Tài liệu Báo cáo khoa học: Molecular modeling and functional characterization of the monomeric primase–polymerase domain from the Sulfolobus solfataricus plasmid pIT3 doc

Molecular modeling and functional characterization
of the monomeric primase–polymerase domain from
the Sulfolobus solfataricus plasmid pIT3
Santina Prato
1
, Rosa Maria Vitale
2
, Patrizia Contursi
1
, Georg Lipps
3
, Michele Saviano
4
, Mose
´
Rossi
1,5
and Simonetta Bartolucci
1
1 Dipartimento di Biologia Strutturale e Funzionale, Universita
`
degli Studi di Napoli Federico II, Naples, Italy
2 Istituto di Chimica Biomolecolare, CNR, Pozzuoli, Naples, Italy
3 Institute of Biochemistry, University of Bayreuth, Germany
4 Istituto di Biostrutture e Bioimmagini, CNR, Naples, Italy
5 Istituto di Biochimica delle Proteine, CNR, Naples, Italy
In all cell types, chromosomal DNA replication is a
complex process entailing three enzymatic activities:
helicase activity for double-helix unzipping and prim-
ase and DNA polymerase for RNA primer de novo
synthesizing and elongation respectively [1,2].

Based on the biochemical data accumulated to date,
archaeal DNA replication involves a smaller number
of polypeptides at each stage of the process and is thus
just a simpler form of the much more complex eukary-
otic replication machinery [3–6]. Nonetheless, Archaea
are not simply ‘mini Eukarya’. A better definition
would be ‘a mosaic of eukaryal and bacterial systems
with specific archaeal features’. Aspects worth men-
tioning in this respect are the promiscuous nature of
the nucleic acid functions performed by archaeal
primases and the dual, template-dependent and
Keywords
DNA replication; pIT3 plasmid; primase–
polymerase domain; Sulfolobus; terminal
transferase
Correspondence
S. Bartolucci, Dipartimento di Biologia
Strutturale e Funzionale, Universita
`
degli
Studi di Napoli Federico II, Complesso
Universitario di Monte S. Angelo, Via
Cinthia, 80126, Naples, Italy
Fax: +39 0816 79053
Tel: +39 0816 79052
E-mail:
(Received 4 April 2008, revised 23 June
2008, accepted 4 July 2008)
doi:10.1111/j.1742-4658.2008.06585.x
A tri-functional monomeric primase–polymerase domain encoded by the

plasmid pIT3 from Sulfolobus solfataricus strain IT3 was identified using a
structural–functional approach. The N-terminal domain of the pIT3 repli-
cation protein encompassing residues 31–245 (i.e. Rep245) was modeled
onto the crystallographic structure of the bifunctional primase–polymerase
domain of the archaeal plasmid pRN1 and refined by molecular dynamics
in solution. The Rep245 protein was purified following overexpression in
Escherichia coli and its nucleic acid synthesis activity was characterized.
The biochemical properties of the polymerase activity such as pH, tempera-
ture optima and divalent cation metal dependence were described. Rep245
was capable of utilizing both ribonucleotides and deoxyribonucleotides for
de novo primer synthesis and it synthesized DNA products up to several kb
in length in a template-dependent manner. Interestingly, the Rep245 prim-
ase–polymerase domain harbors also a terminal nucleotidyl transferase
activity, being able to elongate the 3¢-end of synthetic oligonucleotides in a
non-templated manner. Comparative sequence–structural analysis of the
modeled Rep245 domain with other archaeal primase–polymerases revealed
some distinctive features that could account for the multifaceted activities
exhibited by this domain. To the best of our knowledge, Rep245 typifies
the shortest functional domain from a crenarchaeal plasmid endowed with
DNA and RNA synthesis and terminal transferase activity.
Abbreviations
AEP, archaeo-eukaryotic replicative primases; dNTP, deoxyribonucleotide; MD, molecular dynamics; prim–pol, primase–polymerase; TdT,
terminal deoxyribonucleotidyl transferase; TP, template ⁄ primer.
FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS 4389
-independent activities that these enzymes perform in
addition to primer synthesis. For example, Sulfolobus
DNA primase has the additional catalytic property of
performing 3¢-terminal nucleotidyl transferase activity
[7,8], and archaeal replicative primases can use deoxy-
ribonucleotides (dNTPs) as a substrate for synthesizing

in vitro DNA strands up to several kb in length [8–10].
Despite their unique multifunctional nature, archaeal
DNA primases share a number of features with eukar-
yal ones and are consequently subsumed within the
superfamily of structurally related proteins called
archaeo-eukaryotic replicative primases (AEPs) [11].
Primase–polymerases (prim–pols) are a novel family of
AEPs which are sporadically found in both bacterio-
phages and crenarchaeal and Gram-positive bacterial
plasmids. In a recent description, they are said to be
typified by the RepA-like protein ORF904 encoded by
the pRN1 plasmid from the hyperthermophilic archa-
eon Sulfolobus islandicus [12,13]. Prim–pols catalyze
both a DNA polymerase and a primase reaction
(hence the name). They are often fused with superfam-
ily III helicases or encoded by genes in proximity to
those encoding such helicases [12]. It has been sug-
gested that both these primases and the associated heli-
cases are the constituent elements of the replication
initiation complex of the corresponding plasmids [12].
Available structural data on the small primase subunit
of the euryarchaeote Pyrococcus furiosus (Pfu) [14], the
S. solfataricus (Sso) [15] and Pyrococcus horikoshii
(Pho) [16] heterodimeric primase complexes and the
prim–pol domain from S. islandicus plasmid pRN1
[13] reveal that the novel fold in the N-terminal mod-
ules of the catalytic cores of AEPs and prim–pols is
unrelated to that of other known polymerases, whereas
the RRM-like fold encompassed by their C-terminal
units is also reported for the catalytic modules of other

polymerases [11]. Furthermore, the conservation of
catalytic aspartate residues and their 3D arrangement
suggest that the catalysis mode is probably comparable
with the two-metal-ion mechanism of both RNA and
DNA synthesis [17].
In a previous study, we reported the findings of an
analysis of the complete sequence of the cryptic plas-
mid pIT3 isolated from the crenarchaeon S. solfatari-
cus strain IT3 [18]. The fully sequenced plasmid
contains six ORFs, the largest of which (ORF915)
spans over half the plasmid genome and encodes a
putative 100 kDa replication protein designated as
RepA [18]. Bioinformatic analyses of the predicted
amino acid sequence showed that the C-terminal half
of the RepA of the pIT3 plasmid is sequence-similar to
the helicases of the phage-encoded superfamily III pro-
teins. The N-terminal half of the pIT3 protein RepA
shows little sequence similarity to both the related
RepA of crenarchaeal plasmids and the ORF904 pro-
tein of the plasmid pRN1, which is the only enzyme
biochemically characterized to date in Sulfolobales
plasmids. Despite low sequence identity, multisequence
alignment highlighted major similarities in short
sequence motifs, e.g. two conserved aspartates in a
local group of hydrophobic amino acid residues which
are known to serve as ligands for divalent cations and
as tags revealing the presence of DNA polymerases in
the active site [18–20].
In this study, we report on the structural and func-
tional characterization of the shortest tri-functional

recombinant prim–pol domain encoded by a crenar-
chaeal plasmid identified to date. Using an approach
combining homology modeling, molecular simulations
and biochemical analysis, we identified a number of
structural features which are likely to account for
diverse nucleic acid synthesis functions associated with
the 1–245 N-terminal domain of the putative replica-
tion protein from the S. solfataricus plasmid pIT3.
Furthermore, a longer variant (Rep516) comprising
the 1–516 N-terminal residues of the pIT3 full-length
replication protein was designed and its nucleic acid
synthetic activity was compared with that exhibited by
Rep245.
Results
Homology modeling and structure–sequence
analysis
The N-terminal domain comprising residues 31–245 of
the orf915-encoded putative replication protein of the
plasmid pIT3 was predicted to be the minimum-length
sequence containing all the functionally relevant struc-
tural motifs [18]. This domain (without the 30 N-termi-
nal residues) was modeled onto the crystallographic
structure of the orf904-encoded bifunctional prim–pol
domain of the archaeal plasmid pRN1 (PDB entry
1RN1) [13], which following PSI-BLAST sequence
search against PDB and FUGUE server fold recogni-
tion was found to be the best possible structural tem-
plate. In point of fact, this template was found to be
the only prim–pol domain from archaeal plasmids that
had been structurally characterized to date.

