Research Article

Cryo-EM structure of the replisome reveals multiple interactions coordinating DNA synthesis

Arkadiusz W. Kulczyk, Arne Moeller, Peter Meyer, Piotr Sliz, and Charles C. Richardson
PNAS March 7, 2017 114 (10) E1848-E1856; first published February 21, 2017 https://doi.org/10.1073/pnas.1701252114

  1. Contributed by Charles C. Richardson, January 27, 2017 (sent for review December 7, 2016; reviewed by James M. Berger and Nicholas E. Dixon)

Significance

The antiparallel nature of the two strands in duplex DNA poses a topological problem for their simultaneous synthesis. The “trombone” model of the replication fork postulates that the lagging-strand forms a loop such that the leading- and lagging-strand replication proteins contact one another. The replisome then can move in one direction along the DNA while synthesizing both strands. Physical interactions between the replication proteins and DNA coordinate processive synthesis of the leading and lagging strands. Here, we present the structure of a functional replisome from bacteriophage T7. Our structural and biochemical analyses provide an explanation of the mechanisms governing coordination of leading- and lagging-strand synthesis.

Abstract

We present a structure of the ∼650-kDa functional replisome of bacteriophage T7 assembled on DNA resembling a replication fork. A structure of the complex consisting of six domains of DNA helicase, five domains of RNA primase, two DNA polymerases, and two thioredoxin (processivity factor) molecules was determined by single-particle cryo-electron microscopy. The two molecules of DNA polymerase adopt a different spatial arrangement at the replication fork, reflecting their roles in leading- and lagging-strand synthesis. The structure, in combination with biochemical data, reveals molecular mechanisms for coordination of leading- and lagging-strand synthesis. Because mechanisms of DNA replication are highly conserved, the observations are relevant to other replication systems.

The reverse polarity and antiparallel nature of the two strands in duplex DNA pose a topological problem for their simultaneous synthesis. The “trombone” model of DNA replication postulates that the lagging strand forms a loop such that the leading- and lagging-strand replication proteins contact one another (1). This replication loop contains a nascent Okazaki fragment and allows for coordination of leading- and lagging-strand synthesis. The bacteriophage T7 replisome has been studied extensively (2). Thus, the macromolecular complex of T7 replication proteins provides an excellent model for structural analysis of a replisome. Notably, the human mitochondrial and the T7 replication systems are similar (3).

The phage T7 replisome is relatively simple. Four proteins are sufficient for reconstitution of the functional replisome, yet the assembled replisome recapitulates all of the key features of more complex prokaryotic and eukaryotic systems (2, 4). These proteins are the following: DNA polymerase (gp5) and its processivity factor, Escherichia coli thioredoxin (trx), single-stranded DNA (ssDNA)-binding protein (gp2.5), and the bifunctional DNA primase-helicase (gp4).

Gp4 forms multiple oligomeric forms (5). The negative stain EM revealed that hexamers and heptamers are the most abundant (∼22% and ∼78%, respectively). Upon addition of ssDNA, the equilibrium between the two oligomeric forms changes (∼77% protein rings are now hexameric, and ∼18% are heptameric), indicative of DNA binding (6). The hexamer is an enzymatically active form of gp4. It binds to ssDNA, hydrolyzes nucleotides, translocates on ssDNA, and unwinds dsDNA (7⇓–9). In contrast, the heptamer does not bind to ssDNA (6).

Crystal structures of all of the individual T7 replication proteins have been determined (10⇓⇓⇓⇓–15). However, to date there is no structural information on the T7 replisome or its subassemblies. Consequently, intermolecular interactions involved in coordination of DNA synthesis have not been identified. We have therefore used cryo-electron microscopy (cryo-EM) and single-particle reconstruction methods to determine a structure of the T7 replisome.

Results

Replisome Assembly.

We have identified conditions necessary to assemble the T7 replication complex. A stable complex was obtained with genetically altered gp4 and gp5, trx, and DNA resembling a replication fork in buffer containing dTTP (Fig. 1A). Gp4 contains an amino acid substitution, E343Q. Glutamate 343 functions as a catalytic base for hydrolysis of dTTP, an event that leads to translocation on DNA. A crystal structure of the hexameric gp4 reveals that E343 is in position to activate the water molecule for a nucleophilic attack on the nucleoside γ-phosphate (16). Consequently, this altered gp4 (gp4-E343Q) binds dTTP but does not catalyze hydrolysis on DNA to which it is tightly bound. The genetically altered gp5 contains two amino acid substitutions, D5A and E7A, in the active site of the 3′-5′ exonuclease domain. The altered protein exhibits a polymerization activity similar to the activity of wild-type DNA polymerase, but its exonuclease activity is reduced by a factor of 106 (10).

