Download Physical interaction between proliferating cell nuclear antigen and

GTC_572.fm Page 911 Monday, August 19, 2002 9:23 PM
Physical interaction between proliferating cell nuclear
antigen and replication factor C from Pyrococcus furiosus
S79September
O
Crystal
Matsumiya
riginal
Article
et al.ofLtd
PCNA-PIP
peptide complex
Blackwell
Oxford,
Genes
GTC
©
1356-9597
Blackwell
tostructure
UK
Cells
Science,
2002
Science
Ltd
Shigeki Matsumiya,1 Sonoko Ishino,2 Yoshizumi Ishino2,a,* and Kosuke Morikawa1,*
1
Department of Structural Biology, and 2Department of Molecular Biology, Biomolecular Engineering Research Institute (BERI), 6-2-3,
Furuedai, Suita, Osaka 565-0874, Japan
Abstract
Background: Proliferating cell nuclear antigen
(PCNA), which is recognized as a DNA polymerase
processivity factor, has direct interactions with
various proteins involved in the important genetic
information processes in Eukarya. We determined
the crystal structure of PCNA from the hyperthermophilic archaeon, Pyrococcus furiosus (PfuPCNA) at
2.1 Å resolution, and found that the toroidal ringshaped structure, which consists of homotrimeric
molecules, is highly conserved between the Eukarya
and Archaea.This allowed us to examine its interaction
with the loading factor at the atomic level.
Results: The replication factor C (RFC) is known as
the loading factor of PCNA on to the DNA strand.
P. furiosus RFC (PfuRFC) has a PCNA binding motif
(PIP-box) at the C-terminus of the large subunit
(RFCL). An 11 residue-peptide containing a PIP-box
sequence of RFCL inhibited the PCNA-dependent
primer extension ability of P. furiosus PolI in a
Introduction
DNA replication is a central process of the cell cycle that
is essential for the maintenance of life. Proteins with
common roles in the DNA replication mechanism have
been extensively characterized in Bacteria and Eukarya,
and they are mostly well conserved, despite differences
in their amino acid sequences (reviewed in Kornberg &
Baker 1992; Stillman 1994; Kelman & O’Donnell 1995;
Communicated by: Fumio Hanaoka
*Corrrespondence: E-mail: [email protected],
[email protected]
a
Present address: Laboratory of Protein Chemistry and
Engineering, Department of Genetic Resources Technology,
Faculty of Agriculture, Kyushu University, 6-10-1 Hakozaki,
Higashi-ku, Fukuoka-shi, 812-8581, Japan.
© Blackwell Science Limited
concentration-dependent manner. To understand
the molecular interaction mechanism of PCNA
with PCNA binding proteins, we solved the crystal
structure of PfuPCNA complexed with the PIP-box
peptide. The interaction mode of the two molecules
is remarkably similar to that of human PCNA and a
peptide containing the PIP-box of p21WAF1/CIP1. Moreover, the PIP-box binding may have some effect
on the stability of the ring structure of PfuPCNA
by some domain shift.
Conclusions: Our structural analysis on PfuPCNA
suggests that the interaction mode of the PIP-box
with PCNA is generally conserved among the
PCNA interacting proteins and that the functional
meaning of the interaction via the PIP-box possibly
depends on each protein. A movement of the Cterminal region of the PCNA monomer by PIP-box
binding may cause the PCNA ring to be more rigid,
suitable for its functions.
Waga & Stillman 1998).The replicative DNA polymerases
require an elongation factor called the ‘sliding clamp’
for processive DNA synthesis. The eukaryotic proliferating cell nuclear antigen (PCNA), the bacterial DNA
polymerase III β-subunit, and the bacteriophage T4
gene 45 protein (gp45) are known to be sliding clamps
for their respective DNA polymerases. The amino acid
sequences of these sliding clamps are quite different from
one another. However, the crystal structures of the yeast
and human PCNAs (Krishna et al. 1994; Gulbis et al.
1996), the E. coli β-subunit (Kong et al. 1992), and the
gp45 proteins of the T4 and RB69 bacteriophages
(Shamoo & Steitz 1999; Moarefi et al. 2000) have a
common ring-shaped structure with a pseudo six-fold
symmetry. The homotrimer of PCNA and gp45, and
the homodimer of the β-subunit molecules encircle the
double-stranded DNA (dsDNA) strand in its central
Genes to Cells (2002) 7, 911– 922
911
GTC_572.fm Page 912 Monday, August 19, 2002 9:23 PM
S Matsumiya et al.
cavity in a topological manner, and can freely slide along
dsDNA. The sliding clamps interact directly with the
replicative DNA polymerases (Pol δ and Pol ε for
PCNA, Pol III for the β-subunit, and gp43 (Pol) for
gp45) and enhance their processive DNA synthesis.
Recent sequence analyses of the complete genomes
from several organisms in Archaea, the third domain of
life, confirmed that the proteins relating to the genetic
information system, including DNA replication, are
structurally more similar to eukaryotic proteins than
those from Bacteria (reviewed in Edgell & Doolittle
1997; Ishino & Cann 1998; Cann & Ishino 1999; Leipe
et al. 1999). We cloned a gene encoding a sequence
which was homologous to eukaryotic PCNA from
the hyperthermophilic euryarchaeote, Pyrococcus furiosus,
expressed it in Escherichia coli, and characterized the
purified gene product (Cann et al. 1999). The protein
interacted with both DNA polymerases I (Pol BI) and
II (Pol D) in this organism and enhanced their DNA
synthesizing activities; therefore, we designated it as
PfuPCNA. Characterizations of the PCNA homologues
in other archaeal organisms, including Sulforobus solfataricus
(a crenarchaeote) and Methanothermobacter thermoautotrophicus
(an euryarchaeote), have also been reported, and they
also stimulate DNA polymerization (De Felice et al.
