RESEARCH ARTICLE

cryo-EM structures of the E. coli

replicative DNA polymerase reveal its
dynamic interactions with the DNA sliding
clamp, exonuclease and t
Rafael Fernandez-Leiro†, Julian Conrad†, Sjors HW Scheres*, Meindert H Lamers*

MRC Laboratory of Molecular Biology, Cambridge, United Kingdom

Abstract The replicative DNA polymerase PolIIIa from Escherichia coli is a uniquely fast and
processive enzyme. For its activity it relies on the DNA sliding clamp b, the proofreading
exonuclease e and the C-terminal domain of the clamp loader subunit t. Due to the dynamic nature
of the four-protein complex it has long been refractory to structural characterization. Here we
present the 8 Å resolution cryo-electron microscopy structures of DNA-bound and DNA-free states
of the PolIII-clamp-exonuclease-tc complex. The structures show how the polymerase is tethered to
the DNA through multiple contacts with the clamp and exonuclease. A novel contact between the
polymerase and clamp is made in the DNA bound state, facilitated by a large movement of the
polymerase tail domain and tc. These structures provide crucial insights into the organization of the
catalytic core of the replisome and form an important step towards determining the structure of
the complete holoenzyme.
DOI:10.7554/eLife.11134.001

*For correspondence: scheres@

mrc-lmb.cam.ac.uk (SHS);
[email protected]
(MHL)
Introduction
In Escherichia coli, DNA replication is highly efficient with speeds of up 600–1000 nucleotides per
†
These authors contributed second (Mok and Marians, 1987; Mcinerney et al., 2007), >100,000 basepairs (bp) synthesized per
equally to this work binding event (Yao et al., 2009), and an error rate of ~1 per million (Bloom et al., 1997). Impor-
Competing interest: See tantly, DNA replication is greatly complicated by the antiparallel orientation of the two DNA strands
page 13 that need to be replicated simultaneously. To do so, DNA replication is performed by a large multi-
protein complex termed the DNA polymerase III holoenzyme that synthesizes the leading strand in a
Funding: See page 13
continuous manner, while the lagging strand is synthesized in short fragments of ~1000 bp. The
Received: 25 August 2015 holoenzyme is composed of 10 subunits (a, b, e, q, d, d’, g, t, c, ψ), that together with the helicase
Accepted: 23 October 2015 DnaB and the RNA primase DnaG form the replisome with a combined molecular weight of 1 MDa.
Published: 24 October 2015 The replisome can be divided into three functional subcomplexes that together catalyze a series of
Reviewing editor: Stephen C events. The helicase DnaB separates the two DNA strands (Mok and Marians, 1987) and transiently
Kowalczykowski, University of associates with the RNA primase DnaG that synthesizes short RNA primers required for DNA synthe-
California, Davis, United States sis at the lagging strand (Wu et al., 1992). The clamp loader subcomplex (d, d’, g, t, c, ψ) loads the
DNA sliding clamp b, the processivity factor for the DNA polymerase, onto the DNA
Copyright Fernandez-Leiro et
(Stukenberg et al., 1991). It furthermore connects the leading and lagging strand polymerases via
al. This article is distributed under
its t subunits (McHenry, 1982; Onrust et al., 1995). Finally, DNA synthesis is performed by the
the terms of the Creative
Commons Attribution License, polymerase subcomplex that contains the DNA polymerase III a (PolIIIa), the DNA sliding clamp b,
which permits unrestricted use the proofreading exonuclease e, and the C-terminal domain of the clamp loader subunit t. The activ-
and redistribution provided that ity of PolIIIa is poor in isolation (Maki and Kornberg, 1985) and is greatly enhanced by its associ-
the original author and source are ated proteins. For error-free DNA synthesis the polymerase relies on the exonuclease e that removes
credited. any misincorporated bases and decreases the error rate of DNA replication by 1–2 orders of

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 1 of 16

 Research article Biophysics and structural biology

eLife digest DNA replication is complicated because the two strands that form its “double
helix” structure run in opposite directions and need to be replicated at the same time. One of the
new strands, the leading strand, is built continuously. While the other strand, called the lagging
strand, is made in stretches that are about 1000-times shorter and run in the opposite direction from
the leading strand. This means that the enzyme that builds the new strands of DNA (called DNA
polymerase) must be repeatedly released and repositioned when it builds the lagging strand,
however it is not fully understood how this achieved.
Fernandez-Leiro, Conrad et al. have now used a technique called cryo-electron microscopy to
reveal the three-dimensional structure of a DNA polymerase from a bacterium called Escherichia coli
complete with other associated factors and a DNA molecule. These factors include: the “sliding
clamp” that allows the polymerase to slide along the DNA; the “proofreading exonuclease” that
removes mistakes in the newly built DNA strand, and the “processivity switch Tau” that is needed
for the repeated release and repositioning of the polymerase at the lagging strand. These structures
show how the polymerase is bound to the DNA by multiple interactions with the sliding clamp and
exonuclease.
Fernandez-Leiro, Conrad et al. also solved the structure of the same proteins but without the
DNA molecule. This revealed a large structural change between the DNA-bound and DNA-free
states, which provides some clues as to how the polymerase can be quickly released from the DNA
during the repeated cycles of DNA synthesis at the lagging strand. Further research is now needed
to uncover what signals trigger this release of the DNA polymerase.
DOI:10.7554/eLife.11134.002

magnitude (Scheuermann et al., 1983; Lancy et al., 1989). In addition, the exonuclease strengthens
the interactions between the polymerase and clamp as it binds both proteins simultaneously
(Toste Rêgo et al., 2013; Jergic et al., 2013). For processivity, PolIIIa binds to the DNA sliding
clamp (b subunit) (Stukenberg et al., 1991). At the leading strand this interaction is stable and
results in DNA segments of >100.000 bp synthesized per binding event (Yao et al., 2009). At the
lagging strand in contrast, DNA synthesis is discontinuous, with an averaged length of 1000 bp syn-
thesized per fragment, depending on the frequency of the RNA primase activity (Wu, Zechner and
Marians, 1992). This therefore requires repeated binding and release of the polymerase and clamp.
Finally, the C-terminal domain of t (tc) acts as a ’processivity switch’ for the polymerase to enable
repeated binding and release at the lagging strand (Leu et al., 2003; Georgescu et al., 2009). How
this tetrameric complex of PolIIIa-clamp-exonuclease-tc assembles and how it is repeatedly loaded
and released during lagging strand synthesis is poorly understood.
The structures of the helicase-primase subcomplex (Bailey et al., 2007; Wang et al., 2008) and
the clamp loader subcomplex (Jeruzalmi et al., 2001; Simonetta et al., 2009) have been known for
some time. The structure of the PolIIIa-clamp-exonuclease-tc complex on the other hand has
remained elusive due to its dynamic nature that forms a significant hurdle for structure determina-
tion. To overcome this, we have used a combination of site directed mutagenesis and computational
classification of different structural states to determine the cryo-EM structures of the complex in
both a DNA-bound and a DNA-free state to 8 Å resolution. The well defined features of the cryo-
EM maps enable the unambiguous fitting of the crystal structures of the individual proteins, reveal-
ing the unique interactions between the four proteins and DNA. In the DNA-bound complex, the
polymerase is tethered to the DNA through multiple contacts with the clamp. The interaction with
the clamp is further stabilized by the exonuclease that is wedged between the two proteins and
forms a second, indirect interaction between polymerase and clamp. Strikingly, a large conforma-
tional change in the polymerase switches its tail domain from interacting with the clamp in the DNA-
bound structure, to more than 30 Å away from the clamp in the DNA-free structure. Finally, the
processivity switch tc binds the tail of the polymerase and appears to sequester the polymerase tail
away from the clamp in the DNA-free structure. Hence, our structures provide crucial insights into
the regulation of the replicative DNA polymerase PolIIIa by its associated proteins clamp, exonucle-
ase and tc. They furthermore form a crucial step towards determining the structure of the complete
DNA polymerase III holoenzyme.