Despite low sequence identity (29% for the N-termi-
nal 32–103 region, but  17% for the modeled
sequence as a whole), the pairwise alignment in the
modeling procedure (Fig. 1A) shows no gaps and ⁄ or
insertions of more than two residues, highly conserved
residues (highlighted in yellow) are evenly distributed
among archaeal plasmids prim–pol domains, and both
Analysis of the pIT3 prim–pol domain S. Prato et al.
4390 FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS
the acidic residues D101, D103 and D166 and the
adjacent H138 are present in the active site. Moreover,
the construction of a reasonable model for the Rep245
prim–pol domain (as we designate it from now on)
from the pRN1 prim–pol structure was supported by
both the reliable FUGUE server score value (12.45,
with a recommended cut-off of 6) and the secondary
structure profile (data not shown), both of which point
to considerable fold similarity. To build the Rep245
model, we performed 16 pairwise and multiple
alignments of template and target sequences and used
deleted versions of the template structure. In overall
terms, the final model selected by reference to quality
score indices (Modeller objective function, Procheck
and 3D profile) was in agreement with the template.
Its rmsd value was 0.391 A
˚
and had been derived from
backbone superimposition at the Ca atom level in the
following regions: 31–60, 61–123, 128–130, 136–141,
150–159, 164–184, 199–230 and 233–244 of the Rep245

protein, i.e. all regions except those with gaps ⁄ inser-
tions. In the Rep245 model, all secondary structure
template elements were conserved except the b11
strand which connects the a5 and a6 helices in the
pRN1 prim–pol protein. Because of a two-residue gap
in the corresponding region of the Rep245 sequence,
this finding had not been predicted in phd and prof
secondary structure prediction programs (data not
shown). Fold stability was assessed by energy-minimiz-
ing the model thus selected and subjecting it to 1.5 ns
molecular dynamics (MD) simulation in water. Snap-
shots saved every 15 ps were seen to be best fitted at
the heavy atom backbone level with an rmsd value of
1.04 A
˚
. The larger fluctuations we expected actually
occurred in the 183–201 loop region, whereas second-
ary structure content and distribution were found to
undergo no change during the simulation. Compara-
tive analysis of the resulting model (Fig. 1B) and the
template structure revealed that two structural ele-
ments which are highly conserved in prim–pol domains
were absent from the prim–pol domain of pIT3: the
A
B
Fig. 1. Structure-based sequence alignment
of Rep245 prim–pol domain (31–245).
(A) Sequence alignment between 1RNI and
Rep245 prim–pol domains. Secondary struc-
ture elements of the Rep245 model are

reported above the alignment and colored
according to the ribbon representation (cyan
cylinders for a helices, light-cyan cylinders
for 310 helices and light-blue arrows for
b strands). Highly conserved residues within
prim–pol domain sequences from archaeal
plasmids are highlighted in yellow, the three
acidic residues with the histidine of the
active site in red, the loop region in
magenta and the corresponding 1RNI
Zn-stem in gray. Cysteine residues are high-
lighted in green with the disulfide bonds
drawn as green lines. Sequence alignment
of the conserved motif between Pfu-prim-
ase and Rep245 is also reported in the
brown boxed region. (B) Ribbon representa-
tion of Rep245 homology model with a -
helices colored in cyan and b strands in
light-blue. The three acidic residues and the
adjacent histidine are shown as stick bonds
and colored in violet.
S. Prato et al. Analysis of the pIT3 prim–pol domain
FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS 4391
Zn-binding motif and the two disulfide bonds respec-
tively connecting the a4-helix to the b4 strand and the
b9 strand to the b10 strand at the bottom of the
Zn-stem loop in the pRN1 prim–pol structure. How-
ever, because the Zn-stem loop is a fairly self-standing
structure protruding from the interface between the
DNA binding and the active site subdomains, we man-

aged to model the entire domain without it.
Another significant finding concerns the nature of
the acid residues within the active site of Rep245. The
carboxylate triad of Rep245 including the D101, D103
and D166 motif is similar to the triads of X family
DNA polymerases and terminal deoxynucleotidyl
transferases (TdTs) [21], but differs from that of the
pRN1 prim–pol which contains the D111, E113 and
D171 motif. The presence of an aspartic residue in
place of the glutamic one is likely to have functional
implications: a drastic decrease in enzymatic activity
has been observed upon the mutation of aspartate to
glutamate in human terminal TdT enzyme [22].
Structure–function analysis conducted on the
Rep245 prim–pol domain also pointed to K135 and
R186 residues being potentially critical for a putative
primase activity of this domain, because these posi-
tively charged residues: (a) are not conserved in the
pRN1 prim–pol, whose domain performs no primase
activity; and (b) after the best possible fit of the Ca
atoms of the catalytic triad, are positional homologs
of the R148 and K300 residues of P. furiosus archaeal
primase, both of which are known to play a pivotal
role in the activity of this protein [14]. The side-chains
of the first pair of residues, i.e. K135 and R148,
matched almost exactly; those of the second pair were
in close proximity. The R148 residue of the Pfu-prim-
ase is part of a motif which is highly conserved in
archaeal and eukaryotic primases and is also found in
the Rep245 sequence (146-SGRGYH-151 in Pfu-prim

and 133-TGKGYH-138 in Rep245; Fig. 1A), although
not in the prim–pol domain of pRN1. The sequence
similarity observed reflects a comparable spatial
arrangement, because this motif is part of a b-strand-
loop situated close to the active site in either protein.
Again, a strong parallelism was observed for the latter
pair of residues: in the Pfu-primase structure, the
K300 residue is located in a loop left on the active site
and because of its poorly defined electronic density
other authors have suggested that it was likely to
change conformation upon DNA binding [14]; simi-
larly, as in Rep245, the R186 residue lies in the loop
(corresponding to the 1RN1 zinc knuckle motif) posi-
tioned left of the active site, we assumed that it could
plausibly be involved in sequence recognition and
DNA binding.
In sum, sequence–structure analysis highlighted that
the Rep245 domain of the pIT3 plasmid replication
protein shares structural features with other replicative
archaeal and eukaryotic enzymes and suggested simi-
larity at the functional level as well.
Expression and protein purification
Initially, we checked if the orf915 of the pIT3 plasmid
from the archaeal S. solfataricus strain IT3 actually
encoded a DNA polymerase. When the corresponding
protein was produced in E. coli, we found that it could
synthesize DNA products in a template ⁄ primer (TP)-
dependent polymerase reaction.
We designed a truncated variant of the full-length
pIT3 replication protein comprising the N-terminal

amino acids 1–245 and then including the residues pre-
dicted to be responsible for the DNA polymerase and
primase activities, accordingly to the homology model-
ing data (Fig. 2A). As described in the Experimental
Procedures, the deletion gene was amplified using the
PCR of the S. solfataricus plasmid pIT3 [18] and then
cloned into pET-30c(+). In E. coli, the recombinant
protein (from now on Rep245) was highly overexpres-
sed as a fusion with the C-terminal six-residue histidine
tail (LEHHHHHH). The Rep245 obtained from
heated protein extracts was purified to homogeneity in
a two-stage process using, in succession, affinity chro-
matography on HisTrap HP and anionic exchange on
the Q Resource column. SDS ⁄ PAGE analysis revealed
a single band with an expected molecular mass of
 29 kDa (Fig. 2B; lane 5). To assess the quaternary
structure of purified Rep245, we conducted analytical
gel filtration on SuperdexÒ 75 PC 3.2 ⁄ 30. The protein
was eluted at a volume consistent with a monomeric
form (data not shown). As a further purification step a
PhenomenexÒ C4 (with a linear gradient 5–70% aceto-
nitrile and trifluoroacetic acid 0.05%) reverse-phase
column was used.
In addition, a longer variant comprising the N-ter-
minal residues 1–516 (Rep516) and lacking the C-ter-
minal ATP⁄ GTP-binding site motif A was also
designed and the truncated protein was purified under
the same conditions as described for the Rep245
(Fig. 2A,C; lane 1).
Biochemical characterization of Rep245 DNA