1999; Kelman & Hurwitz 2000). These results suggest
that the basic role of the sliding clamps in processive
DNA synthesis is conserved across the three biological
domains.
We crystallized the PfuPCNA protein and solved its
three-dimensional structure. It turned out that the ringshaped structure with the pseudo-sixfold symmetry of
the sliding clamps is clearly conserved, and that the
archaeal and eukaryotic PCNAs are particularly similar
to each other (Matsumiya et al. 2001). In addition, we
demonstrated that PfuPCNA interacts functionally
with mammalian replication proteins, and showed that
PfuPCNA stimulated calf thymus DNA polymerase δ
activity. Moreover, in the case of a circular DNA
template, human Replication Factor C (RFC) enhanced
the PfuPCNA-dependent activity of Pol δ (Ishino et al.
2001).The functional interaction of Thermococcus fumicolans
PCNA with calf thymus Pol δ has also been reported
(Henneke et al. 2000).These findings further support the
idea that the structure-function analyses of the archaeal
replication proteins greatly contribute to the understanding of the molecular recognition mechanisms
employed within human cells.
In addition to the role of the sliding clamp for DNA
polymerases, PCNA interacts with many proteins
involved in important cell cycle processes, including
DNA repair and apoptotic pathways under cell cycle
912
Genes to Cells (2002) 7, 911– 922
control (reviewed in Jonsson & Hubscher 1997; Kelman
1997; Kelman & Hurwitz 1998; Tsurimoto 1998; Warbrick
1998; Tsurimoto 1999; Warbrick 2000). Therefore,
studies on the interaction mechanisms between PCNA
and other proteins are very important to understand
the molecular mechanisms of the sequential reactions
under precise cell cycle control. Interestingly, a conserved
sequence motif, Qxx(L/I/M)xxF(F/Y) (x = any residue),
is found in many PCNA-interacting proteins, and the
motif is now called the PCNA interacting protein box
(PIP-box) (Warbrick et al. 1998).The crystal structure of
a short peptide containing the PIP-box of the p21WAF1/CIP1
protein, a cyclin-dependent kinase related protein,
complexed with human PCNA, confirmed that the PIPbox interacts with a hydrophobic pocket on the PCNA
surface (Gulbis et al. 1996). The RFC works for loading
of the PCNA on to the DNA template-primer, and the
direct interaction between RFC and PCNA is necessary
for this function (reviewed in Mossi & Hubscher 1998).
We demonstrated that P. furiosus RFC enhanced the
PfuPCNA-dependent DNA synthesis by P. furiosus Pol I
and Pol II (Cann et al. 2001).The biochemical characterization of two more archaeal RFCs (from Methanothermobacter thermoautotrophicus and Sulfolobus solfataricus)
have also been reported (Kelman & Hurwitz 2000;
Pisani et al. 2000). All of the archaeal RFCs consist of
two different proteins (large and small subunits). In the
large subunit (RFCL) of PfuRFC, a candidate PIP-box
sequence is found at residues 470– 477 of the 479 residues
in the protein. In this study, we determined the crystal
structure of PfuPCNA complexed with a peptide
containing the PIP-box sequence of the RFCL, and
analysed the interaction at the atomic level.
Results
The PIP-box of RFCL specifically binds to
PfuPCNA
An in vitro pull-down assay in our previous study showed
that PfuPCNA and RFCL have a physical interaction
with each other (Cann et al. 1999).We found a PIP-box-like
sequence in RFCL, but not in RFCS, of the archaeal
RFC complex (Fig. 1).Therefore, the PIP-box may make
an important contribution to the direct interaction
between RFC and PCNA for the processive DNA
synthesis. To investigate this prediction, a peptide corresponding to the C-terminal 11 amino acids of P. furiosus
RFCL,
Acetyl-Lys-Gln-Ala-Thr-Leu-Phe-Asp-PheLeu-Lys-Lys, was synthesized. When this peptide was
added to an in vitro DNA synthesis reaction with
P. furiosus Pol I, PCNA and RFC, under the conditions
© Blackwell Science Limited
GTC_572.fm Page 913 Monday, August 19, 2002 9:23 PM
Crystal structure of PCNA-PIP peptide complex
Figure 1 PIP-box sequences conserved in the archaeal RFCL.
The PIP-box-like sequence motifs found in the C-terminal
region of the RFC large subunit from four archaeal genus are
aligned. The consensus residues are given in bold. ‘h’ and ‘a’
represent moderately hydrophobic residues (L, I, M) and highly
hydrophobic residues with aromatic side chains (F, Y), respectively.
‘x’ indicates any residue.
described in the Experimental procedures, the strand
synthesis became obviously inhibited with increasing
amounts of the peptide added to the reaction mixtures
(Fig. 2a). To show the effect of PfuPCNA on P. furiosus
Pol I activity most clearly, reactions were carried out in
the presence of 100 mm NaCl, in which the activity of
Pol I by itself was inhibited as shown previously (Cann
et al. 2001; Ishino et al. 2001). Furthermore, a priming
site, from which the DNA synthesis is paused at about
0.7 kb by the second secondary of the template DNA,
was chosen to show the effect of RFC on the extension
reaction more clearly. P. furiosus Pol I also has a PIP-box-like
sequence in its C-terminal region (763-QVGLTSWL-770),
as shown previously (Cann et al. 1999), the RFCL peptide
may compete the PIP-box binding site in PfuPCNA with
Pol I. As we found previously (Cann et al. 1999, 2001),
PfuPCNA can load on to the circular DNA without
any help from a clamp-loader, in contrast to the eukaryotic
PCNA, although the loading efficiency is drastically
enhanced by the addition of PfuRFC (Fig. 2, lanes 2, 3).