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 2 of 16

 Research article Biophysics and structural biology

Results
Structure determination of the PolIIIa-clamp-exonuclease-t500 complex
The interaction between PolIIIa and the clamp is weak, in the order of 1 mM (Toste Rêgo et al.,
2013), and is not sufficient to maintain an intact complex at the low concentrations used for cryo-
EM. Therefore, to stabilize the complex we altered the sequences of the clamp binding motifs of
PolIIIa and the exonuclease to increase the affinity for the clamp. For this we used sequences
derived from the translesion DNA polymerase UmuC and the DNA replication initiation factor Hda
that out of a panel of 15 peptide sequences were the most potent inhibitors of the interaction
between the polymerase and clamp (Wijffels et al., 2004) (see Materials and methods for more
details). The obtained complex is >100 fold more stable than the wild-type complex (Figure 1—fig-
ure supplement 1A) This stabilized complex of PolIIIa, clamp and exonuclease was used together
with t500 (the polymerase-binding domain of t: residues 500–643) and a 25 base pair (bp) DNA sub-
strate to prepare samples for cryo-EM (Figure 1—figure supplement 2A,B). Three structurally dis-
tinct groups of particles could be identified from a single data set (63,215 particles). Two of these
represent the PolIIIa-clamp-exonuclease-t500 with and without DNA bound (Figure 1, Videos 1 and
2). The third class contains DNA too, but in this complex the tail domain of the polymerase and t500
are not visible due to structural heterogeneity. The DNA-bound (5663 particles) and DNA-free
(16,970 particles) structures were refined to 8.0 and 8.3 Å resolution, respectively (see Figure 1—fig-
ure supplement 2 for details). The remaining particles (40,582) were classified into the third class in
which the tail domain is not visible. Due to the larger number of particles, this structure was refined
to 7.3 Å resolution. As this structure is otherwise identical to the complete DNA-bound complex, it
will not be discussed further.

Overall structure of the complex

The cryo-EM maps enable the unambiguous fitting of the high-resolution structures of the individual
subunitsinto the cryo-EM maps (Figure 1B,C). No conformational changes were required for the fit-
ting of the clamp, exonuclease or t500, while the polymerase was divided into five domains that were
independently fitted into density as rigid bodies (see Figure 1—figure supplement 3). None of the
loops were modified, with the exception of the clamp binding motifs of the polymerase and exonu-
clease that were modeled after existing crystal structures of clamp-bound peptides from Pol II and
Pol IV (Georgescu et al., 2008a; Bunting et al., 2003). B-form DNA was used for the DNA sub-
strate, except for the last four base pairs that deviate from B-form DNA and were modeled after the
DNA substrate from the Thermus aquaticus (Taq) PolIIIa crystal structure (Wing et al., 2008).
We describe the DNA-free complex first (Figure 2). The overall conformation of PolIIIa resembles
that of the X-ray structure of E. coli and Taq PolIIIa (Lamers et al., 2006; Bailey et al., 2006) and
reveals only a ~15˚ rotation of the fingers domain between the two structures (Figure 1—figure sup-
plement 3). PolIIIa interacts with the clamp through the internal clamp binding motif (residues 920–
924) (Dohrmann and McHenry, 2005; Toste Rêgo et al., 2013) that binds in the canonical binding
pocket of the clamp (Figure 2B). Immediately after the clamp binding motif the density for the poly-
merase disappears, and resumes ~10 residues later, just before the oligonucleotide/oligosaccharide
binding (OB) domain, indicating that this region of the polymerase is flexible (Figure 2A, left and
middle panel).
On the other side of the complex, across the opening of the clamp, the PHP domain of the poly-
merase comes close to, but makes no contacts with the clamp (Figure 2A, left panel). Instead, the
exonuclease is wedged between the clamp and the thumb domain of PolIIIa (Figure 2A, right
panel). The catalytic domain of the exonuclease is in direct contact with the polymerase thumb
domain whereas the contact with the clamp is mediated via a canonical clamp binding motif that is
located immediately downstream of the catalytic domain (Toste Rêgo et al., 2013; Jergic et al.,
2013). This clamp binding motif is bound to the pocket of the clamp in a manner similar to the poly-
merase in the other half of the clamp (cf. Figure 2B,C) and hence both pockets of the dimeric clamp
are occupied in the ternary complex. Downstream of the clamp binding motif the tail of the exonu-
clease follows the contours of the polymerase PHP domain, where it is disordered for a stretch of
~15 residues that were shown to be mobile by NMR (Ozawa et al., 2013). Finally, the C-terminal
helix of the exonuclease that mediates most of the binding to the polymerase (Ozawa et al., 2013)

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 3 of 16

 Research article Biophysics and structural biology

Figure 1. Cryo-EM structures of the E. coli PolIIIa-clamp-exonuclease-t500 complex. (A) Surface representation of the three structures, shown at 5 s. Left
to right: DNA-free, DNA-bound, and DNA-bound without tail. Colors indicate the position of the different proteins (B) Individual structures of PolIIIa,
clamp, exonuclease, and t500 fitted into the cryo-EM map (shown in grey at 5 s) (C) Detailed views of the cryo-EM map (shown in grey mesh at 6 s). Left
panel: exit channel of the clamp in the DNA-free structure showing the ‘DNA-free’ map. Middle panel: bottom view of the polymerase showing the
‘DNA-free’ map. Right panel: detail of the DNA showing the ‘DNA-bound, no tail’ map. See also Videos 1 and 2.
DOI: 10.7554/eLife.11134.003
The following figure supplements are available for Figure 1:
Figure supplement 1. Characterization of improved clamp binding mutants.
DOI: 10.7554/eLife.11134.004
Figure supplement 2. Microscopy data analysis and validation.
DOI: 10.7554/eLife.11134.005
Figure supplement 3. Rigid body movements in PolIIIa.
DOI: 10.7554/eLife.11134.006

packs tightly against the PHP domain of PolIIIa (Figure 2A, left and right panel), similar to the crystal
structure of the PolIIIa-PHP domain and C-terminal helix of the exonuclease (Ozawa et al., 2013).
Hence, in the ternary complex the exonuclease simultaneously binds the polymerase and clamp. By
doing so it strengthens the association between the two proteins (Toste Rêgo et al., 2013), which is