polymerase activity
Based on the results of structure–sequence analysis, we
characterized the functions of the Rep245 protein and
tried to determine optimal DNA polymerase activity
conditions.
Analysis of the pIT3 prim–pol domain S. Prato et al.
4392 FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS
The pH dependence of DNA polymerase activity
was investigated in the 5.0–10.0 range using the hetero-
polymeric 40 ⁄ 20-mer TP (Table 1). As shown in
Fig. 3A and Fig. S1, Rep245 was found to be active
over a broad pH range with maximal DNA template
elongation at pH 8.0.
Because all polymerases require divalent cations for
catalysis, we tested the effect of metal ions on enzyme
activity. The influence of Mg
2+
,Mn
2+
and Zn
2+
ions
on the synthesis function of Rep245 was assessed on
TP heteropolymeric DNA as a template (Fig. 3B).
First, because the protein was unable to perform DNA
synthesis without a metal ion activator (Fig. 3B) we
concluded that Rep245 polymerase activity was strictly
dependent on divalent cations. Second, because DNA
synthesis started promptly after the addition of 1 mm
MgCl

2
, reached a peak in the presence of Mg
2+
ions
at 5 mm and was seen to diminish at higher ion con-
centrations, we concluded that the activating metal
preferably used by Rep245 for its DNA polymerase
activity was Mg
2+
at concentrations between 5
and 10 mm (Fig. S1). With Mn
2+
as a cofactor, the
DNA polymerase activity of Rep245 was found to be
optimal at lower ion concentrations (1–2.5 mm) and to
decrease noticeably at increasing amounts of Mn
2+
.
Furthermore, Zn
2+
cations do not support the DNA
polymerization activity of Rep245.
The thermophilicity of Rep245 was characterized by
investigating its polymerase activity at increasing tem-
peratures utilizing the TP heteropolymeric DNA sub-
strate. As shown in Fig. 3C, the peak reached at 65 °C
was followed by rapid decreases in activity at higher
temperatures. This behavior may be traced to melting
synthesis products and ⁄ or enzyme inactivation. A
gel profile of the products is shown in Fig. S1.

Thus, to verify if this unexpectedly low thermophi-
licity level was correlated to structural protein unfold-
ing, far-UV CD spectroscopy was used to assess the
structural stability of the Rep245 mutant. Following
30 min incubation at 60, 70 and 80 °C, we recorded
the CD spectra of the incubated Rep245 samples at
these temperatures. The absence of thermal unfolding
transitions provided evidence that temperature
increases did not result in detectable changes in the
secondary structure of the Rep245 protein (data not
shown). Based on this finding, we could rule out that
the loss of DNA polymerase activity sparked off by
temperature increases in the tested range was to be
traced to thermal enzyme inactivation.
30 prim-pol 245
915
1
Walker A motif
RepA
A
B
C
Rep245
1
245
6His
prim-pol
Rep516
6His
1

516
prim-pol
M1 23 4 5kDa
66
29
45
36
Rep245
29
24
20
kDa
66
M12
Rep516
Rep245
29
45
36
24
20
14.2
Fig. 2. Schematic representation and production of truncated vari-
ants of the replication protein (RepA) of the plasmid pIT3 from Sulf-
olobus solfataricus, strain IT3. (A) RepA, Rep245 and Rep516
indicate the full-length residues, 1–245 and 1–516 truncated
proteins, respectively. The constructs represent the C-terminally
His-tagged proteins. The prim–pol domain and putative heli-
case ⁄ NTPase domain are indicated in gray and black respectively.
(B) Purification of the recombinant Rep245 protein. SDS ⁄ PAGE of

protein extracts at various stages of the purification of Rep245.
Lane M, molecular mass markers; lane 1, crude extract from unin-
duced Escherichia coli control culture; lane 2, crude extract from
induced E. coli (pET-Rep245) cells; lane 3, heat-treated sample; lane
4, eluate from the nickel affinity chromatography; lane 5, eluate
from the Resource-Q cation-exchange column. (C) Purified trun-
cated proteins. SDS ⁄ PAGE of purified Rep245 and Rep516 pro-
teins. Lane M, molecular mass markers; lane 1 and 2, purified
C-His
6
-tagged Rep516 (59 kDa) and Rep245 (29 kDa), respectively.
Table 1. DNA substrates used in this study. The position of the
radioactive label is marked with an asterisk.
Template-primer used for polymerase assay
TP 40 ⁄ 20-mer
40-mer 3¢-GCGCCTCTAACGAAGATAGGATCCGTGTGTCTTAGCTTCC-5¢
20-mer *5¢-CGCGGAGATTGCTTCTATCC-3¢
Oligonucleotides used for TdT assay
TEMP
20-mer *5¢-CGAACCCGTTCTCGGAGCAC-3¢
oligo(dT)
28
S. Prato et al. Analysis of the pIT3 prim–pol domain
FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS 4393
Eventually, heat resistance tests conducted by assay-
ing residual polymerase activity after 15 min incuba-
tion at temperatures between 50 and 80 °C showed
that Rep245 was fairly stable even after incubation at
80 °C, when its residual activity was found to be 60%
of the corresponding level of non-preincubated samples

(Fig. 3D).
Rep245 can synthesize RNA and DNA primers
Next, we addressed the question if Rep245 could display
primase activity. Significantly, following incubation
with M13 mp18 single-stranded DNA in the presence of
a ribonucleotide mixture containing [
32
P]ATP[aP],
Rep245 was actually found to be capable of synthesizing
an alkali-labile 16-base RNA primer as well as a less
abundant 20-mer oligoribonucleotide. RNA primer for-
mation was found to be a specific activity because it
was not detected in the absence of Rep245 (Fig. 4A).
Surprisingly, Rep516, the longer variant comprising
the N-terminal residues 1–516 (Fig. 2A,C, lane 1), was
found to be capable of de novo synthesis of larger
molecular size RNA products (Fig. 4B, lane 1). These
RNA primers formed on the M13 mp18 can be
elongated by Rep516 and Taq DNA polymerase when
further incubation in the presence of dNTPs was
performed (Fig. 4B, lanes 3 and 4). When Rep516 was
omitted, neither a ribonucleotide primer nor elongation
products were observed (Fig. 4B, lane 2).
Another point we set out to investigate was whether
Rep245 could use dNTPs as a substrate for primer
synthesis. For this purpose, primase reactions with
dNTPs as substrates were performed on M13 mp18
single-stranded DNA at temperatures between 5 and
90 °C. Under these reaction conditions, the Rep245
protein was found to efficiently synthesize and elongate

DNA primers into longer products (Fig. 4C). Temper-
ature increases were seen to influence the size of DNA
products: small amounts of DNA primers between 16
and 20 nucleotides in size were synthesized at 30 °C; in
the temperature range between 40 and 65 °C, DNA
primer formation was both more clearly observable
and accompanied by the appearance of longer DNA
products. Because no product was observed when
the protein was not included in the reaction mixture,
this reaction was clearly template dependent and
specific.
The fact that the Rep245 variant retained the capabil-
ity of the RepA full-length protein of synthesizing and
elongating DNA products, although with a reduced
80
100
120
20
40
60
Relative acitivity (%)
Temperature (°C)
0
40 50 60 70 80 90
60
80
100
Relative activity (%)
20
40

0
5678910
pH
80
100
120
Mg
(2+)
80
100
20
40
60
Residual activity (%)
Mn
(2+)
Zn
(2+)
Relative acitivity (%)
20
40
60
0
NP 50 60 70 80
Pre-incubation T (°C)ion concentration (m
M)
0
0 1 2.5 5 10 50
AC
BD