This self-loading activity of PfuPCNA can be explained
by the smaller number of intersubunit hydrogen bonds
in PfuPCNA, as compared with those in the yeast and
human PCNA crystal structures, as described earlier
(Matsumiya et al. 2001). Therefore, the PfuPCNAdependent DNA synthesis reaction of PolI without RFC
were compared in the presence and absence of the PIPbox peptide of RFCL. As shown in Fig. 2b, the PIP-box
peptide clearly inhibited strand synthesis with increasing
amounts of the peptide added to the reaction mixture.
On the other hand, a peptide having a sequence unrelated
to the PIP-box affected no inhibition at all. In this case,
a priming site different from that in Fig. 2a, from which
DNA synthesis proceeds more smoothly with less
pausing, was used to show the effect of PCNA clearly.
This result indicates that the RFCL PIP-box peptide
binds to the appropriate position on PfuPCNA and
competes with PolI for the binding position. Greater
amounts of the RFC peptide were needed for the
inhibition of DNA synthesis in Fig. 2b as compared with
in Fig. 2a. This observation suggests that the PIP-box
Figure 2 Inhibition of the PCNAdependent DNA synthesis activity of PolI
by the PIP-box peptide. (a) The PIP-box
peptide was added in increasing amounts
(20, 50, 200, 500 and 1000 pmol) to the
primer extension reaction containing
1 pmol of 32P-labelled primer, 0.2 µg of
M13 mp18ssDNA, 0.5 U of PolI, 2 pmol of
PfuPCNA (trimer), and 2 pmol of PfuRFC,
and the reaction mixtures were incubated
at 70 °C for 4 min. The reaction products
were analysed by alkaline agarose gel electrophoresis followed by autoradiography.
(b) The primer extension reaction, as shown
in (a), was done without PfuRFC. The
PIP-box peptide and the control peptide
were each added in increasing amounts (50,
500, 2500 and 5000 pmol). The synthesized
products were analysed by alkaline agarose
gel electrophoresis. Size markers were
prepared by labelling BstPI-digested λDNA
with 32P-ATP.
© Blackwell Science Limited
Genes to Cells (2002) 7, 911– 922
913
GTC_572.fm Page 914 Monday, August 19, 2002 9:23 PM
S Matsumiya et al.
of PolI has an affinity to PfuPCNA which is greater than
that of RFCL.
Structure determination of the PfuPCNA-RFCL
peptide complex
The complex crystals of PfuPCNA with the C-terminal
11-mer peptide of RFCL were obtained under conditions similar to those used for the uncomplexed
PfuPCNA crystals (Matsumiya et al. 2001).The results of
the data processing and the structure analysis are summarized in Table 1. Both the uncomplexed PCNA and the
PCNA-peptide complex were crystallizd in an identical
space group (P63) and with similar unit cell constants
(a = 89.682, c = 63.269 Å for the uncomplexed PCNA
and a = 91.847, c = 64.144 Å for the PCNA-peptide
complex).The molecular packing of PCNA in the complex crystal is nearly identical to that of the uncomplexed
PCNA crystal. The toroidal structure of the PCNA
trimer is retained in the complex crystal (Fig. 3). One
C-terminal peptide of RFCL is bound on every PCNA
molecule in the trimer complex, as in the case of the
human PCNA-p21WAF1/CIP1 C-terminal peptide complex
(Gulbis et al. 1996).
Table 1 Summary of the crystal structure analysis
Data collection
Space group
Unit cell (Å)
Total reflections (100.00 –2.30 Å)
Unique reflections (100.00 –2.30 Å)
Completeness (100.00 –2.30 Å) (%)
Completeness (2.38 –2.30 Å) (%)
Average redundancy (100.00 –2.00 Å)
Average redundancy (2.38 –2.30 Å)
Average I/s(I) (100.00 –2.30 Å)
Average I/s(I) (2.38 –2.30 Å)
R meas (100.00 –2.30 Å)
R meas (2.38 –2.30 Å)
P63
a = 91.847
c = 64.144
60 038
13 867
99.9
99.1
4.4
3.8
19.2
2.2
0.078
0.345
Structure refinement (50.00 –2.30 Å)
Reflections used for refinement
Reflections used for cross validation
R
R free
Rmsd of bond lengths (Å)
Rmsd of bond angles (°)
Rmsd of dihedral angles (°)
Rmsd of improper angles (°)
Average B factor (Å2 )
914
Genes to Cells (2002) 7, 911– 922
12 451
1342
0.2348
0.2913
0.007
1.3
25.7
0.74
38.2
Coarse view of the peptide
The PCNA ring has two distinct faces, and the RFCL
peptide is located on the front face containing the Cterminal tail (Fig. 3). The N-terminus of the peptide
(acetyl-Lys469 –Ala471) is connected with the C-terminus
of PfuPCNA (Ala244 –Val247) through hydrogen bonds
(Fig. 4).A short helix at Leu473–Asp475 is placed on the
hydrophobic pocket formed by Met41, Arg45– Leu48
and Leu242, and Phe476 is surrounded by Glu224 –
Pro226 and Pro245. In this crystal, the last three residues,
Leu477–Lys479 at the C-terminus of the RFCL peptide,
could not be located because of structural disorder.