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 4 of 16

 Research article Biophysics and structural biology

required for processive DNA synthesis (Jergic et al., 2013). Previously, we built an approximate
model for the polymerase, clamp and exonuclease complex, using distance restraints provided by
chemical cross-linking coupled to mass-spectrometry (Toste Rêgo et al., 2013). When we map the
same cross-links onto the cryo-EM model we find an improved fit of the cross-links with the model,
due to the conformational changes in the complex as well as the detailed information about the
interactions between the proteins that could not be modeled based on the cross-links alone (Fig-
ure 2—figure supplement 1).
The NMR structure of residues 500–621 of t (Su et al., 2007) can be fitted into the density
between the tail and fingers domain of the polymerase (Figure 2D and Figure 2—figure supple-
ment 2A). The C-terminal end of t is in contact with the tail of the polymerase. Unfortunately, the
last 23 residues of t that bind the polymerase (Jergic et al., 2007) are not part of the NMR structure
and are therefore not present in our model. A second contact is found between the globular domain
of t500 and the fingers domain of the polymerase. This contact is mediated by residues 530–535 and
562–566 of t500 and residues 657–667 of PolIIIa (Figure 2—figure supplement 2B). The position of
t500 in this complex is different from the position of the C-terminal domain of t in complex with Taq
PolIIIa, where it is located at the opposite side of the polymerase tail (Figure 2—figure supplement
3) (Liu et al., 2013). It must be noted though that the C-terminal domain of t from E. coli and Taq
share no sequence or structural homology and therefore engage with the polymerase in different
ways.

DNA binding in the PolIIIa-clamp-exonuclease-t500 complex

In the DNA-bound complex (Figure 3), the entire length of the 25 base pair duplex is in contact with
protein (Figure 3A). The position of the DNA is similar to that of the DNA in the crystal structure of
Taq PolIIIa and Geobacillus kaustophilus PolC (Wing et al., 2008; Evans et al., 2008) (Figure 3—
figure supplement 1). No density is observed for the 4 nucleotide (nt) single stranded overhang on
the template strand indicating that this part of the DNA is flexible. In the complex, all contacts to
the DNA are mediated by the thumb, palm and fingers domains of the polymerase and the inner sur-
face of the clamp. No contacts to the DNA are made by the polymerase OB domain, the exonucle-
ase, or t500. The most extensive DNA contacts occur at the primer 3’ end in polymerase active site
where the thumb, palm and fingers domain of the polymerase contact the first 9 base pairs of the
DNA duplex. It is also here that the only non-backbone contact is made by a loop of the thumb (resi-
dues 464–470), which is inserted into the major groove of the DNA (Figure 3B).
Away from the active site, the tip of the fingers domain (i.e. little finger [Lamers et al., 2006] or b
binding domain [Bailey et al., 2006]), makes additional contacts to the DNA. Here, several positively
charged residues (K831, K872, R876, R877, K881) as well as the positive charge of the helix dipole of

Video 1. Structure of the DNA-free complex of PolIIa- Video 2. Structure of the DNA-bound complex of
clamp-exonuclease-t500, Related to Figure 1. Fitting of PolIIa-clamp-exonuclease-t500, Related to Figure 1.
the high-resolution structures into the cryo-EM map of Fitting of the high-resolution structures into the cryo-
the DNA-free complex. EM map of the DNA-bound complex.
DOI: 10.7554/eLife.11134.007 DOI: 10.7554/eLife.11134.008

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 5 of 16

 Research article Biophysics and structural biology

Figure 2. Multiple contacts between the subunits hold the complex together. (A) Three different views of the DNA-free complex of PolIIIa-clamp-
exonuclease-t500 showing extensive contacts between the polymerase and other subunits. Missing loops in PolIIIa (residues 927–936) and exonuclease
(residues 190–207) are shown in dots. Dashed boxes indicate views shown in panels B-D. (B) Modified clamp binding motif of PolIIIa (QLDLF: shown in
sticks) modeled into the binding pocket of the clamp. (C) Modified clamp binding motif of the exonuclease (QLSLPL: shown in sticks) modeled into the
second binding pocket of the dimeric clamp. (D) t500 simultaneously binds the fingers and tail domain of the polymerase. The C-terminal residues of
t500 (residues 622–643: not modeled) bind to the tail of the polymerase, while the globular domain of t500 binds to the polymerase fingers domain (see
Figure 2—figure supplement 2 for more details).
DOI: 10.7554/eLife.11134.009
The following figure supplements are available for Figure 2:
Figure supplement 1. Previously determined cross-links fit accurately with the cryo-EM model.
DOI: 10.7554/eLife.11134.010
Figure supplement 2. Details of the interactions between t500 and the PolIIIa fingers domain.
DOI: 10.7554/eLife.11134.011
Figure supplement 3. Comparison of t binding in E. coli and Taq PolIIIa.
DOI: 10.7554/eLife.11134.012

two helices (residues 842–856 and 875–886) are pointed towards the DNA backbone (Figure 3A). At
this position, the OB domain is in close proximity of the DNA but makes no contacts with it
(Figure 3C). Instead, the OB domain forms a bridge between the PolIIIa fingers domain, thumb
domain, and the exonuclease. Furthermore, while the isolated OB domain has been shown to bind
to ssDNA (Georgescu et al., 2009), in this complex the OB domain is ~40 Å away from the ssDNA
template. The DNA furthermore interacts with the clamp that surrounds the DNA like a nut around a
bolt (Figure 3D). Several non-specific contacts are made to the backbone of the DNA providing an
electrostatic cushion for the DNA to pass through as it leaves the complex. Compared to the crystal
structure of the isolated clamp bound to DNA (Georgescu et al., 2008a) the clamp is rotated by
~20˚, resulting in an almost perpendicular orientation (~80˚) with respect to the DNA. In the 19 Å,
negative stain EM structure of Pyrococcus furiosus PolB, the only other known structure of a DNA
polymerase in complex with clamp and DNA, the DNA runs straight through the clamp as well
(Mayanagi et al., 2011).

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 6 of 16

 Research article Biophysics and structural biology

Figure 3. The DNA has extensive contacts with PolIIIa and clamp. (A) Overview of the DNA-bound complex. The
N-termini of the two helices that point at the DNA backbone are colored in blue. Potential DNA interacting side
chains are shown in sticks. The tail of PolIIIa, the exonuclease and t500 are omitted for clarity. Arrow indicates
viewpoint in panel B (B) Polymerase active site, with the DNA held between thumb, palm and fingers domain.
Polymerase active site residues are indicated with magenta spheres. Arrow indicates viewpoint in panel C (C) DNA
interactions downstream of the active site. The OB domain is positioned on top of the DNA but does not make
any contacts with it. (D) DNA exit channel in the clamp with positively charged residues within 10 Å of the DNA
indicated in magenta sticks. Note that the positions of the side chains have not been refined and should be seen
as approximate positions. (E) In the DNA-bound complex, the clamp is at a ~80˚ angle from the DNA. A dashed
line indicates the position of the clamp alone bound to a DNA substrate. (Georgescu et al., 2008a). The other
subunits (PolIIIa, exonuclease, t500) are shown in light grey for clarity.
DOI: 10.7554/eLife.11134.013
The following figure supplements are available for Figure 3:
Figure supplement 1. Comparison of DNA binding by C family DNA polymerases.
DOI: 10.7554/eLife.11134.014
Figure supplement 2. Pol IIIa has more extensive DNA interactions than other DNA polymerases.
DOI: 10.7554/eLife.11134.015

