Lipman , D .J. ( 1991) A workbench
for multiple alignment construction and analysis. Proteins 9,180–
190.
29. Allegrini, S., Pesi, R., Tozzi, M.G., Fiol, J.C., Johnson, R.B. &
Eriksson, S. (1997) Bovine cytoso lic IMP/GMP-spec ific nucleoti-
dase: cloning and expression of active enzyme in Escherichia coli.
Biochem. J. 328, 483–487.
30. Bradford, M. (1979) A rapid and sensitive method for t he quan-
titation of microgram quantities of protein utilizing the prin ciple
of protein-dye binding. Anal. Biochem. 72, 2 48–254.
31. Laemmli, U.K. (1970) Cleavage of structural proteins during the
assembly of the head of bacteriophage T4. Nature (London) 227,
680–685.
32. Cecconi, I., Scaloni, A., Rastelli, G., Moroni, M., Vilardo, P.,
Costantino, L., Cappiello, M., Garland, D., Carper, D., Petrash,

J.M.,DelCorso,A.&Mura,U.(2002)Oxidativemodificationof
aldose reductase induced by copper ion. Definition of th e metal–
protein interaction mechanism. J. Biol. Chem. 277, 42017–42027.
33. D’Ambrosio, C., Talamo, F., Vitale, M.R., Amodeo, P., Tell, G.,
Ferrara, L. & Scaloni, A. ( 2003) Probing the dimeric structure of
porcine aminoacylase 1 by mass spectrometric and modeling
procedures. Biochemistry 42, 4430–4443.
34. Plapp, B.V. (1995) Site-directed mutagenesis: a tool for s tudying
enzyme catalysis. Methods Enzymol. 249, 91–119.
35. Pesi, R., Baiocchi, C., Allegrini, S., Moretti, E ., Sgarrella, F.,
Camici, M. & Tozzi, M.G. (1998) Identification, separation and
characterisation of tw oformsofcytosolic5¢-nucleotidase/
nucleoside phosphotransferase in c alf thymus. Biol. Chem. 379,
699–704.
36. Haniu, M., A cklin, C., Kenney, W.C. & Rohde, M.F. ( 1994)
Direct assignment of disulfide bonds by Edman degradation o f
selected peptide fragments. Int. J. Pept. Protein Res. 43, 81–86.
37. Li, Y.F., Hata, Y., Fujii, T., Hisano, T., Nishihara, M., Kurihara,
T. & Esaki, N. (1998) Crystal structures of reaction intermediates
of L-2-haloacid dehalogenase and implications for the reaction
mechanism. J. Biol. Chem. 273, 15035–15044.
38. Wang, W., Kim, R., Jancarik, J., Yokota, K. & Kim, S.H. (2001)
Crystal structure of phosphoserine phosphatase from Methano-
coccus jannaschii, a hyperthermophile, at 1.8 A resolution. Struc-
ture 9, 65–71.
39. Collet,J.F.,Stroobant,V.&VanShaftingen,E.(1999)Mechan-
istic studies of PSP, an enzime related to P-Type ATPases. J. Biol.
Chem. 274, 33985–33990.
40. Clarke, D.M., Loo, T.W. & MacLennan, D.H. (1990) Functional
consequences of alterations to a mino acids located in the nucleo-

tide binding domain of the Ca
2+
-ATPase of sarcoplasmic reticu-
lum. J. Biol. C hem. 265, 22223–22227.
41. Maruyama, K. & MacLennan, D.H. (1988) Mutation of aspartic
acid-351, lysine-352, and lysine-515 alters the Ca
2+
transport
activity of the Ca
2+
-ATPase expressed in COS-1 cells. Proc. Natl
Acad. Sci. USA 85, 3314–3318.
42. Maruyama, K., Clarke, D.M., Fujii, J., Inesi, G., Loo, T.W. &
MacLennan, D.H. (1989) Functional consequences of alterations
to amino acids lo cated in t he catalytic center ( isoleucine 348
to threonine 357) and nucleotide-binding domain of the Ca
2+
-
ATPase of sarcoplasmic reticulum. J. Biol. Chem. 264, 13038–
13042.
43. Morais, M.C., Zhang, W., Baker, A.S., Zhang, G., Dunaway-
Mariano, D. & Allen, K.N. (2000) The crystal structure of Bacill us
cereus phosphonoacetaldehyde hydrolase: insight into catalysis of
phosphorus bond cleavage and catalytic diversification within the
HAD enzyme superfamily. Bioche mistr y 39, 10385–10396.
Ó FEBS 2004 Cytosolic 5¢-nucleotidase II mechanism (Eur. J. Biochem. 271) 4891