Structural definition of the PIP-box
In the peptide derived from the PIP-box of RFCL, the
side chain of Gln470 is connected to Gly200 and the
Ala244–Arg246 of PfuPCNA through direct or indirect
(via water) hydrogen bonds (Fig. 4). Gln470 is so deeply
involved in the intermolecular interaction that the
replacement of this glutamine with any other residue
will be impossible.The short helical structure at Leu473–
Asp475 directs the side chains of Leu473 and Phe476
towards the hydrophobic pocket of PfuPCNA, and
therefore, these hydrophobic side chains are important
for these residues at positions 473 and 476.
Three groups of residues participating in the PIP-box
recognition have been identified from mutation studies
of human PCNA: the centre loop between βC1 and
βD1, the interdomain connecting loop (ID-loop)
between βI1 and βA2, and the C-terminal tail after βI2
(Tsurimoto 1998, 1999). These findings were consistent
with the crystal structures of the human PCNAp21WAF1/CIP1 C-terminal peptide complex (Gulbis et al.
1996) and the bacteriophage RB69 gp45-gp43 Cterminal peptide complex (Shamoo & Steitz 1999). In
these complexes, the C-termini of the peptides were
extended into the ID-loops of the PCNA and gp45
proteins, respectively (Fig. 5). In the PfuPCNA-RFCL
peptide complex, the residues in the centre loop and
the C-terminus of PCNA connect to the peptide
through hydrogen bond. However, the ID-loop is not
fixed along with the C-terminus of RFCL. Both of
p21CIP1/WAF1 and DNA polymerase δ, but not RFC,
require the ID-loop of PCNA for their interactions in
eukaryotes, and therefore, the failure in modelling the
ID-loop of PCNA and the C-terminus of the RFCL
peptide into the PfuPCNA-RFCL peptide complex
crystal suggests that the intermolecular interactions
between these sites are not important for the binding
of PfuRFC on to PfuPCNA.
© Blackwell Science Limited
GTC_572.fm Page 915 Monday, August 19, 2002 9:23 PM
Crystal structure of PCNA-PIP peptide complex
Figure 3 Overall structure of the trimer
of the PfuPCNA/PfuRFCL C-terminal
peptide complex. Each molecule is indicated with the symmetry operator and drawn
in a different colour.
Figure 4 Schematic representation of the intermolecular interactions observed in the PfuPCNA/PfuRFCL C-terminal peptide
complex. The residues corresponding to the C-terminal region of RFCL (469–479) are indicated in yellow. The interacting residues in
PfuPCNA are shown in pink for C-terminal tail, green for the centre loop, blue for hydrophobic pocket, and orange for not previously
classified. The intermolecular hydrogen bonds (N···O or O···O distance ≤ 3.5 Å) are shown by dashed lines.
Movement of the C-terminal domain of PfuPCNA
by peptide binding
The three-dimensional structural alignment of PfuPCNA
and the PfuPCNA-RFCL peptide complex in the trimer
form reveals that the C-terminal domain moves signifi© Blackwell Science Limited
cantly upon binding of the RFCL peptide, while the Nterminal domain remained unaffected. The C-terminal
domain of the complex swayed about 10 degrees backward
at the front side and slid to the inner side of the ring at the
opposite face (rear side) (Fig. 6).The structural framework
of the PCNA C-terminal domain is retained in the
Genes to Cells (2002) 7, 911– 922
915
GTC_572.fm Page 916 Monday, August 19, 2002 9:23 PM
S Matsumiya et al.
Figure 5 Comparison of the three-dimensional structures of PCNA-PIP-box peptide complexes. The PfuRFCL peptide is blue,
the human p21 peptide is green, the bacteriophage RB69 gp43 peptide is red, the PfuPCNA is cyan, the human PCNA is yellow, and
the RB69 gp45 is magenta.The N- and C-termini are marked by the letters N and C (for the sliding clamps) or N′ and C′ (for peptides)
with subscripts indicating the species (p for P. furiosus, h for human and r for RB69).
Figure 6 Structural comparison between complexed (bold lines) and uncomplexed (thin lines) PfuPCNA molecules in the trimeric state.
For clarity, only one PCNA-peptide complex is shown. In the complexed form, the N-terminal domain, C-terminal domain, and RFCL
peptide are indicated in blue, green and red, respectively.
916
Genes to Cells (2002) 7, 911– 922
© Blackwell Science Limited
GTC_572.fm Page 917 Monday, August 19, 2002 9:23 PM
Crystal structure of PCNA-PIP peptide complex
complex crystal, despite the large movement caused by
complex formation.The intermolecular interface at the
outer surface of the PfuPCNA trimeric ring is composed
of the anti-parallel β-strands, βI1 and βD2. Although the
C-terminal peptide of RFCL does not interact directly with
βI1 or βD2, the movement of the C-terminal domain of
PCNA is perturbed at the intermolecular interaction.
Four intermolecular main chain hydrogen bonds (N···O
distance ≤ 3.3 Å) are observed in the uncomplexed
PfuPCNA crystal, suggesting a weaker intermolecular
interaction in PfuPCNA, as compared with the yeast
and human PCNAs, which have seven and eight hydrogen bonds at the intermolecular interface, respectively
(Matsumiya et al. 2001). In the crystal of the PfuPCNA–
RFCL peptide complex, the end of the C-terminal
domain moves towards the centre of the ring, and
three more hydrogen bonds (Thr106O···Gly185N,
Thr108N···Lys178O and Thr108O···Lys178N) were
emerged, in addition to the four that were observed
in the uncomplexed PfuPCNA (Thr110N···Glu176O,
Thr110O···Glu176N,
Arg112N···Glu174O
and
Arg112O···Glu174N) (Fig. 7a,c).