Fig. 3. Effects of pH, divalent cations and temperature on Rep245 polymerase activity. Polymerase activity was assayed on TP heteropoly-
meric 40 ⁄ 20-mer DNA as the substrate. Reaction products were separated on a 20% polyacrylamide ⁄ urea gel and quantified by PhosphoIm-
ager. (A) Graphical representation of the pH dependence. Buffer systems (25 m
M final concentration and pH measured at 65 °C) were as
follows: Na-acetate (pH 5.0, 5.4 and 5.8), Tris ⁄ HCl (pH 6.5, 7.0, 7.5 and 8.0) and glycine ⁄ NaOH (pH 8.6, 9.0 and 9.6). (B) Dependence of
Rep245 polymerase activity on metal ions. The results are the means of three independent experiments. (C) The dependence of polymerase
activity on the temperature was determined by assaying the enzyme in the standard reaction mixture at the indicated temperatures. (D)
Thermal stability of Rep245 was tested by pre-incubating the enzyme for 20 min at the indicated temperatures (NP, not pre-incubated);
enzyme residual activity was then assayed on TP heteropolymeric 40 ⁄ 20-mer DNA, as described in Experimental procedures.
Analysis of the pIT3 prim–pol domain S. Prato et al.
4394 FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS
specific activity value (0.607 nmol dNTPsÆmin
)1
Æmg
)1
protein i.e.  20% of the corresponding level of the
RepA full-length protein’s polymerase activity measured
by the DE-81 filter binding assay) was evidence that our
structural homolog model included an active DNA
polymerase and primase domain within the N-terminal
1–245 amino acids of the pIT3 replication protein.
Furthermore, the progressive accumulation of smal-
ler length products observed for Rep245 might point
to high-frequency enzyme–DNA dissociation during
catalysis as a result of the higher temperatures. When
Rep516 was tested under identical assay conditions we
observed a more pronounced increase in RNA ⁄ DNA
synthesis. As shown in Fig. 4C, Rep516 mainly synthe-
sized larger molecular size DNA products that had not
entered the polyacrylamide gel; a negligible accumula-

tion of smaller products was only observed at 80 and
90 °C, suggesting that Rep516 was more active than
Rep245 in performing DNA synthesis. Hence the dif-
ferent efficiency in de novo RNA ⁄ DNA synthesis can
be ascribed to additional residues responsible for the
lesser frequency with which this enzyme is dissociated
from DNA during catalysis.
Taken together, these findings indicate that besides
performing RNA primer synthesis activity, the Rep245
and Rep516 proteins can both incorporate dNTPs for
de novo primer synthesis and elongate these primers
into larger DNA products, though the efficiency to
make long products of Rep516 is higher than that of
the smaller Rep245 variant and is comparable with the
wild-type protein. In conclusion, the Rep245 domain
contains the catalytic residues required for both
primase and polymerase activities.
Rep245 performs 3¢-terminal nucleotidyl
transferase activity
During our primase activity test, we observed that
following incubation with poly(dT), Rep245 syn-
thesized greater than template-length DNA primers
(data not shown). To establish whether the protein
could also perform a non-template synthesis function
we resolved to verify whether different 5¢-end labeled
oligonucleotides underwent elongation in the presence
of unlabeled (d)NTPs. For this purpose, individual
DNA substrates were incubated with Rep245 and
separately supplied with each of the four (d)NTPs. As
shown in Fig. 5, Rep245 was found to preferentially

incorporate dATP and dGTP used for the test at the
3¢-end of the 28-mer homo-oligomer (oligodT) and
20-mer heteropolymeric (TEMP) substrates, respec-
tively (for sequence details see Table 1), albeit at
different levels of efficiency (Fig. 5A,C). Interestingly,
template
KOH
A
C
B
Rep245
+
–
+
–
+–
+
+
++–
–
1234
16 nt
20 nt
28 nt
28 nt
20 nt
35 n
t
ATP
ATP

Temperature [°C]
Rep516
C
5
50
70
65
80
Rep245
5
50
70
65
80
C
16 nt
20 nt
28 nt
dATP
d
Fig. 4. Primase activity of Rep245 and Rep516 proteins. (A) RNA
primer synthesis. Reaction mixtures, containing M13 single-
stranded circular DNA, NTPs including [
32
P]ATP[aP], and Rep245
(or Rep516), were incubated at 60 °C for 30 min. ss20-mer, ss28-
mer and ss35-mer oligonucleotides were 5¢ labeled with
[
32
P]ATP[cP] and used as markers. (B) Rep516 synthesized and

elongated RNA primers (lane 1) that can be extended to longer
products by further 30 min incubation in the presence of 0.2 m
M
dNTPs (lane 3) or 0.2 mM dNTPs and 0.5 U Taq DNA polymerase
(lane 4). Neither primer nor extension products were seen when
Rep516 was omitted from the reaction with Taq polymerase (lane
2). (C) DNA primer synthesis and their elongation. The primase
activities of Rep245 and Rep516 proteins were assayed between 5
and 90 °C for 30 min on M13 single-stranded DNA, with dNTPs
including [
32
P]dATP[aP] as substrates. The approximate size of the
bands (in nucleotides) is indicated on the right-hand side of each
panel.
S. Prato et al. Analysis of the pIT3 prim–pol domain
FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS 4395
when ribonucleotides were included in the reaction
mixtures, Rep245 was able to elongate synthetic oligo-
nucleotides, although it showed no preferential use of
any rNTPs in the transferase activity (Fig. 5B,D). The
longer variant Rep516 was also tested for nucleotidyl
transferase activity under identical experimental condi-
tions. As already described for DNA and RNA syn-
thesis, Rep516 proved more efficient than Rep245 in
elongating the 3¢-ends of synthetic oligonucleotides
(data not shown).
Because our enzymatic assays were conducted at
60 °C, a temperature at which hairpin loop-like DNA
structures are likely to be fairly unstable, we were able
to rule out that the elongation products observed had

been produced in a template-directed fashion. More-
over, the evidence that nucleotide addition was not
governed by the sequence of the substrates used for
these assays was further supported by the finding that
Rep245, when incubated with each of the above DNA
oligonucleotides, proved able to incorporate all of the
four (d)NTPs tested.
Discussion
In this study, we describe the structure–function
analysis of a 1–245 N-terminal domain of the puta-
tive replication protein encoded by the pIT3 plasmid
from S. solfataricus, the shortest fully functional
prim–pol domain from a crenarchaeal plasmid identi-
fied and characterized to date. To model the N-ter-
minal domain of the pIT3 replication protein
encompassing residues 31–245 (i.e. Rep245) we used
as a template the resolved crystal structure of the
prim–pol domain of the protein ORF904 from the
pRN1 plasmid of S. islandicus, which had been iden-
tified via both fold recognition and sequence search
against the PDB data bank [13]. In structural terms,
the pIT3 prim–pol domain mainly differs from that
of pRN1 because it has no Zn-stem motif and lacks
two disulfide bonds (one of which is located at the
bottom of the Zn-stem). However, a MD simulation
on the Rep245 model showed that the absence of
the two disulfide bridges did not affect the overall
protein fold. The Zn-binding motif is a structural
feature conserved in all archaeal primase–eukaryotic
primases characterized to date [13,23]. By virtue of

its length and within-domain location, the loop
region of the pIT3 prim–pol domain which replaces
the Zn-stem motif could play a comparable role to
that ascribed to the Zn-stem motif in DNA interac-
tion [24]. A sequence–structure comparison of the
Rep245 model with other archaeal primase–polyme-
rases revealed the conservation of motifs which were
either absent from the pRN1 prim–pol domain or
slightly different from those occurring therein. These
differences may account for the fairly different func-
tions performed by the prim–pol domain of the pIT3
plasmid in vitro, i.e. DNA and RNA synthesis and
3¢-terminal nucleotidyl transferase activity.
Accordingly, we used the modeled pIT3 prim–pol
structure in designing the truncated Rep245 protein
containing the residues predicted to be responsible for
polymerase and primase catalysis, and reported on the
functional characterization of the main functions of
this protein.
C
G
U
A
dC
dG dT
dAAB
CD
28-mer
0312 4 0 312 4
0312 4 0 312 4