PolIIIa is an extremely fast DNA polymerase with DNA synthesis speeds of up to 600–1000 nt/s
(Mok and Marians, 1987; Mcinerney et al., 2007). In contrast, the E. coli DNA polymerases Pol I,
Pol II, and Pol IV have synthesis speeds of 15, 10, and 1 nt/s, respectively (Schwartz and Quake,
2009; Indiani et al., 2009). Furthermore, in isolation PolIIIa has a surprisingly low affinity for DNA
when compared to the other E. coli DNA polymerase (Figure 3—figure supplement 2A). Because
of its weak binding to DNA, PolIIIa must therefore have developed a different way to keep itself cor-
rectly positioned on the DNA during rapid DNA synthesis. To this effect, PolIIIa may have evolved

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 7 of 16

 Research article Biophysics and structural biology

its uniquely long fingers domain that is more than twice as long as the other E. coli DNA polymer-
ases (Figure 3—figure supplement 2B). These polymerases have considerably smaller fingers
domains, use shorter regions of DNA contacts, and have shorter predicted distances between the
polymerase active site and the clamp (Figure 3—figure supplement 2B). In PolIIIa, the number of
base pairs between the 3’ end of the primer and the opening of the clamp is 22, while the predicted
number of base pairs for Pol I, Pol II, and Pol IV is ~15. This unusually long DNA-protein contact
appears to be well suited to accurately position the DNA without requiring tight binding that could
slow down the translocation of the DNA. At the same time, the sequence-independent backbone
contacts and the perfectly straight B-form DNA may facilitate the rapid exit of the DNA from the
active site. The active site itself is wrapped tightly around the DNA where PolIIIa is the only polymer-
ase that inserts a loop of the thumb domain into the major groove of the DNA, while the thumb
domains of Pol I, Pol II, and Pol IV only have backbone contacts with the DNA (Figure 3—figure sup-
plement 2C). Hence, it seems plausible that this combination of unique contacts with the DNA may
have evolved to support the high speeds of DNA synthesis by PolIIIa without compromising
accuracy.

DNA binding induces a large conformational change in the complex

To enable the many contacts with the DNA, the complex undergoes extensive conformational
changes from the DNA-free to the DNA-bound state. Most prominent is a ~35˚ rotation of the poly-
merase tail and t500, which move from a position over the polymerase active site to a position adja-
cent to the sliding clamp (Figures 1,4 and Video 3). The simultaneous movement of the polymerase
tail and tc results in a 70 Å displacement of the globular domain of t500. The tail of PolIIIa consists of
the OB domain (residues 960–1071) and the C-terminal t-binding domain (residues 1079–1160).
Together with t500 they form a rigid structure that shows few changes between the DNA free and
DNA bound structure, indicating that the interaction between t500 and the tail of PolIIIa must be sta-
ble (Figure 4C). As a result of the repositioning of the polymerase tail, the contact between t500 and
the fingers domain of the polymerase is broken and a new contact between the OB domain and the
clamp is forged (Figure 4B and Video 3). The OB domain makes two new contacts with the clamp
via a short helix (1035–1043) and a long protruding loop (1003–1013) (Figure 4D,E). These motifs
contact the clamp at loops 24–28 and 275–278, respectively. Hence in the DNA-bound complex, the
polymerase has three points of contact to the clamp: one via the canonical clamp binding motif (resi-
dues 920–925: Figure 2B); one indirectly via the exonuclease (Figure 2C); and one contact via the
OB domain (Figure 4D,E). Previously, it has been shown that a triple mutation in OB domain result
in reduced DNA synthesis (Georgescu et al., 2009) which was attributed to the loss of ssDNA bind-
ing by the OB domain. However, in our structure the OB domain is far away (~40 Å) from the ssDNA
overhang of the template strand. Instead, the mutations (R1004S, K1009S, R1010S) are located at
the interface between the OB domain and the clamp (Figure 4E) and therefore could weaken the
interaction between the polymerase and clamp, providing an alternative explanation for the reduced
DNA synthesis.

Discussion
The E. coli replisome consists of 12 different proteins that can be divided into three subcomplexes:
the helicase-primase complex, the clamp loader complex, and the PolIIIa-clamp-exonuclease-tc com-
plex. The structures of two of the three subcomplexes have been determined previously: the heli-
case-primase complex (Bailey et al., 2007; Wang et al., 2008), and the clamp loader complex
(Jeruzalmi et al., 2001; Simonetta et al., 2009). The structure of the PolIIIa-clamp-exonuclease-t500
complex on the other hand has remained elusive due to its dynamic nature. The cryo-EM structures
of the PolIIIa-clamp-exonuclease-t500 complex presented in this work finally reveal the nature of the
interactions in the ternary complex and are a crucial step forward towards determining the structure
of the complete bacterial replisome.
Our cryo-EM structures furthermore provide a crucial insight into the structural organization of
the replicative DNA polymerase and its associated proteins clamp, exonuclease and t500. They show
how the clamp and exonuclease tether the polymerase to the DNA through multiple contacts.
Importantly, they also reveal a large conformational change where the tail of the polymerase moves
from interacting with the clamp in the DNA-bound state, to a position 35 Å away from the clamp in

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 8 of 16

 Research article Biophysics and structural biology

Figure 4. DNA binding induces large conformational changes in the polymerase. (A) Clamp binding by PolIIIa in the DNA-free complex. Arrows
indicate movement of the PolIIIa tail (see also Video 3). (B) Clamp binding by PolIIIa in the DNA-bound complex. Dashed boxes indicate views shown
in panel D and E (C) Comparison of the PolIIIa-tail - t500 interaction in the DNA-free (in grey) and DNA-bound structure. (D and E) Detailed view of the
clamp - PolIIIa OB domain interaction. Interacting regions at the interface are indicated in thick coil in magenta (clamp: residues 24–24 and residues
275–278) and red (PolIIIa-OB domain: residues 1035–1043 and residues 1003–1013). Residues mutated in Georgescu et al (Georgescu et al., 2009) are
shown in sticks and labeled with outlined boxes. Note that the positions of the side chains have not been refined and should be seen as approximate
positions.
DOI: 10.7554/eLife.11134.016

the DNA-free state. What may be the role for such a conformational change? On the lagging strand,
the polymerase repositions to a newly primed site every ~1000 bp. To do so, the polymerase needs
to release both clamp and DNA. We propose that the switch-like movement of the polymerase tail
may play a part in the release and consequent repositioning of the polymerase at the end of the
Okazaki fragment. A hypothetical model describing how this could work is presented in Figure 5.
During DNA synthesis, the tail of the polymerase is bound to the clamp, stabilizing the interaction
between polymerase and clamp (marked with ‘1’ in Figure 5). This confirmation may be further sta-
bilized by the presence of a DNA binding region immediately upstream of t500 (marked with ‘2’)
(Jergic et al., 2007). Upon encounter of a release signal, t500 rebind to the polymerase fingers
domain (marked with ‘3’) thus sequestering the polymerase tail away from the clamp (marked with
‘4’) and initiating the dissociation of the polymerase from clamp and DNA. What could serve as the
release trigger? Two non-exclusive models have been proposed (Li and Marians, 2000). In the ‘colli-
sion’ model, the encounter with the dsDNA of the previous Okazaki fragment induces the release of
the polymerase. In support of this model, it has been shown that a decreasing gap size between the
3’ terminus of the lagging strand and the 5’ end of the previously synthesized Okazaki fragment pro-
motes release of the polymerase (Leu et al., 2003; Georgescu et al., 2009; Dohrmann et al.,
2011). Two possible sensors for the decreasing gap on the lagging strand have been suggested.
The ssDNA binding properties of the OB domain in the tail of the polymerase has been proposed to