The movement of the C-terminal domain also
rearranges intermolecular side chain ion pairs. In the
case of the uncomplexed PfuPCNA, seven ion pairs
involving five residues (Arg82, Lys84, Arg109, Asp143
and Asp147) exist on the inner side of the PCNA ring,
and three pairs between Arg112 and Glu174 were
observed on the outer side. In the PfuPCNA-RFCL
peptide complex, the ion pairs were rearranged. Six ion
pairs involving six residues (Arg82, Lys84, Arg109,
Glu139, Asp143 and Asp147) are formed on the inner
side of the ring, although the ion pairs on the outer side
disappeared (Fig. 7b,d, and Table 2). The residues which
were newly involved in the intermolecular hydrogen
bonds and the ion pairs are located on the rear side of
the PCNA ring, and the ion pairs observed only in the
uncomplexed PCNA crystal are located on the front
side of the ring. Although the number of intermolecular
ion pairs decreased upon complex formation, the ion pair
network on the inner side of the ring was more extended
in the complex, as compared with the uncomplexed
form. Therefore, the ring structure of PfuPCNA may be
stabilized by PIP-box binding.
Discussion
We determined the crystal structure of P. furiosus PCNA
complexed with a peptide containing the PIP-box,
derived from the P. furiosus RFC large subunit. In the
complex crystal, the interaction mode of the PIP-box
with the PCNA was highly conserved, as observed in the
© Blackwell Science Limited
p21WAF1/CIP1 PIP-box complexed with human PCNA
and the RB69 sliding clamp complexed with the
PIP-box of RB69 DNA polymerase (Gulbis et al. 1996;
Shamoo & Steitz 1999).
The PIP-box of PfuRFC is located at the C-terminus
of RFCL.The sequence alignment of the RFC subunits
suggests that RFCL has a core structure similar to the
three domains found in the crystal structure of RFCS
(Oyama et al. 2001). The PIP-box is connected to the
core by a long chain (of ≈ 70 residues) composed mainly
of Glu and Lys. To examine the role of the PIP-box in
RFCL for the clamp loading function of PfuRFC, we
made a deletion mutant protein of RFCL that lacks the
C-terminal 12 amino acids The mutant was combined
with RFCS, and the RFC complex (PfuRFC∆12) was
compared with the wild-type PfuRFC for stimulation
of the PfuPCNA-dependent DNA synthesis activity
of Pol I. However, no critical difference was detected
between the wild-type and the truncated RFC proteins
for stimulation of DNA synthesis. This result was in
contrast with the case of PolI, which completely lost the
PfuPCNA-dependency on its DNA synthesis activity by
the truncation of C-terminal 30 residues (PolI ∆1 mutant
as published in Komori & Ishino 2000) containing a PIPbox (data not shown). These observations suggest that
the PIP-box of RFCL is not essential for the PfuRFC
to interact with PfuPCNA, at least in stimulations of
the in vitro primer extension. It is known that not all
the known PCNA-binding proteins contain PIP-box
sequences, and no PIP-box-like sequence is found
in the eukaryotic RFC proteins. Other motifs, such
as a replication factory targeting sequence (RFTS)
(Montecucco et al. 1998) and KA-box (Xu et al. 2001),
have also been proposed as the interacting sequences
in the PCNA-binding proteins. Further analyses are
required to understand the interaction mechanism of
PfuPCNA and PfuRFC, including the physiological function of the PIP-box in the archaeal RFCL. In E. coli, the
δ subunit in the clamp loader complex utilizes the
edge of a helix within the N-terminal domain to interact
with the hydrophobic pocket of the β subunit (sliding
clamp) and to open the ring ( Jeruzalmi et al. 2001). The
RFCS protein of P. furiosus shares a similar three-domain
structure with E. coli δ (Oyama et al. 2001; Jeruzalmi
et al. 2001). It may be assumed that the PfuRFC uses
the RFCL PIP-box as an anchor to one subunit in the
trimeric PCNA ring, and when the ring needs to be
opened, the core RFC complex approaches other PCNA
subunits for the main interaction. Although the anchor is
not essential for the function of PfuRFC on PfuPCNA,
attachment of PfuRFC to the DNA polymerase–PCNA
complex during DNA strand synthesis would provide
Genes to Cells (2002) 7, 911– 922
917
GTC_572.fm Page 918 Monday, August 19, 2002 9:23 PM
S Matsumiya et al.
Figure 7 Hydrogen bonds and ion pairs observed in the PfuPCNA crystals. Hydrogen bond (N···O ≤ 3.3 Å) and ion pairs (N···O ≤
4.0 Å) observed in the PfuPCNA trimer, with and without the PIP-box peptide are shown as red dashed lines and blue dashed lines,
respectively, in each figure. (a) Outside view of the PCNA–peptide complex. (b) Inside view of the PCNA–peptide complex. (c) Outside
view of the uncomplexed PCNA. (d) Inside view of the uncomplexed PCNA.
918
Genes to Cells (2002) 7, 911– 922
© Blackwell Science Limited
GTC_572.fm Page 919 Monday, August 19, 2002 9:23 PM
Crystal structure of PCNA-PIP peptide complex
Figure 7 Continued
© Blackwell Science Limited
Genes to Cells (2002) 7, 911– 922
919
GTC_572.fm Page 920 Monday, August 19, 2002 9:23 PM
S Matsumiya et al.