28-mer
dC
dG dT
dA
C
GU
A
20-mer
20-mer
Fig. 5. Rep245 has a 3¢-terminal nucleotidyl-transferase activity.
TdT activity was assayed at 60 °Con5¢-end-labeled oligo(dT)
28
(A,
B) and a random 20-mer (C, D) oligonucleotides (see Table 1 for
details of the sequence), as described in Experimental procedures.
Reaction products were separated on 20% polyacrylamide ⁄ urea
gels and radioactivity was detected by autoradiography. Lanes 1–4
of each gel were loaded with reaction mixtures containing only the
indicated (d)NTPs in addition to the DNA template and the protein,
whereas lane 0 contains a control reaction without protein.
Analysis of the pIT3 prim–pol domain S. Prato et al.
4396 FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS
All known DNA polymerases require divalent
cations for catalysis. The main function of the metal
activator is to coordinate incoming nucleoside triphos-
phate substrates with the catalytic site of the DNA
polymerase molecule [17]. Mg
2+
is thought to be the
divalent metal cation employed by most polymerases

for in vivo catalysis [1]. Similarly, the DNA polymerase
activity of Rep245 was found to be dependent on diva-
lent cations, especially Mg
2+
ions which probably act
as physiological metal activators, in a broad optimum
concentration range between 5 and 10 mm. By con-
trast, polymerase activity is stimulated by Mn
2+
ions
at low concentrations (1.0–2.5 mm) and strongly inhib-
ited at higher concentrations. The ability of polymeras-
es to use Mn
2+
instead of Mg
2+
as a required
cofactor is well established [25]. However, the bio-
chemical properties of polymerases are altered as a
result of replacing Mg
2+
with Mn
2+
, which reduces
substrate selection stringency and incorporation fidelity
[26].
Thermal activity analysis of Rep245 revealed an
optimal temperature of 65 °C, i.e.  10 °C lower than
the growth temperature of the natural host S. solfatari-
cus strain IT3 harboring the pIT3 plasmid. Hence,

additional extrinsic factors such as post-translational
modifications, compatible solutes, molecular chaper-
ones and other heat shock factors present in the S. sol-
fataricus cytosol may be involved in protecting the
enzyme against thermal denaturation and guaranteeing
its performance in vivo [27]. Our data clearly show that
DNA polymerase activity of the Rep245 was resistant
to heat treatment. Hence, it is highly unlikely that such
a temperature-stable activity stems from an E. coli-
derived protein present in the enzyme preparation.
Moreover, we carried out a Rep245 mock purification
of an E. coli culture expressing an unrelated protein
and were not able to detect any DNA polymerase or
primase activities.
Bacterial and eukaryotic primases synthesize primers
of defined lengths regardless of template sequence
[1,2]. The typical length of RNA primers produced by
the eukaryotic heterodimeric primase is 6–15 nucleo-
tides [1,28]. It has previously been reported that the
N-terminal (255 residues) prim–pol domain of the pro-
tein ORF904 from the archaeal pRN1 plasmid does
not retain any primase activity, although in this
bifunctional domain the same active site is responsible
for both DNA polymerase and primase activity [13].
By contrast, our study reveals that Rep245 retains its
primase activity, synthesizes primers of  16 nucleo-
tides and is able to incorporate dNTPs for primer
synthesis. The typical length of Rep245-synthesized
DNA primer is 16–20-mer, plus a few 28-mers. DNA
products of defined lengths suggest that Rep245 is

inherently able to count the number of bases
incorporated.
A reasonable structural interpretation of the primase
activity of Rep245 suggests involvement of the K135
and R186 residues, which have counterparts in Pfu-
primase, although not in the pRN1 prim–pol protein.
In archaeal and eukaryotic primases, the K135 residue
(the counterpart of R148 in Pfu-primase) is part of a
highly conserved motif which is absent from the pRN1
prim–pol domain (see alignment in Fig. 1A). The
sequence similarity observed reflects a similar spatial
arrangement, because this motif is part of a b-strand-
loop situated close to the active site in either protein.
Similarly, both the R186 residue in the Rep245 domain
and K300, its counterpart in Pfu-primase, were
contained in a loop that is plausibly involved in DNA
recognition and binding and is positioned left of
the active site [14].
Rep245 is both capable of de novo synthesis of
DNA primers and of elongating them. Long DNA
extension products were observed on the ssDNA tem-
plate when dNTPs were used as substrates, although
primase activity was found to prevail over DNA elon-
gation at higher temperatures. Such reduced DNA
elongation activity might either depend on dissociation
of the Rep245 prim–pol ⁄ ssDNA template complex or
on the fact that Rep245 translocation along the
substrate is probably hindered by the absence of the
additional amino acids needed to stabilize the enzyme–
DNA complex. This explanation seems to be

supported by experimental evidence pointing to
enhanced Rep245 primase activity and better synthesis
product accumulation at higher temperatures. In light
of these observations, we designed a longer variant
comprising the 1–516 N-terminal residues (Rep516)
and investigated its biochemical properties. As we
anticipated, in RNA⁄ DNA synthesis Rep516 proved
more active than Rep245, in that it generated new and
extended DNA and RNA products which were up to
several kb in length.
Hence we suggest that: (a) the additional 271 N-ter-
minal amino acids were necessary to stabilize the grip
of the polymerase on its DNA substrate, and the
enzyme is also able to perform continuous strand
synthesis; or (b) the polymerase activity of Rep245 is
stimulated to a large extent by inclusion of the extra-
portion of the protein in Rep516.
The Rep245 protein typifies the shortest functional
domain among those endowed with primase and poly-
merase activities.
Based on the design of the Rep245 and Rep516
mutants and comparison of their polymerase activities,
S. Prato et al. Analysis of the pIT3 prim–pol domain
FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS 4397
we were able to account for the promiscuous nature of
the synthesis functions performed by the prim–pol
domain and to discriminate between the functions of
in vitro primase and polymerase.
Another finding of our biochemical analysis was that
Rep245 is able to elongate the 3¢-end of DNA mole-

cules in a non-templated manner. To our knowledge,
this is the first evidence that a prim–pol domain
encoded by a crenarchaeal plasmid is intrinsically able
to perform 3¢-terminal nucleotidyl transferase activity.
Similarly, DNA primase from the S. solfataricus
crenarcheon has been shown to synthesize DNA in a
template-independent manner [7,8]. Interestingly, this
property is shared by the X family of human DNA
polymerases, which includes the TdT enzymes and two
additional members, Pol k [29] and Pol l [30]. The
latter two enzymes are functionally malleable to the
point of carrying out various nucleic acid synthesis
reactions on a wide range of substrates [31–33]. Fur-
thermore, like the TdT enzyme [34], the Rep245 protein
can incorporate ribo- and deoxynucleotides in vitro.
A noteworthy finding is that this functional equivalence
is matched by structural relationships between the
catalytic subunit of archaeal primases and the active
site of the X family of polymerases [23]. Indeed, unlike
the pRN1 prim–pol protein whose motif is DXE D,
the Rep245 protein, the X family of DNA polymerases
and the TdT enzymes have the DXD D motif in the
carboxylate triad in common. An additional major
finding reported previously in the literature is a drastic
reduction in enzymatic activity observed when the sec-
ond aspartic residue in the human TDT enzyme motif
is mutated to glutamate [22].
Thanks to the modular architecture of the replication
protein from the pIT3 plasmid, we were able to design
Rep245 and Rep516 truncated proteins and to charac-

terize their multifunction nature, thus demonstrating
that the main activities required for DNA replication
are included in a single-chain polypeptide. This inde-
pendent protein organization suggests a mechanistic
coupling of earlier DNA replication steps such as
primer synthesis and its elongation and, hence, the
autonomy of the plasmid from the host replication
apparatus. This is particularly important for environ-
mental plasmid survival and transfer into new hosts.
The promiscuous nature of the prim–pol domains might
be an atavistic feature evidencing a continuous link
between primase and polymerase activities and the ori-
ginal core replicon of primordial cells. In light of this
suggestion, it seems plausible that prim–pol proteins are
evolutionary precursors acting both as primases and
DNA polymerases, whereas the proteins descended
from them evolved distinct and specific activities.
Within this scenario, the structural and functional simi-
larities between AEP superfamily proteins might be
indicators of this evolutionary interconnection.
Experimental procedures
Materials
PCR grade (d)NTPs were from Roche Applied Science
(Monza, Italy). Radioactive nucleotides [
32
P]dATP[aP]
(3000 CiÆmmol
)1
), [
32