Fig. 1.
Fig. 1.

cryo-EM analysis and assembly of the replisome. (A) A fork-shaped DNA is formed by folding of ssDNA such that a stable tetra-loop is formed in the middle of the sequence flanked by a short double-helical region; two oligonucleotides are annealed to serve as primers. (B and C) Melting profiles of the tetra-loop DNA. Samples containing different concentrations of the fork-shaped DNA are represented by colored curved lines. Melting curves (B) and their first derivatives (C) show a biphasic transition with the melting temperatures (Tm) of 53 ± 2 °C. (D) Fluorescence anisotropy measurements of replisome formation. The replication complex assembled in the presence of gp4-E343Q is four times more stable than the complex assembled with wild-type gp4. Apparent binding constants are 0.2 ± 0.01 µM and 0.8 ± 0.05 µM, respectively. (E) A representative cryo-EM image. (Scale bar, 10 nm.) (F) The 2D class averages. (Scale bar, 10 nm.) (G) A calculated 3D map of the replisome. The initial model for the 3D calculation was generated using the atomic coordinates of the helicase domain of gp4 filtered to a resolution of 60 Å. (H) Fourier Schell Correlation (FSC). The resolution of 13.8 Å is based on the gold-standard FSC0.5 criterion. (I) Comparison of reprojections from the 3D structure with the 2D class averages. (Scale bar, 10 nm.)

The DNA consists of a 77-nt oligonucleotide containing two self-complementary regions that anneal to form a tetra-loop and a 5-bp double-helical region in the middle of the sequence with a 32-nt 5′-overhang and a 31-nt 3′-overhang (Fig. 1A). These two overhangs mimic the lagging and the leading strand, respectively. Although the sequence is capable of forming intermolecular polymers as well as the desired intramolecular monomeric species, the monomeric form is favored by inclusion of the tetra-loop sequence. The tetra-loop forms a highly stable turn conformation only in the monomeric form that we previously confirmed by 1D and 2D 1H-NMR (17). We assessed the folding properties of an oligonucleotide resembling a replication fork using fluorescence melting spectroscopy in the presence of a fluorescent dye that intercalates into double-stranded DNA. The decrease in fluorescence as a function of increasing temperature suggests folding of the template and formation of the duplex (Fig. 1B). The melting profiles show a biphasic transition indicative of formation of a single monomeric species in solution with a melting temperature of 53 ± 2 °C (Fig. 1C). In the control experiment, a similar DNA lacking a stable tetra-loop sequence does not exhibit a biphasic melting transition (Fig. 1B). The 5′-overhang of the folded template contains the primase recognition sequence 5′-GGGTC-3′. The lagging-strand primer is a tetraribonucleotide complementary to the primase recognition sequence, and it has an additional two nucleotides at the 3′-end that base-pair with the template: a deoxyribonucleotide at position 5 and a dideoxynucleotide at position 6. The dideoxynucleotide at the 3′-end of the primer prevents primer extension by gp5/trx. The leading-strand primer is a 21-nt oligonucleotide, also containing a dideoxynucleotide at the 3′-end (Fig. 1A).

Upon addition of gp5/trx and gp4-E343Q to the fork DNA in the presence of dTTP, a stable replication complex is formed as evidenced by an increase in fluorescence anisotropy of a fluorescein-labeled template. Gp4-E343Q binds to the replication complex assembled on the DNA with an affinity fourfold higher than that measured for binding of gp4 (0.2 ± 0.01 µM and 0.8 ± 0.05 µM, respectively), (Fig. 1D). We assessed the binding of individual replication proteins to the fork DNA in a filter-binding experiment, in which the leading-strand primer was radioactively labeled at the 5′-end. Dissociation constants determined for gp5/trx and gp4-E343Q are 0.07 ± 0.03 µM and 0.28 ± 0.05 µM, respectively (SI Appendix, Fig. S1). A stable replication complex of gp4-E343Q and gp5 is formed with Kd of 0.2 ± 0.04 µM, consistent with fluorescence anisotropy measurements. We have previously characterized gp4–gp5/trx–DNA interactions on the lagging and leading strand using a variety of biophysical methods (5, 18). The fork DNA was designed such that it combined DNA constructs used in these experiments.

Preparation of the Cryo-Electron Microscopy Sample and Determination of the Replisome Structure.