Table 2 List of intermolecular ion pairs (d ≤ 4.0 Å)
Ionic bonded atoms
PfuPCNA
inner side
Arg82 Nη1...Asp147 Oδ1
Arg82 Nη1...Asp143 Oδ1
Lys84 Nζ...Asp143 Oδ1
Lys84 Nζ...Asp143 Oδ2
Arg109 Nη2...Asp147 Oδ1
Arg109 Nη2...Asp147 Oδ2
Arg109 Nη1...Asp143 Oδ2
outer side
Arg112 Nη1...Glu174 Oε1
Arg112 Nη1...Glu174 Oε2
Arg112 Nη2...Glu174 Oε1
PfuPCNA/RFCL peptide
inner side
Arg82 Nη2...Asp147 Oδ2
Lys84 Nζ...Asp143 Oδ1
Lys84 Nζ...Asp143 Oδ2
Lys84 Nζ...Glu139 Oε2
Arg109 Nη2...Asp147 Oδ1
Arg109 Nη1...Asp143 Oδ1
Å
3.04
3.81
3.23
3.49
3.28
3.78
3.91
3.74
3.73
3.93
2.88
2.48
3.22
3.50
3.99
3.81
very efficient unloading and reloading of the clamp. An
anchor of the clamp loader to the DNA polymerase III
complex and DnaB helicase using the carboxy-terminal
region of the τ subunit has been demonstrated in E. coli
replisome (Studwell-Vaughan & O’Donnell 1991; Kim
et al. 1996; Yuzhakov et al. 1996).
A structural comparison of the PfuPCNA alone and
PfuPCNA-PIP-box peptide complex revealed the
movement of the C-terminal region upon peptide
binding. The movement causes two intermolecular
main chain hydrogen bonds between βI1 and βD2
(The108N···Lys178O and Thr108O···Lys178N) and
one between adjacent loops (Thr106O···Gly185N). The
hydrogen bond equivalent to the Thr106O···Gly185N
of PfuPCNA-RFCL peptide complex (Glu109O···
Ser183N) is observed in the crystal structure of the human
PCNA-p21 peptide complex (Gulbis et al. 1996), but
not in yeast PCNA without a peptide (Krishna et al.
1994). Since the three-dimensional structure of human
PCNA without a PIP-box peptide is not yet available
yet, it is not clear whether the additional intermolecular
hydrogen bond results from the movement of the Cterminal domain induced by peptide binding. However,
if the interaction mode of the PCNA-PIP-box is generally conserved among PCNA binding proteins, then it
can be assumed that the binding of the PIP-box peptides
stabilizes the toroidal structure of the PCNA trimer by
920
Genes to Cells (2002) 7, 911– 922
increasing the number of hydrogen bonds at the intermolecular interface. Inhibition of the DNA synthesis
reaction by the PIP-box peptide as shown in Fig. 2 may
be caused by binding competition between the PIP-box
peptide and PolI to PfuPCNA, as described above.
However, if the PIP-box peptide binding stabilizes the
PfuPCNA structure, it would also be possible that the
peptide binding inhibits self-loading of PfuPCNA on to
DNA, and subsequently blocks the processive DNA
synthesis.To clarify whether the binding of the PIP-box
functionally stabilizes the PfuPCNA ring structure,
chemical cross-linking and gel-filtration analyses were
carried out. However, to date no difference in the trimerformation efficiency of PfuPCNA in the presence or
absence of the RFCL PIP-box peptide has been observed
from these experiments (data not shown).
It has not yet been confirmed that the domain-shift
caused by PIP-box binding contributes to the physiological function of PCNA at this stage. It would be reasonable to assume that stabilization of the PCNA ring by
a physical interaction with other proteins will provide
some advantages for PCNA function. If the PCNA ring
is stabilized by binding of the PIP-box in the DNA
polymerase, the rigid ring structure of PCNA complexed
with the polymerase will be more suitable for the
processive DNA synthesis. In the case of RFC, stabilization
of the PCNA ring structure is disadvantageous for the
clamp-loading and unloading. The PIP-box of PfuRFC
may work just for anchoring PfuPCNA during the
DNA synthesis and some different interactions may
work positively for the ring-opening of PfuPCNA, as
discussed above. Direct binding analyses of the PIP-box
of PolI and PfuPCNA with DNA strands and also
elucidation of the clamp-loading mechanism should be
done to clarify this issue.
In conclusion, our structural study on PfuPCNAPIP-box peptide revealed that the interaction mechanism
of PCNA and the PIP-box sequences are well conserved
among the PCNA binding proteins and the PCNA ring
structure may become more rigid by PIP-box binding.
Further studies are important to understand how the
structural change of PCNA will contribute to its
function in the cells.
Experimental procedures
In vitro primer extension using P. furiosus PolI with
PfuPCNA and PfuRFC
P. furiosus PolI, PCNA(M73L), and RFC were prepared as
described earlier (Komori & Ishino 2000; Cann et al. 2001;
Matsumiya et al. 2001). There was no difference between the
© Blackwell Science Limited
GTC_572.fm Page 921 Monday, August 19, 2002 9:23 PM
Crystal structure of PCNA-PIP peptide complex
wild-type PfuPCNA and PCNA(M73L) for stimulating PolI
activity, and therefore, the crystal analysis was carried out using
PCNA(M73L) (Matsumiya et al. 2001). Hereafter, PCNA(M73L)
will be referred to as PfuPCNA. An in vitro primer extension
reaction using M13 mp18 single-stranded DNA annealed with
a 32P-labelled primer was carried out under previously described
conditions (Cann et al. 2001).The synthetic C-terminal peptide of
PfuRFCL (acetyl-Lys-Gln-Ala-Thr-Leu-Phe-Asp-Phe-Leu-LysLys) was obtained from Dr T.Tanaka (Biomol. Eng. Res. Institute).
Functioning as a peptide with a sequence unrelated to the
PIP-box, Arg-Arg-Leu-Ile-Glu-Asp-Ala-Glu-Tyr-Ala-Ala-Arg-Gly
(obtained from Peptide Institute Inc., Osaka, Japan) was used for
the competition experiment.