P]ATP[aP] (3000 CiÆmmol
)1
) and
[
32
P]ATP[cP] (3000 CiÆmmol
)1
) were purchased from Per-
kin–Elmer (Waltham, MA, USA). The expression vector
pET-30c(+) was supplied by Novagen (Milan, Italy).
Homology modeling and MD calculations
Sequence search against PDB using psi-blast [35] identified
the crystallographic structure of ORF904 bifunctional
DNA primase–polymerase from the archaeal plasmid
pRN1 at 1.85 A
˚
of resolution (PDB entry 1RNI) [13], as
the best template for Rep245 (32–103, 29% of identity). A
sequence search by fold recognition as implemented in the
FUGUE server [36] also identified the same protein which
was then selected as the best template (Z-score 12.41). To
build the Rep245 model, 16 pairwise and multiple align-
ments between the template and target sequences were
proved, also using modified versions of template structure.
The alignments were carried out with clustal w v. 1.83
[37] and manually edited in order to better align secondary
structure elements of the template with the consensus for
the target sequence deriving from phd and prof secondary
structure prediction programs [38], along with the structural
alignment deriving from FUGUE server. For each align-

ment, modeller v. 6.2 [39] was used to construct 50
homology models (Q31–Q245) and their quality was
assessed by using procheck v. 3.5.4 [40] and the 3D profile
of insightii (Accelrys Software Inc., San Diego, CA,
USA). The best model was completed by addition of all
hydrogen atoms and underwent energy minimization
followed by MD simulation in explicit solvent with the
sander module of the amber 8 package [41], using
PARM99 force field [42].
To perform MD simulation in solvent, the minimized
model was confined in a truncated octahedron box (x, y,
z =80A
˚
) filled with TIP3P water molecules and counteri-
ons (Na
+
) to neutralize the system. The solvated molecule
was then energy minimized through 1000 steps with the
solute atoms restrained to their starting positions using a
force constant of 10 kcalÆmol
)1
ÆA
˚
)1
prior to MD simula-
tion. After this, it was subjected to 90 ps restrained MD
(5 kcalÆmol
)1
ÆA
˚

)1
) at constant volume, gradually heating to
300 K, followed by 60 ps restrained MD (5 kcalÆmol
)1
ÆA
˚
)1
)
at constant pressure to adjust the system density. The
Analysis of the pIT3 prim–pol domain S. Prato et al.
4398 FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS
production MD simulation was carried out at 300 K at
constant pressure for 1.5 ns, with a time-step of 1.5 fs. The
bonds involving hydrogens were constrained using the
shake algorithm [43]. The snapshots were saved every
10 000 steps and analyzed with molmol [44].
Construction of bacterial expression plasmids
Truncated variants of the orf915 gene encoding a putative
replication protein (RepA) were amplified by PCR from
S. solfataricus plasmid pIT3 as a template [18], using
HF Taq DNA polymerase (Roche Applied Science). A
deletion mutant, which contains the N-terminal residues
1–245, was amplified with primers F-245 (5¢-CGGTGCC
GC
CATATGGATAGTTTC-3¢) and R-245 (5¢-CTCGAG
CTGTTCTTTCCT-3). A N-terminal variant comprising
residues 1–516 was obtained with primers F-245 (see above)
and R-516 (5¢-
CTCGAGAGGCTCACGGGC-3¢). The
primers F-245, R-245 and R-516 introduce NdeI and XhoI

restriction sites (underlined), respectively.
The PCR products were purified with QIAquick PCR
purification kit (Qiagen, SpA, Milan, Italy) and cloned in
pGEMTeasy vector (Promega, Milan, Italy). The nucleotide
sequences of both DNA strands of the inserts were verified.
The NdeI–XhoI fragments were cloned into the same sites
of the expression vector pET-30c(+) to obtain the recombi-
nant plasmids pETRep245 and pETRep516, containing
both an in-frame fusion with the six histidine C-terminal
tag.
Expression and purification of the deletion
proteins
Escherichia coli BL21-CodonPlusTM(DE3)-RIL (Strata-
gene, La Jolla, CA, USA) cells carrying pETRep245 were
grown at 37 °C in 1 L of Luria–Bertani medium supple-
mented with kanamycin and chloramphenicol. When the
cultures reached D
600
= 0.6, protein expression was
induced by addition of isopropyl b-d-thiogalactoside at a
final concentration of 1 mm. Cells were grown for an addi-
tional 6 h at 37 °C and then harvested by centrifugation,
suspended in 20 mm sodium phosphate pH 7.4, 500 mm
NaCl, protease inhibitor cocktail (Complete EDTA-free,
Roche) and disrupted by ultrasonication (Sonicator Ultra-
sonic liquid processor; Heat System Ultrasonics Inc., Plain-
view, NY, USA). After ultracentrifugation (160 000 g for
30 min), the crude extract was heated at 65 °C for 20 min,
and the denatured proteins were removed by centrifugation
(15 000 g for 30 min). The heat-resistant fraction was con-

centrated (Amicon, Millipore Corp., Bedford, MA, USA)
and loaded on to a HiTrapÒ affinity column (GE Health-
care Europe GmbH, Milan, Italy) pre-equilibrated in
20 mm sodium phosphate pH 7.4, 500 mm NaCl, 10 mm
imidazole (buffer A). The column was equilibrated with
buffer A containing 20 mm imidazole and the recombinant
proteins were eluted with buffer A supplemented with
250 mm imidazole. The active fractions were pooled and
dialyzed against 20 m m Tris ⁄ HCl pH 9.0 (buffer B). The
dialyzed sample was loaded on to a Resource Q column
(GE Healthcare) developed with a linear gradient of 0–
400 mm NaCl in 20 mm Tris ⁄ HCl pH 9.0. Pooled fractions
were extensively dialyzed against 20 mm sodium phosphate
pH 7.4 and 20% glycerol (buffer C). The mutant Rep516
was purified by the same procedure used for the Rep245.
The purification was monitored after each step by Coomas-
sie Brillant Blue and silver-stained SDS ⁄ PAGE gels. Protein
concentrations were determined by the Bradford assay, and
enzyme stocks (typically 1 mgÆmL
)1
in buffer C) were
stored at )20 °C.
DNA substrates
DNA oligonucleotides were synthesized by MWG-Biotech-
nologies AG (Ebersberg, Germany). When appropriate,
labeling at the 5¢-end was performed using [
32
P]ATP[cP] and
T4 polynucleotide kinase (Roche Applied Science). Unincor-
porated nucleotide was removed with a NickÔ column (GE

Healthcare). The DNA heteropolymeric TP for DNA poly-
merase assays, was prepared by annealing the 5¢-labeled
oligonucleotide to an unlabeled complementary strand in a
1 : 1 molar ratio. The mixture was incubated in a thermo-
cycler with the following temperature profile: 2 min at 95 °C,
5 min at 65 °C, followed by slow cooling at room tempera-
ture. The sequences of the different oligonucleotides used in
this work are listed in Table 1.
DNA polymerase activity
The standard reaction mixture (10 lL) for he DNA poly-
merase assay contained a polymerase buffer (25 m m
Tris ⁄ HCl pH 7.5, 1 mm dithiothreitol, 5 mm MgCl
2
), 5 nm
annealed 5¢-labeled TP used as the substrate, 0.2 mm each
dNTPs, 1–3.4 lm protein of Rep245 protein. The reactions
were allowed to proceed for 10–30 min at 65 °C and
stopped by adding 5 lL of denaturing gel loading buffer
(95% v ⁄ v formamide, 20 mm EDTA, 0.05% bromophenol
blue, 0.05% xylen cyanol). Samples were heated for 3 min
at 95 °C, and then subjected to electrophoresis on a dena-
turing gel (20% polyacrylamide containing 8 m urea). The
radioactive signals were visualized by autoradiography and
quantified using a Molecular Dynamics Bio-Rad Phosphor-
Imager (quantity one software).
DNA polymerase activity was also measured on homo-
polymeric TP (dA
12
⁄ dT
28