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 9 of 16

 Research article Biophysics and structural biology

play a role in the sensing of the ssDNA vs

dsDNA (Georgescu et al., 2009). Yet our cryo-
EM models show that the OB domain is ~40 Å
away from the ssDNA and that it is involved in
binding to the sliding clamp. Alternatively, it has
been proposed that the C-terminal fragment of
t may act as the sensor as it is required to
release the polymerase from a decreasing gap
size (Leu et al., 2003). Indeed, it was found that
the region in t, immediately upstream of t500,
has DNA binding affinity (Jergic et al., 2007).
Contesting the collision model is the observa-
tion that the release of the polymerase by a
decreasing gap size is too slow (t1/2 = 110 s) for
the frequency at which the lagging strand poly-
Video 3. DNA binding induces large conformational merase is re-positioned (every 1–2 s)
changes in the complex, Related to Figure 4. Linear (Dohrmann et al., 2011), suggesting that addi-
morphing of the DNA-free to DNA-bound state
tional or alternative release factors are required.
showing the large conformational change between the
The alternative ‘signaling’ model therefore pro-
two states.
DOI: 10.7554/eLife.11134.017
poses that the trigger comes from one of the
other components of the replisome such as the
RNA primase DnaG, based on the observation
that the increased primase concentration results
in shorter lagging strand fragments (Wu, Zechner and Marians, 1992; Li and Marians, 2000). How-
ever, it was recently shown that the presence of a primer alone is sufficient to induce release at the
lagging strand and that activity of the primase is not required (Yuan and McHenry, 2014). How the
presence of the RNA primer is signaled to the polymerase remains unclear. Yet, the fact that the t
protein is both part of the clamp loader complex that positions clamps onto the primer and simulta-
neously binds the polymerase makes this a suitable conveyor of the signal. While our structures do
not discriminate between the type of release trigger for the lagging strand polymerase, they do now
provide the means to test the precise workings of the molecular switch that enables the release of
the polymerase.

Materials and methods

Materials
All chemicals and oligonucleotides were purchased from Sigma-Aldrich (Gillingham,
United Kingdom) and chromatography columns from GE healthcare (Little Chalfont, United
Kingdom).

Protein expression and purification

To increase binding to the clamp, amino acid residues 920–924 of E. coli PolIIIa were changed by
site directed mutagenesis from QADMF to QLDLF, while in the exonuclease residues 182–187 were
changed from QTSMAF to QLSLPL, based on sequences described in (Wijffels et al., 2004). All pro-
teins were expressed in E. coli (DE3) BL21. PolIIIa, clamp and exonuclease were expressed and puri-
fied as described before (Toste Rêgo et al., 2013). t500 was purified by Histrap HP column,
Resource S column, and a Superdex 75 gel filtration column. His-tags were removed by proteolytic
cleavage with human rhinovirus 3C protease. Proteins were flash frozen in liquid nitrogen and stored
at -80˚.

Gel filtration analysis

Proteins were analyzed by gel filtration using a 2.4 mL Superdex 200 Increase column (GE health-
care) in 25 mM Hepes pH 7.5, 150 mM NaCl, and 2 mM DTT. PolIIIa-clamp-exonuclease complex
was assembled at 10, 1, and 0.1 mM and 50 mL injected onto the column.

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 10 of 16

 Research article Biophysics and structural biology

Figure 5. Schematic representation for a possible role of the conformational changes in the polymerase. During processive DNA synthesis, the tail of
the polymerase is attached to the clamp (indicated with ‘1’) and pulls t500 away from the polymerase fingers domain. This conformation may be further
stabilized by the presence of a DNA binding region immediately upstream of t500 (indicated with ‘2’; see text for more details).Upon encounter of a
release trigger, the contact between t500 and the polymerase fingers domain is restored (indicated with ‘3’), and the contact between the clamp and
polymerase tail is broken (indicated with ‘4’). The release trigger may either come from a collision with the previous Okazaki fragment (indicated with
‘Collision’), or a signal from other replisome components via the flexible linker of t (indicated with ‘Signaling’). Once the polymerase-tail clamp contact
has been broken, the two remaining contacts between the clamp and polymerase-exonuclease are not enough to keep the polymerase bound to the
clamp. The polymerase is released from clamp and DNA and can be repositioned to a newly primed site to reinitiate DNA synthesis.
DOI: 10.7554/eLife.11134.018

Electro mobility shift assay

A DNA substrate identical to the substrate used for the cryo-EM samples was used, with the excep-
tion of a 6-carboxyfluorescein (6-FAM) at the 5’ end of the primer strand and a phosphorothioate
link at the 3’ terminal bond to prevent exonuclease digestion. 5 nM DNA was incubated with 2.5 mM
polymerase (E. coli Pol I (Klenow fragment), Pol II, Pol IIIa, or Pol IV) for 10 min at room temperature.
Reaction mixtures contained 20 mM Tris pH 7.5, 4% glycerol, 5 mM DTT, 40 mg/ml BSA, and 40 mM
Potassium Glutamate. Half of the sample was separated on a native 6% acrylamide gel and imaged
on a Typhoon laser scanner (GE Healthcare). The remaining half of the sample was analyzed on a
denaturing 4–12% SDS acrylamide gel and stained with Coomassie blue.

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 11 of 16

 Research article Biophysics and structural biology

Sample preparation for cryo-EM

The PolIIIa-clamp-exonuclease-t500 protein complex was assembled from the individual components
to a final concentration of 15 mM and purified on a 2.4 mL Superdex 200 Increase gel filtration col-
umn (GE Healthcare) in 25 mM Hepes pH 7.5, 50 mM Potassium Glutamate, 3 mM Magnesium Ace-
tate, and 2 mM DTT. The peak fraction (~4 mM) was retrieved and incubated for 5 min with 20 mM of
a 25 bp DNA substrate with a 4 nt overhang (template: 50-TCAGGAGTCCTTCGTCCTAGTACTAC-
TCC-30, primer: 50-GGAGTAGTACTAGGACGAAGGACTC-30) for 5 min at room temperature. Subse-
quently, 0.1 volume of 0.05% (V/V) Tween 20 was added and incubated for another 5 min before the
samples were pipetted onto glow-discharged holey carbon cryo-EM grids (Quantifoil Cu R1.2/1.3),
and frozen in liquid ethane using a Vitrobot (FEI, Hillsboro, OR).