Crystal structure analysis
A solution, containing 0.4 mm of the protein and 1.8 mm of the
peptide, was used for the hanging drop vapour diffusion crystallization experiment.The drop (1.0 µL of the protein solution and
1.0 µL of the precipitant solution) was equilibrated against 500 µL
of the precipitant solution. Colourless hexagonal co-crystals
of 0.15 × 0.15 × 0.15 mm were obtained using the precipitant
solution containing 100 mm sodium citrate, pH 5.5, 2.4 m ammonium sulphate, and 10% (v/v) glycerol. These crystals could
be flash-cooled under a nitrogen stream without cryoprotection.
Diffraction data were collected at 104 K on the BL-6B beamline
at the Photon Factory, using 1.0000 Å radiation and a Weissenberg
camera imaging plate (Sakabe et al. 1997).The data were processed
using HKL (Otwinowski & Minor 1997), and 13 793 reflections
within the resolution range of 2.3–50.0 Å were used for the
structure analysis.
Structure determination
The initial crystal structure of the PfuPCNA(M73L)-PfuRFCL
C-terminal peptide complex was determined by the molecular
replacement method, using the CNS program suite (Brünger et al.
1998) with uncomplexed PfuPCNA(M73L) (Matsumiya et al.
2001) (PDB code 1GE8) as the search model. The molecular
model of the RFCL peptide was built from the difference Fourier
map using O ( Jones et al. 1991), and the overall structure of the
complex was refined using CNS. Met1, Glu120–Met125, and
Glu248–Glu249 for PCNA, and Leu477–Lys479 for the RFCL
peptide were not modelled, because of the poor electron density
at these regions. The coordinates and structural factors have been
deposited in the Protein Data Bank under the accession code
1ISQ.
Preparation of a truncated mutant RFC
∆C12)
(PfuRFC∆
To delete the region corresponding to the C-terminal 12 amino
acids from the rfcL gene, PCR was done using pTRFLhis (rfcL
is inserted into pET28a′) as described earlier (Cann et al. 2001).
The amplified gene fragment was inserted into the NdeI-XhoI sites
of pET15b, and the resultant plasmid was designated as pRFL∆12
© Blackwell Science Limited
after conformation of the nucleotide sequence. The expression of
the gene in E. coli and the preparation of the mutant PfuRFC
complex using RFC∆12 were performed exactly as previously
described (Cann et al. 2001).
Acknowledgements
We thank M. Yuasa for assistance with the preparation of the
PfuRFC proteins. We also thank N. Sakabe for help with the use
of the Photon Factory BL-6B beamline.This work was supported
in part by New Energy and Industrial Technology Development
Organization (NEDO) in the Japanese Government.
References
Brünger, A.T., Adams, P.D., Clore, G.M., et al. (1998) Crystallography
and NMR system: a new software suite for macromolecular
structure determination. Acta Crystallogr. D54, 905 –921.
Cann, I.K.O. & Ishino, Y. (1999) Archaeal DNA replication:
identifying the pieces to solve a puzzle. Genetics 152, 1249 –1267.
Cann, I.K.O., Ishino, S., Hayashi, I., et al. (1999) Functional
interactions of a homolog of proliferating cell nuclear antigen
with DNA polymerases in Archaea. J. Bacteriol. 181, 6591–6599.
Cann, I., Ishino, S., Yuasa, M., et al. (2001) Biochemical analysis
of replication factor C from the hyperthermophilic archaeon
Pyrococcus furiosus. J. Bacteriol. 183, 2614–2623.
De Felice, M., Sensen, C.W., Charlebois, R.L., Rossi, M. &
Pisani, F.M. (1999) Two DNA polymerase sliding clamps from
the thermophilic archaeon Sulfolobus solfataricus. J. Mol. Biol.
291, 47–57.
Edgell, D.R. & Doolittle, W.F. (1997) Archaea and the origin(s)
of DNA replication proteins. Cell 89, 995 –998.
Gulbis, J.M., Kelman, Z., Hurwitz, J., O’Donnell, M. & Kuriyan, J.
(1996) Structure of the C-terminal region of p21WAF1/CIP1
complexed with human PCNA. Cell 87, 297–306.
Henneke, G., Raffin, J.P., Ferrari, E., Jonsson, Z.O., Dietrich, J.
& Hubscher, U. (2000) The PCNA from Thermococcus fumicolans
functionally interacts with DNA polymerase δ. Biochem. Biophys.
Res. Commun. 276, 600–606.
Ishino, Y. & Cann, I.K.O. (1998) The euryarchaeotes, a subdomain
of Archaea, survive on a single DNA polymerase: fact or farce?
Genes Genet. Syst. 73, 323–336.
Ishino, Y., Tsurimoto, T., Ishino, S. & Cann, I.K.O. (2001)
Functional interactions of an archaeal sliding clamp with
mammalian clamp loader and DNA polymerase δ. Genes Cells
6, 699–706.
Jeruzalmi, D., Yurieva, O., Zhao, Y., et al. (2001) Mechanism of
processivity clamp opening by the δ subunit wrench of the clamp
loader complex of E. coli DNA polymerase III. Cell 106, 417–428.
Jones, T.A., Zou, J.-Y., Cowan, S.W. & Kjeldgaard, M. (1991)
Improved methods for building protein models in electron
density maps and the location of errors in these models. Acta
Crystallogr. A47, 110–119.
Jonsson, Z.O. & Hubscher, U. (1997) Proliferating cell nuclear
antigen: more than a clamp for DNA polymerase. Bioessays 19,
967–975.
Genes to Cells (2002) 7, 911– 922
921
GTC_572.fm Page 922 Monday, August 19, 2002 9:23 PM
S Matsumiya et al.
Kelman, Z. (1997) PCNA: structure, functions, and interactions.
Oncogene 14, 629 – 640.