). Protein (0.5 lm) was incubated
with TP, 10 lm dNTPs and [
32
P]dATP[aP] in 25 mm
Tris ⁄ HCl (pH 7.5), 1 mm dithiothreitol, 5 mm MgCl
2
in a
10 lL reaction. Aliquots of the reactions were pipetted
onto a DE81 filter; unincorporated dNTPs were removed
by washing with 0.5 m sodium phosphate, pH 7.0, and
filters were counted.
S. Prato et al. Analysis of the pIT3 prim–pol domain
FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS 4399
Determination of pH and divalent ion optima for
polymerase activity
The influence of pH on Rep245 polymerase activity was
determined using the standard polymerase assay in buffer
solutions whose pH was adjusted to the desired value at
65 °C. The buffer systems used (25 mm final concentra-
tion) were as follows: Na-acetate (pH 5.0, 5.4 and 5.8),
Tris ⁄ HCl (pH 6.5, 7.0, 7.5 and 8.0) and glycine ⁄ NaOH
(pH 8.6, 9.0 and 9.6). Reactions were carried out for
30 min at 65 °C and the products resolved by denaturing
gel electrophoresis, as previously described. The amount
of radioactivity in each lane was quantified by Phosphor-
Imager.
The dependence of the Rep245 polymerase activity on
divalent cations was determined using Mg
2+
,Mn

2+
and
Zn
2+
ions at concentrations in the range 0–50 mm. For
both studies, Rep245 protein was used at a concentration
of 3.4 lm.
Thermophilicity and thermostability
Thermophilicity was evaluated in the temperature range
40–90 °C by measuring the polymerase activity of the
Rep245 under standard conditions on TP heteropolymeric
DNA, as a substrate.
Thermal stability was measured by pre-incubating the
Rep245 enzyme (3.4 lm) at 50, 60, 70 and 80 °C. After
15 min, aliquots of the incubated samples were centrifuged
in a microcentrifuge (Eppendorf) at 16 000 g to eliminate
any precipitated material, and assayed at 65 °C under stan-
dard conditions.
Primase activity
RNA primer synthesis was checked by incubating an assay
mix (10 lL) containing polymerase buffer, 100 lm each of
CTP, GTP and UTP, and 10 lm [
32
P]ATP[aP], 0.25 lgof
M13 mp18 ssDNA, 1.7 lm Rep245 or 0.8 lm Rep516 at
60 °C for 30 min. Primer elongation was carried out at 60 °C
upon further 30 min incubation in the presence of 0.2 mm
dNTPs and 0.5 U of Taq DNA polymerase (Promega).
To analyze the effect of temperature on primase activity,
Rep245 and Rep516 proteins were incubated with 0.25 lg

of M13 mp18 ssDNA, 100 lm each of dCTP, dGTP and
dTTP, and 10 lm [
32
P]dATP[aP], in the temperature range
5–90 °C for 30 min. The primase reactions were resolved
onto 10 and 20% denaturing polyacrylamide gel, followed
by autoradiography.
Terminal transferase activity
The 3¢-terminal nucleotidyl-transferase activity was assayed
on 5¢-labeled oligo(dT)
28
and a random 20-mer oligonucleo-
tide (Table 1), with 0.1 mm unlabeled (d)NTPs and 1.7 lm
Rep245 or 0.8 lm Rep516 in the polymerase assay buffer.
Samples were incubated at 60 ° C for 30 min, the aliquots
were analyzed by electrophoresis using denaturing 20% poly-
acrylamide gels and the radioactivity was detected by using a
PhosphorImager. Commercial TdT was used according to
the manufacturer’s protocol (Roche Applied Science).
CD spectrum
CD in the far-UV region was performed with a thermo-
stated Jasco J-815 spectropolarimeter using 0.1 cm path
length quartz cuvettes. The concentration of Rep245 pre-
pared in 10 mm sodium phosphate (pH 7.4) was 5 lm. CD
spectra were recorded at 60, 70 and 80 °C between 190 and
260 nm with a step increase of 0.2 nm, and a bandwidth of
1 nm. Thermal stability of Rep245 was measured by incu-
bating the protein at 60, 70 and 80 °C for 30 min and then
recording the CD spectra of the incubated samples at indi-
cated temperatures.

Acknowledgements
We are grateful to Dr Mariagrazia Squillace for excel-
lent technical assistance and helpful discussion and to
Dr Gabriella Fiorentino, Dr Pietro Amodeo and Dr
Francesca Maria Pisani for critical reading of this
manuscript. This work was grant-aided by the Euro-
pean Union within the framework of the ‘Screen’
projects (contract QLK3-CT-2000-00649), by the
Ministero dell’Universita
`
e della Ricerca Scientifica
(Progetti di Rilevante Interesse Nazionale 2003, U.O.
Simonetta Bartolucci). Support from the Regional
Center of Competence (CRdC ATIBB, Regione Camp-
ania, Naples) is also gratefully acknowledged.
References
1 Kornberg A & Baker TA (1992) DNA Replication.
Freeman, New York.
2 Frick DN & Richardson CC (2001) DNA primases.
Annu Rev Biochem 70, 39–80.
3 Myllykallio H, Lopez P, Lopez-Garcia P, Heilig R,
Saurin W, Zivanovic Y, Philippe H & Forterre P (2000)
Bacterial mode of replication with eukaryotic-like
machinery in a hyperthermophilic archaeon. Science
288, 2212–2215.
4 Kelman Z & White MF (2005) Archaeal DNA replica-
tion and repair. Curr Opin Microbiol 8, 669–676.
5 Dionne I, Robinson NP, McGeoch AT, Marsh VL,
Reddish A & Bell SD (2003) DNA replication in the
hyperthermophilic archaeon Sulfolobus solfataricus.

Biochem Soc Trans 3, 674–676.
Analysis of the pIT3 prim–pol domain S. Prato et al.
4400 FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS
6 Barry ER & Bell SD (2006) DNA replication in the
Archaea. Microbiol Mol Biol Rev 4, 876–887.
7 De Falco M, Fusco A, De Felice M, Rossi M &
Pisani FM (2004) The DNA primase of Sulfolo-
bus solfataricus is activated by substrates containing
a thymine-rich bubble and has a 3¢-terminal nucleo-
tidyl-transferase activity. Nucleic Acids Res 32,
5223–5230.
8 Lao-Sirieix SH & Bell SD (2004) The heterodimeric
primase of the hyperthermophilic archaeon
Sulfolobus solfataricus possesses DNA and RNA prim-
ase, polymerase and 3¢-terminal nucleotidyl transferase
activities. J Mol Biol 344, 1251–1263.
9 Bocquier AA, Liu L, Cann IK, Komori K, Kodha D &
Ishino Y (2001) Archaeal primase: bridging the gap
between RNA and DNA polymerases. Curr Biol 11,
452–456.
10 Liu L, Komori K, Ishino S, Bocquier AA, Cann IK,
Kodha D & Ishino Y (2001) The archaeal DNA
primase: biochemical characterisation of the p41–p46
complex from Pyrococcus furiosus. J Biol Chem 276,
45484–45490.
11 Iyer LM, Koonin EV, Leipe DD & Aravind L (2005)
Origin and evolution of the archaeo-eukaryotic primase
superfamily and related palm-domain proteins: struc-
tural insights and new members. Nucleic Acids Res 33,
3875–3896.