Data collection and image processing

All data was collected using a Titan Krios electron microscope (FEI) operated at 300 kV equipped
with a K2 summit direct electron detector (Gatan, Pleasaston, CA). Although this detector was
mounted after a Gatan Imaging Filter (GIF), the filter was not used to remove any inelastic scatter-
ing. Images were collected in single-electron counting mode at a calibrated magnification of
28.571x (1.76 Å/pixel), using a flux of 2 e/Å2/sec and a total dose of 40 e/Å2 over a total of 20
frames. Frames were aligned and averaged using whole-image movement correction using MOTIO-
NCORR (Li et al., 2013). Contrast transfer function parameters were calculated using CTFFIND3
(Mindell and Grigorieff, 2003). All subsequent particle picking and data processing was performed
using Relion-1.3 (Scheres, 2012), with the exception of the generation of the initial model, which
was done using Eman2 (Tang et al., 2007). A total of 1350 micrographs were recorded from which
>550,000 particles were picked automatically in Relion. After 2D classification, a large number of
spurious particles as well as particles that show free polymerase or free clamp were removed, yield-
ing a dataset of ~90,000 particles. After 3D classification a another ~27,000 were removed to yield a
final dataset of 63,215 particles. From these, six 3D classes were calculated that were subsequently
merged into the final three 3D classes of ’DNA-free’ (16,970 particles), ’DNA-bound’ (5663 particles)
and ’DNA-bound, no tail’ (40,582 particles). Particle-based movement correction and per-frame B-
factor weighting to account for radiation damage and unresolved particle movement was performed
in the later stages of refinement using the particle polishing option in Relion (Scheres, 2014).
Reported resolutions are based on the gold-standard FSC-0.143 criterion (Scheres and Chen, 2012)
and FSC-curves were corrected for the convolution effects of a soft mask using high-resolution
noise-substitution (Chen et al., 2013). All density maps were sharpened by applying a negative B-
factor that was estimated using automated procedures (Rosenthal and Henderson, 2003). We
believe that the resolution of these reconstructions is limited by both the relatively small size of the
complex (250 kDa), which hampers accurate alignment and classification, and the inherent flexibility
of this four-protein and DNA complex. Still, the maps are of excellent quality, with individual helices,
b-sheets, and loops clearly visible in the map (Figure 1C).

Fitting of the crystal structures into the cryo-EM map

Individual crystal or NMR structures were manually placed into the cryo-EM map in PyMOL
(Schrödinger, LLC 2010) and subsequently rigid-body fitted into the density using Coot
(Emsley et al., 2010). PDB codes of the fitted structures are: PolIIIa: 2HNH (Lamers et al., 2006),
clamp: 2POL (Kong et al., 1992), exonuclease: 1J54 (Hamdan, et al., 2002), t500: 2AYA (Su et al.,
2007). The C-terminal tail of Eco PolIIIa that is lacking in the crystal structure (2HNH) was modeled
as described in (Toste Rêgo et al., 2013). The PolIIIa structure was divided into five domains that
were further fitted independently into density as rigid bodies (see Figure 1—figure supplement 3B,
C). These domains were: PHP (residues 1–280), palm-fingers (residues 281–432 + 510–810), thumb
(residues 433–509), tip-of-fingers (residues 811–928) and C-terminal tail (residues 929–1160). Clamp
binding motifs of PolIIIa and exonuclease were manually built into the clamp in Coot guided by the
crystal structures of clamp-bound peptides from Pol II and Pol IV, (Bunting et al., 2003;
Georgescu et al., 2008b; Jeruzalmi et al., 2001). The DNA substrate was generated with Coot,
and the last four base pairs of the DNA were adjusted guided by the DNA from Taq Pol IIIa
(Wing et al., 2008).

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 12 of 16

 Research article Biophysics and structural biology

Comparison of DNA polymerase structures

The following crystal structures of C family DNA polymerases were used to compare DNA binding
and t binding. DNA bound Taq PolIIIa (PDB code: 3E0D [Wing et al., 2008]), t bound Taq PolIIIa
(PDB code: 4IQJ [Liu et al., 2013]), DNA bound G. kaustophilus PolC (PDB code: 3F2B
[Evans et al., 2008]). Crystal structures of bacterial DNA polymerases in complex with DNA were
used to compare the distance between the polymerase active site and the opening to the clamp.
The following structures were used: T. aquaticus DNA Pol I (PDB code: 1QTM [Li et al., 1999]), E.
coli Pol II (PDB code: 3K57 [Wang and Yang, 2009]), E. coli PolIIIa (this work), and E. coli Pol IV
(PDB code: 4IRD [Sharma et al., 2013]). For the structures of Pol I, Pol II, and Pol IV, the sliding
clamp (PDB code: 2POL [Kong et al., 1992]) was manually placed close to the clamp binding
sequences in the different polymerases, taking care not to cause any clashes with other parts of the
polymerase.

Acknowledgements
We thank members of the Lamers lab as well as Xiaochen Bai, Paula da Fonseca, and Nigel Unwin
for useful suggestions. We also thank David Barford and Ana Toste-Rêgo for critically reading the
manuscript. This work was supported by the UK Medical Research Council through grants
U105197143 to MHL and MC_UP_A025_1013 to SHWS Cryo-EM density maps have been deposited
with the Electron Microscopy Data Bank (accession numbers EMD-3198, EMD-3201, and EMD3202),
and coordinates of the models have been deposited with the Protein Data Bank (PDB entry codes
5FKV, 5FKU, and 5FKW)

Additional information
Competing interests
SHWS: Reviewing editor, eLife. The other authors declare that no competing interests exist.

Funding
Funder Grant reference number Author
Medical Research Counc U105197143 Meindert H Lamers
Medical Research Council MC_UP_A025_1013 Sjors HW Scheres

The funders had no role in study design, data collection and interpretation, or the decision to
submit the work for publication.

Author contributions
RFL, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or
revising the article; JC, Acquisition of data, Analysis and interpretation of data; SHWS, Analysis and
interpretation of data, Drafting or revising the article; MHL, Conception and design, Analysis and
interpretation of data, Drafting or revising the article, Contributed unpublished essential data or
reagents

Author ORCIDs
Rafael Fernandez-Leiro, http://orcid.org/0000-0002-7941-0357
Sjors HW Scheres, http://orcid.org/0000-0002-0462-6540

References
Bailey S, Eliason WK, Steitz TA. 2007. Structure of hexameric DnaB helicase and its complex with a domain of
DnaG primase. Science 318:459–463. doi: 10.1126/science.1147353
Bailey S, Wing RA, Steitz TA. 2006. The structure of T. aquaticus DNA polymerase III is distinct from eukaryotic
replicative DNA polymerases. Cell 126:893–904. doi: 10.1016/j.cell.2006.07.027
Bloom LB, Chen X, Fygenson DK, Turner J, O’Donnell M, Goodman MF. 1997. Fidelity of Escherichia coli DNA
polymerase III holoenzyme: The effects of beta, gamma complex processivity proteins and epsilon

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 13 of 16

 Research article Biophysics and structural biology

proofreadingexonuclease on nucleotide misincorporation efficiencies. Journal of Biological Chemistry 272:

27919–27930. doi: 10.1074/jbc.272.44.27919
Bunting KA. 2003. Structural basis for recruitment of translesion DNA polymerase Pol IV/DinB to the beta-clamp.
The EMBO Journal 22:5883–5892. doi: 10.1093/emboj/cdg568
Chen S, McMullan G, Faruqi AR, Murshudov GN, Short JM, Scheres SHW, Henderson R. 2013. High-resolution
noise substitution to measure overfitting and validate resolution in 3D structure determination by single particle
electron cryomicroscopy. Ultramicroscopy 135:24–35. doi: 10.1016/j.ultramic.2013.06.004
Dohrmann PR, Manhart CM, Downey CD, Mchenry CS. 2011. The rate of polymerase release upon filling the gap
between okazaki fragments is inadequate to support cycling during lagging strand synthesis. Journal of
Molecular Biology 414:15–27. doi: 10.1016/j.jmb.2011.09.039
Dohrmann PR, Mchenry CS. 2005. A bipartite polymerase–processivity factor interaction: only the internal b
binding site of the a subunit is required for processive replication by the DNA polymerase III holoenzyme.
Journal of Molecular Biology 350:228–239. doi: 10.1016/j.jmb.2005.04.065
Emsley P, Lohkamp B, Scott WG, Cowtan K. 2010. Features and development of coot . Acta Crystallographica
Section D. Biological Crystallography 66:486–501. doi: 10.1107/S0907444910007493
Evans RJ, Davies DR, Bullard JM, Christensen J, Green LS, Guiles JW, Pata JD, Ribble WK, Janjic N, Jarvis TC.
2008. Structure of PolC reveals unique DNA binding and fidelity determinants. Proceedings of the National
Academy of Sciences of the United States of America 105:20695–20700. doi: 10.1073/pnas.0809989106
Georgescu RE, Kim S-S, Yurieva O, Kuriyan J, Kong X-P, O’Donnell M. 2008a. Structure of a sliding clamp on
DNA. Cell 132:43–54. doi: 10.1016/j.cell.2007.11.045
Georgescu RE, Kurth I, Yao NY, Stewart J, Yurieva O, O’Donnell M. 2009. Mechanism of polymerase collision
release from sliding clamps on the lagging strand. The EMBO Journal 28:2981–2991. doi: 10.1038/emboj.2009.
233
Georgescu RE, Yurieva O, Kim S-S, Kuriyan J, Kong X-P, O’Donnell M, . 2008b. Structure of a small-molecule
inhibitor of a DNA polymerase sliding clamp. Proceedings of the National Academy of Sciences of the United
States of America 105:11116–11121. doi: 10.1073/pnas.0804754105
Hamdan S, Carr PD, Brown SE, Ollis DL, Dixon NE. 2002. Structural basis for proofreading during replication of
the escherichia coli chromosome. Structure 10:535–546. doi: 10.1016/S0969-2126(02)00738-4
Hwang H, Kim H, Myong S. 2011. Protein induced fluorescence enhancement as a single molecule assay with
short distance sensitivity. Proceedings of the National Academy of Sciences of the United States of America
108:7414–7418. doi: 10.1073/pnas.1017672108
Indiani C, Langston LD, Yurieva O, Goodman MF, O’Donnell M. 2009. Translesion DNA polymerases remodel the
replisome and alter the speed of the replicative helicase. Proceedings of the National Academy of Sciences of
the United States of America 106:6031–6038. doi: 10.1073/pnas.0901403106
Jergic S, Horan NP, Elshenawy MM, Mason CE, Urathamakul T, Ozawa K, Robinson A, Goudsmits JMH, Wang Y,
Pan X, Beck JL, van Oijen AM, Huber T, Hamdan SM, Dixon NE. 2013. A direct proofreader–clamp interaction
stabilizes the Pol III replicase in the polymerization mode. The EMBO Journal 32:1322–1333. doi: 10.1038/
emboj.2012.347
Jergic S, Ozawa K, Williams NK, Su X-C, Scott DD, Hamdan SM, Crowther JA, Otting G, Dixon NE. 2007. The
unstructured C-terminus of the tau subunit of Escherichia coli DNA polymerase III holoenzyme is the site of
interaction with the alpha subunit. Nucleic Acids Research 35:2813–2824. doi: 10.1093/nar/gkm079
Jeruzalmi D, O’Donnell M, Kuriyan J. 2001. Crystal structure of the processivity clamp loader gamma (g) complex
of E. coli DNA polymerase III. Cell 106:429–441. doi: 10.1016/S0092-8674(01)00463-9
Jeruzalmi D, Yurieva O, Zhao Y, Young M, Stewart J, Hingorani M, O’Donnell M, Kuriyan J. 2001. Mechanism of
processivity clamp opening by the delta subunit wrench of the clamp loader complex of E. coli DNA
polymerase III. Cell 106:417–428. doi: 10.1016/S0092-8674(01)00462-7
Kong X-P, Onrust R, O’Donnell M, Kuriyan J. 1992. Three-dimensional structure of the beta subunit of E. coli
DNA polymerase III holoenzyme: a sliding DNA clamp. Cell 69:425–437. doi: 10.1016/0092-8674(92)90445-I
Lamers MH, Georgescu RE, Lee S-G, O’Donnell M, Kuriyan J. 2006. Crystal structure of the catalytic alpha
subunit of E. coli replicative DNA polymerase III. Cell 126:881–892. doi: 10.1016/j.cell.2006.07.028
Lancy ED, Lifsics MR, Kehres DG, Maurer R. 1989. Isolation and characterization of mutants with deletions in
dnaQ, the gene for the editing subunit of DNA polymerase III in Salmonella typhimurium. Journal of
Bacteriology 171:5572–5580.
Leu FP, Georgescu R, O’Donnell M. 2003. Mechanism of the E. coli tau processivity switch during lagging-strand
synthesis. Molecular Cell 11:315–327. doi: 10.1016/S1097-2765(03)00042-X
Li X, Marians KJ. 2000. Two distinct triggers for cycling of the lagging strand polymerase at the replication fork.
Journal of Biological Chemistry 275:34757–34765. doi: 10.1074/jbc.M006556200
Li X, Mooney P, Zheng S, Booth CR, Braunfeld MB, Gubbens S, Agard DA, Cheng Y. 2013. Electron counting
and beam-induced motion correction enable near-atomic-resolution single-particle cryo-EM. Nature Methods
10:584–590. doi: 10.1038/nmeth.2472
Li Y, Mitaxov V, Waksman G. 1999. Structure-based design of taq DNA polymerases with improved properties of
dideoxynucleotide incorporation. Proceedings of the National Academy of Sciences of the United States of
America 96:9491–9496. doi: 10.1073/pnas.96.17.9491
Liu B, Lin J, Steitz Thomas A. 2013. Structure of the PolIIIa-tc-DNA complex suggests an atomic model of the
replisome. Structure 21:658–664. doi: 10.1016/j.str.2013.02.002
Maki H, Kornberg A. 1985. The polymerase subunit of DNA polymerase III of Escherichia coli. II. Purification of
the alpha subunit, devoid of nuclease activities. The Journal of Biological Chemistry 260:12987–12992.