Kelman, Z. & Hurwitz, J. (1998) Protein–PCNA interactions:
a DNA-scanning mechanism? Trends Biochem. Sci. 23, 236–
238.
Kelman, Z. & Hurwitz, J. (2000) A unique organization of the
protein subunits of the DNA polymerase clamp loader in
the archaeon Methanobacterium thermoautotrophicum ∆H. J. Biol.
Chem. 275, 7327–7336.
Kelman, Z. & O’Donnell, M. (1995) DNA polymerase III
holoenzyme: structure and function of a chromosomal replicating
machine. Annu. Rev. Biochem. 64, 171–200.
Kim, S., Dallmann, H.G., McHenry, C.S. & Marians, K.J. (1996)
Coupling of a replicative polymerase and helicase: a tau–DnaB
interaction mediates rapid replication fork movement. Cell 84,
643 – 650.
Komori, K. & Ishino, Y. (2000) Functional interdependence
of DNA polymerizing and 3′→5′ exonucleolytic activities
in Pyrococcus furiosus DNA polymerase I. Protein Eng. 13, 41–
47.
Kong, X.P., Onrust, R., O’Donnell, M. & Kuriyan, J. (1992)
Three dimensional structure of the β subunit of E. coli DNA
polymerase III holoenzyme: a sliding DNA clamp. Cell 69,
425 – 437.
Kornberg, A. & Baker, T.A. (1992) DNA Replication, 2nd edn.
New York: W.H. Freeman.
Krishna, T.S.R., Kong, X.P., Gary, S., Burgers, P.M. & Kuriyan, J.
(1994) Crystal structure of the eukaryotic DNA polymerase
processivity factor PCNA. Cell 79, 1233–1243.
Leipe, D.D., Aravind, L. & Koonin, E.V. (1999) Did DNA
replication evolve twice independently? Nucl. Acids Res. 27,
3389 –3401.
Matsumiya, S., Ishino, Y. & Morikawa, K. (2001) Crystal structure
of an archaeal DNA sliding clamp: proliferating cell nuclear
antigen from Pyrococcus furiosus. Protein Sci. 10, 17–23.
Moarefi, I., Jeruzalmi, D., Turner, J., O’Donnell, M. & Kuriyan, J.
(2000) Crystal structure of the DNA polymerase processivity
factor of T4 bacteriophage. J. Mol. Biol. 296, 1215–1223.
Montecucco, A., Rossi, R., Levin, D.S., et al. (1998) DNA ligase
I is recruited to sites of DNA replication by an interaction with
proliferating cell nuclear antigen: identification of a common
targeting mechanism for the assembly of replication factories.
EMBO J. 17, 3786 –3795.
Mossi, R. & Hubscher, U. (1998) Clamping down on clamps and
clamp loaders—the eukaryotic replication factor C. Eur. J.
Biochem. 254, 209–216.
922
Genes to Cells (2002) 7, 911– 922
Otwinowski, Z. & Minor, W. (1997) Processing of X-ray diffraction data collected in oscillation mode. Meth. Enzymol. 276,
307–326.
Oyama, T., Ishino, Y., Cann, I.K.O., Ishino, S. & Morikawa, K.
(2001) Atomic structure of the clamp loader small subunit from
Pyrococcus furiosus. Mol. Cell 8, 455–463.
Pisani, F.M., DeFelice, M., Carpentieri, F. & Rossi, M. (2000)
Biochemical characterization of a clamp-loader complex homologous to eukaryotic replication factor C from the hyperthermophilic archaeon Sulfolobus solfataricus. J. Mol. Biol. 301, 61–73.
Sakabe, K., Sasaki, K., Watanabe, N., et al. (1997) Large-format
imaging plate and Weissenberg camera for accurate protein
crystallographic data collection using synchrotron radiation.
J. Synchrotron Radiat. 4, 136–146.
Shamoo, Y. & Steitz, T.A. (1999) Building a replisome from
interacting pieces: sliding clamp complexed to a peptide from
DNA polymerase and a polymerase editing complex. Cell 99,
155–166.
Stillman, B. (1994) Smart machines at the DNA replication fork.
Cell 78, 725–728.
Studwell-Vaughan, P.S. & O’Donnell, M. (1991) Constitution of
the twin polymerase of DNA polymerase III holoenzyme.
J. Biol. Chem. 266, 19833–19841.
Tsurimoto, T. (1998) PCNA, a multifunctional ring on DNA.
Biochim. Biophys. Acta 1443, 23–39.
Tsurimoto, T. (1999) PCNA binding proteins. Front. Biosci. 4,
849–858.
Waga, S. & Stillman, B. (1998) The DNA replication fork in
eukaryotic cells. Annu. Rev. Biochem. 67, 721–751.
Warbrick, E. (1998) PCNA binding through a conserved motif.
Bioessays 20, 195–199.
Warbrick, E. (2000) The puzzle of PCNA’s many partners. Bioessays
22, 997–1006.
Warbrick, E., Heatherington, W., Lane, D.P. & Glover, D.M.
(1998) PCNA binding proteins in Drosophila melanogaster: the
analysis of a conserved PCNA binding domain. Nucl. Acids Res.
26, 3925–3932.
Xu, H., Zhang, P., Liu, L. & Lee, M.Y.W.T. (2001) A novel
PCNA-binding motif identified by the panning of a random
peptide display library. Biochemistry 40, 4512 – 4520.
Yuzhakov, A., Turner, J. & O’Donnell, M. (1996) Replisome
assembly reveals the basis for asymmetric function in leading
and lagging strand replication. Cell 86, 877– 886.
Received: 13 May 2002
Accepted: 18 June 2002
© Blackwell Science Limited