12 Lipps G, Rother S, Hartl C & Krauss GA (2003) A
novel type of replicative enzyme harbouring ATPase,
primase and DNA polymerase activity. EMBO J 22 ,
2516–2535.
13 Lipps G, Weinzierl AO, von Scheven G, Buchen C &
Cramer P (2004) Structure of a bifunctional
DNA primase–polymerase. Nat Struct Mol Biol 11,
157–162.
14 Augustin MA, Huber R & Kaiser JT (2001) Crystal
structure of a DNA-dependent RNA polymerase (DNA
primase). Nat Struct Biol 8, 57–61.
15 Lao-Sirieix SH, Nookala RK, Roversi P, Bell SD &
Pellegrini L (2005) Structure of the heterodimeric core
primase. Nat Struct Mol Biol 12, 1137–1144.
16 Ito N, Matsui I & Matsui E (2007) Molecular basis for
the subunit assembly of the primase from an archaeon
Pyrococcus horikoshii. FEBS J 274, 1340–1351.
17 Steitz TA, Smerdon SJ, Jager J & Joyce CM (1994) A
unified polymerase mechanism for nonhomologous
DNA and RNA polymerases. Science 266, 2022–2025.
18 Prato S, Cannio R, Klenk H-P, Contursi P, Rossi M &
Bartolucci S (2006) pIT3, a cryptic plasmid isolated
from the hyperthermophilic crenarchaeon Sulfolobus
solfataricus IT3. Plasmid 56, 35–45.
19 Delarue M, Pock V, Tordo N, Moras D & Argos P
(1990) An attempt to unify the structure of polymerase.
Protein Eng 3, 461–467.
20 Braithwaite DK & Ito J (1993) Compilation, alignment,
and phylogenetic relationships of DNA polymerases.
Nucleic Acids Res 21, 787–802.

21 Delarue M, Boule
´
JB, Lescar J, Expert-Benzac¸ on N,
Jourdan N, Sukumar N, Rougeon F & Papanicolaou C
(2002) Crystal structures of a template-independent
DNA polymerase: murine terminal deoxynucleotidyl-
transferase. EMBO J 21, 427–439.
22 Yang B, Gathy KN & Coleman MS (1994) Mutational
analysis of residues in the nucleotide binding domain of
human terminal deoxynucleotidyl transferase. J Biol
Chem 269, 11859–11868.
23 Lao-Sirieix SH, Pellegrini L & Bell SD (2005) The pro-
miscuous primase. Trends Genet 21, 568–572.
24 Pan H & Wigley DB (2000) Structure of the zinc-bind-
ing domain of Bacillus stearothermophilus DNA prim-
ase. Structure 3, 231–239.
25 Huang Y, Beaudry A, McSwiggen J & Sousa R (1997)
Determinants of ribose specificity in RNA polymeriza-
tion: effects of Mn
2+
and deoxynucleoside monophos-
phate incorporation into transcripts. Biochemistry 44,
13718–13728.
26 Gabbara S & Peliska JA (1996) Catalytic activities asso-
ciated with retroviral and viral polymerases. Methods
Enzymol 275, 276–310.
27 Trent JD, Gabrielsen M, Jensen B, Neuhard J & Olsen J
(1994) Acquired thermotolerance and heat shock proteins
in thermophiles from the three phylogenetic domains.
J Bacteriol 176, 6148–6152.

28 Arezi B & Kuchta RD (2000) Eukaryotic DNA primase.
Trends Biochem Sci 25, 572–576.
29 Garcia-Diaz M, Dominguez O, Lopez-Fernandez LA,
de Lera LT, Saniger ML, Ruiz JF, Parraga M, Garcı
¨
a-
Ortiz MJ, Kirchhoff T, del Mazo J et al. (2000) DNA
polymerase lambda (Pol lambda), a novel eukaryotic
DNA polymerase with a potential role in meiosis.
J Mol Biol 301, 851–867.
30 Dominguez O, Ruiz JF, Lain de Lera T, Garcia-Diaz M,
Gonzalez MA, Kirchhoff T, Martinez AC, Bernad A &
Blanco L (2000) DNA polymerase mu (Pol mu),
homologous to TdT, could act as a DNA mutator in
eukaryotic cells. EMBO J 19, 1731–1742.
31 Ramadan K, Maga G, Shevelev IV, Villani G, Blanco L
& Hubscher U (2003) Human DNA polymerase lambda
possesses terminal deoxyribonucleotidyl transferase
activity and can elongate RNA primers: implications for
novel functions. J Mol Biol 328, 63–72.
32 Ramadan K, Shevelev IV, Maga G & Hubscher U
(2004) De novo DNA synthesis by human DNA poly-
merase k, DNA polymerase k and terminal deoxyribo-
nucleotidyl transferase. J Mol Biol 339, 395–404.
33 Ruiz JF, Jua
´
rez R, Garcı
´
a-Dı
´

az M, Terrados G,
Picher AJ, Gonza
´
lez-Barrera S, Ferna
´
ndez de
Henestrosa AR & Blanco L (2003) Lack of sugar
S. Prato et al. Analysis of the pIT3 prim–pol domain
FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS 4401
discrimination by human Pol mu requires a single
glycine residue. Nucleic Acids Res 31, 4441–4449.
34 Boule¢ J-B, Rougeon F & Papanicolaou C (2001)
Terminal deoxynucleotidyl transferase indiscriminately
incorporates ribonucleotides and deoxyribonucleotides.
J Biol Chem 276, 31388–31393.
35 Altschul SF, Madden TL, Schaffer AA, Zhang J,
Zhang Z, Miller W & Lipman DJ (1997) Gapped
BLAST and PSI-BLAST: a new generation of protein
database search programs. Nucleic Acids Res 25, 3389–
3402.
36 Shi J, Blundell TL & Mizuguchi K (2001) FUGUE:
sequence–structure homology recognition using
environment-specific substitution tables and structure-
dependent gap penalties. J Mol Biol 310, 243–257.
37 Thompson JD, Higgins DG & Gibson TJ (1994)
CLUSTAL W: improving the sensitivity of progressive
multiple sequence alignment through sequence weight-
ing, position-specific gap penalties and weight matrix
choice. Nucleic Acids Res 22, 4673–4680.
38 Rost B, Yachdav G & Liu J (2003) In the Predict

Protein: the Predict Protein Server. Nucleic Acids Res
32(Web Server issue), W321–W326.
39 Sali A & Blundell TJ (1993) Comparative protein
modeling by satisfaction of spatial restraints. J Mol Biol
234, 779–815.
40 Laskowski RA, MacArthur MW, Moss DS & Thornton
JM (1993) PROCHECK: a program to check the
stereochemical quality of protein structures. J Appl
Crystallogr 26, 283–291.
41 Case DA, Darden TA, Cheatham TE, Simmerling CL,
Wang J, Duke RE, Luo R, Merz KM, Wang B,
Pearlman DA et al. (2004) AMBER 8. University of
California, San Francisco.
42 Wang J, Cieplak P & Kollman PA (2000) How well
does a restrained electrostatic potential (RESP) model
perform in calculating conformational energies of
organic and biological molecules? J Omput Chem 21,
1049–1074.
43 Ryckaert J-P, Ciccotti G & Berendsen HJC (1977)
Numerical integration of the Cartesian equations of
motion of a system with constraints: molecular dynam-
ics of n-alkanes. J Computat Phys 23, 327–341.
44 Koradi R, Billeter M & Wu
¨
thrich K (1996) MOLMOL:
a program for display and analysis of macromolecular
structures. Mol Graphics 14, 51–55.
Supporting information
The following supporting information is available:
Fig. S1. The influence of (A) divalent cations, (B) pH

and (C) temperature on Rep245 polymerase activity.
This supporting information can be found in the
online version of this article.
Please note: Blackwell Publishing are not responsible
for the content or functionality of any supporting
information supplied by the authors. Any queries
(other than missing material) should be directed to the
corresponding author for the article.
Analysis of the pIT3 prim–pol domain S. Prato et al.
4402 FEBS Journal 275 (2008) 4389–4402 ª 2008 The Authors Journal compilation ª 2008 FEBS