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 14 of 16

 Research article Biophysics and structural biology

Mayanagi K, Kiyonari S, Nishida H, Saito M, Kohda D, Ishino Y, Shirai T, Morikawa K. 2011. Architecture of the
DNA polymerase b-proliferating cell nuclear antigen (pCNA)-DNA ternary complex. Proceedings of the
National Academy of Sciences of the United States of America 108:1845–1849. doi: 10.1073/pnas.1010933108
McHenry CS. 1982. Purification and characterization of DNA polymerase III’. Identification of tau as a subunit of
the DNA polymerase III holoenzyme. The Journal of Biological Chemistry 257:2657–2663.
McInerney P, Johnson A, Katz F, O’Donnell M. 2007. Characterization of a triple DNA polymerase replisome.
Molecular Cell 27:527–538. doi: 10.1016/j.molcel.2007.06.019
Mindell JA, Grigorieff N. 2003. Accurate determination of local defocus and specimen tilt in electron microscopy.
Journal of Structural Biology 142:334–347. doi: 10.1016/S1047-8477(03)00069-8
Mok M, Marians KJ. 1987. The Escherichia coli preprimosome and DNA B helicase can form replication forks that
move at the same rate. The Journal of Biological Chemistry 262:16644–16654.
Onrust R, Finkelstein J, Turner J, Naktinis V, O’Donnell M. 1995. Assembly of a chromosomal replication
machine: two DNA polymerases, a clamp loader, and sliding clamps in one holoenzyme particle: III. Interface
between two polymerases and the clamp loader. Journal of Biological Chemistry 270:13366–13377. doi: 10.
1074/jbc.270.22.13366
Ozawa K, Horan NP, Robinson A, Yagi H, Hill FR, Jergic S, Xu Z-Q, Loscha KV, Li N, Tehei M, Oakley AJ, Otting
G, Huber T, Dixon NE. 2013. Proofreading exonuclease on a tether: the complex between the E. coli DNA
polymerase III subunits and alpha, epsilon, theta and beta reveals a highly flexible arrangement of the
proofreading domain. Nucleic Acids Research 41:5354–5367. doi: 10.1093/nar/gkt162
Rosenthal PB, Henderson R. 2003. Optimal determination of particle orientation, absolute hand, and contrast
loss in single-particle electron cryomicroscopy. Journal of Molecular Biology 333:721–745. doi: 10.1016/j.jmb.
2003.07.013
Scheres SHW, Chen S. 2012. Prevention of overfitting in cryo-EM structure determination. Nature Methods 9:
853–854. doi: 10.1038/nmeth.2115
Scheres SHW. 2012. RELION: implementation of a Bayesian approach to cryo-EM structure determination.
Journal of Structural Biology 180:519–530. doi: 10.1016/j.jsb.2012.09.006
Scheres SHW. 2014. Beam-induced motion correction for sub-megadalton cryo-EM particles. eLife 3:p.e03665.
doi: 10.7554/eLife.03665.009
Scheuermann R, Tam S, Burgers PM, Lu C, Echols H. 1983. Identification of the epsilon-subunit of escherichia coli
DNA polymerase III holoenzyme as the dnaQ gene product: a fidelity subunit for DNA replication. Proceedings
of the National Academy of Sciences of the United States of America 80:7085–7089. doi: 10.1073/pnas.80.23.
7085
Schrödinger LLC. 2010. The PyMOL molecular graphics system. Version 1.3r1
Schwartz JJ, Quake SR. 2009. Single molecule measurement of the "speed limit" of DNA polymerase.
Proceedings of the National Academy of Sciences of the United States of America 106:20294–20299. doi: 10.
1073/pnas.0907404106
Sharma A, Kottur J, Narayanan N, Nair DT. 2013. A strategically located serine residue is critical for the mutator
activity of DNA polymerase IV from Escherichia coli. Nucleic Acids Research 41:5104–5114. doi: 10.1093/nar/
gkt146
Simonetta KR, Kazmirski SL, Goedken ER, Cantor AJ, Kelch BA, Mcnally R, Seyedin SN, Makino DL, O’Donnell
M, Kuriyan J. 2009. The mechanism of ATP-dependent primer-template recognition by a clamp loader
complex. Cell 137:659–671. doi: 10.1016/j.cell.2009.03.044
Stukenberg PT, Studwell-Vaughan PS, O’Donnell M. 1991. Mechanism of the sliding beta-clamp of DNA
polymerase III holoenzyme. The Journal of Biological Chemistry 266:11328–11334.
Su X-C, Jergic S, Keniry MA, Dixon NE, Otting G. 2007. Solution structure of domains IVa and V of the tau
subunit of Escherichia coli DNA polymerase III and interaction with the alpha subunit. Nucleic Acids Research
35:2825–2832. doi: 10.1093/nar/gkm080
Tang G, Peng L, Baldwin PR, Mann DS, Jiang W, Rees I, Ludtke SJ. 2007. EMAN2: an extensible image
processing suite for electron microscopy. Journal of Structural Biology 157:38–46. doi: 10.1016/j.jsb.2006.05.
009
Toste Rêgo A, Holding AN, Kent H, Lamers MH. 2013. Architecture of the Pol III–clamp–exonuclease complex
reveals key roles of the exonuclease subunit in processive DNA synthesis and repair. The EMBO Journal 32:
1334–1343. doi: 10.1038/emboj.2013.68
Wang F, Yang W. 2009. Structural insight into translesion synthesis by DNA Pol II. Cell 139:1279–1289. doi: 10.
1016/j.cell.2009.11.043
Wang G, Klein MG, Tokonzaba E, Zhang Y, Holden LG, Chen XS. 2008. The structure of a DnaB-family replicative
helicase and its interactions with primase. Nature Structural Molecular Biology 15:94–100. doi: 10.1038/
nsmb1356
Wijffels G, Dalrymple BP, Prosselkov P, Kongsuwan K, Epa VC, Lilley PE, Jergic S, Buchardt J, Brown SE,
Alewood PF, Jennings PA, Dixon NE. 2004. Inhibition of protein interactions with the beta 2 sliding clamp of
Escherichia coli DNA polymerase III by peptides from beta 2 -binding proteins. Biochemistry 43:5661–5671.
doi: 10.1021/bi036229j
Wing RA, Bailey S, Steitz TA. 2008. Insights into the replisome from the structure of a ternary complex of the
DNA polymerase III a-subunit. Journal of Molecular Biology 382:859–869. doi: 10.1016/j.jmb.2008.07.058
Wu CA, Zechner EL, Marians KJ. 1992. Coordinated leading- and lagging-strand synthesis at the Escherichia coli
DNA replication fork. I. Multiple effectors act to modulate Okazaki fragment size. The Journal of Biological
Chemistry 267:4030–4044.

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 15 of 16

 Research article Biophysics and structural biology

Wu CA, Zechner EL, Reems JA, McHenry CS, Marians KJ. 1992. Coordinated leading- and lagging-strand
synthesis at the Escherichia coli DNA replication fork. V. primase action regulates the cycle of okazaki fragment
synthesis. The Journal of Biological Chemistry 267:4074–4083.
Yao NY, Georgescu RE, Finkelstein J, O’Donnell ME. 2009. Single-molecule analysis reveals that the lagging
strand increases replisome processivity but slows replication fork progression. Proceedings of the National
Academy of Sciences of the United States of America 106:13236–13241. doi: 10.1073/pnas.0906157106
Yuan Q, Mchenry CS. 2014. Cycling of the E. coli lagging strand polymerase is triggered exclusively by the
availability of a new primer at the replication fork. Nucleic Acids Research 42:1747–1756. doi: 10.1093/nar/
gkt1098

Fernandez-Leiro et al. eLife 2015;4:e11134. DOI: 10.7554/eLife.11134 16 of 16