Detailed descriptions of the sample preparation protocols are included in Materials and Methods. First, the leading-strand primer and the template were annealed by incubation at 94 °C and by slowly decreasing the temperature to 25 °C. The 1.1-µM primer template was mixed with 6 µM gp4-E343Q in a buffer containing 5 µM lagging-strand primer, 10 mM dTTP, and 0.001% wt/vol n-dodecyl β-d-maltoside. Gp5/trx was added to the mixture after the 10-min incubation at 25 °C. Samples of the replisome were adsorbed onto freshly glow-discharged holey carbon grids and flash-frozen in liquid ethane using a Vitrobot with a controlled temperature and humidity. The protein and DNA concentrations are higher than the measured dissociation constants (Fig. 1D and SI Appendix, Fig. S1), and thus favorable conditions are created on the EM grid for complex formation. To confirm formation of the protein–DNA complex under these conditions, we performed a pull-down experiment with a biotin-labeled fork DNA. The 1 µM DNA was incubated with 2 µM gp4-E343Q and 10 µM gp5/trx in the buffer described above. Replisome samples were separated on streptavidin beads and analyzed by SDS/PAGE (SI Appendix, Fig. S2). The bands representing gp5/trx and gp4-E343Q on the gel confirm protein binding to the biotinylated fork DNA.

The cryo-EM images (Fig. 1E) were collected with a Titan Krios cryo-EM. Subsequent structure calculation and refinement steps were performed using single- and double-precision floating-point compilations of Relion-1.4. We used software and computational infrastructure maintained by SBgrid (19), as well as high-performance computing facilities available through the Stampede Supercomputer at Texas Advanced Computing Center (20).

The initial cryo-EM analysis revealed that ∼95% of helicase rings are hexameric and ∼5% are heptameric. Thus, the initial model for the 3D calculation was generated using the atomic coordinates of a hexameric helicase domain of gp4 [Protein Data Bank (PDB) ID code 1CR0] filtered to a resolution of 60 Å (Fig. 1G). Fig. 1F displays representative 2D class averages. We clearly see additional densities associated with the hexameric ring of DNA helicase likely representing DNA polymerases bound to the replisome. To confirm that these densities are gp5/trx complexes, we labeled thioredoxin with 5 nm Ni-NTA Nanogold (SI Appendix, Fig. S3). Thioredoxin was altered by addition of a poly-His tag at the N terminus. Gp5-trx complex was purified on Ni-NTA agarose (SI Appendix, Fig. S3A) and used to prepare samples for EM. Negative-stain EM reveals gold particles colocalizing with the replisome, confirming the presence of gp5/trx in the complex (SI Appendix, Fig. S3 B and C). We see one or two gold particles associated with the complex (SI Appendix, Fig. S3C). However, this labeling with gold is not quantitative, and gold particles are observed that are not attached to the complex.

We also analyzed samples of the replisome in the absence of DNA using negative-stain EM. When DNA is not present, no binding between gp4-E343Q and gp5/trx is detected, indicating that DNA is required for assembly of the functional gp4–E343Q–gp5/trx complex. In the absence of DNA, ∼70% helicase rings are heptameric, and ∼20% are hexameric, consistent with previous negative-stain EM analysis (5⇓⇓–8), native gel electrophoresis (5, 6, 21), and gel-filtration data (9).

Structural classification with RELION resulted in two 3D maps. One of these maps was then used for further refinements (Fig. 1G). The final 3D map was calculated using ∼50% (79,519) particle images. The presented structure is thus representative of the entire dataset. The angular distribution plot is presented in SI Appendix, Fig. S5. A reported resolution of 13.8 Å is based on the gold-standard 0.5FSC criterion (Fig. 1H). The reprojections from the structure are in good agreement with the 2D class averages (Fig. 1I and SI Appendix, Fig. S6).

Well-defined electron densities allow for unambiguous fitting of two molecules of gp5/trx (PDB ID code 1T7P), six subunits of DNA helicase (PDB ID code 1CR0), and five of six subunits of RNA primase (PDB ID code 1NUI). Automatic docking of atomic coordinates into the cryo-EM map was performed with the Chimera software using the protocol described in SI Appendix. The cross-correlation coefficient (CCC) calculated for simultaneous docking of all PDBs (i.e., two copies of PDB ID code 1T7P, five copies of PDB ID code 1NUI, and PDB ID code 1CR0) is 0.91. The CCCs calculated for individual protein domains range from 0.83 to 0.99 and are presented in SI Appendix, Fig. S7. The local CCCs mapped onto the cryo-EM structure are presented in SI Appendix, Fig. S8.

To determine the handedness in the map of the replisome, we performed an automatic docking of atomic coordinates into a mirror copy of the map across the “y–z” plane. The local CCCs calculated for a chiral copy of the replisome map are shown in SI Appendix, Fig. S8. The model presented in this article represents an enantiomer displaying a significantly better fit than its chiral counterpart (SI Appendix, Fig. S8). We also performed a random canonical tilt experiment (SI Appendix, Fig. S9).

Architecture of the Replisome.

An overall architecture of the T7 replisome is presented in Fig. 2A and Movie S1. The structure reveals the presence of an asymmetric hexameric ring of DNA helicase to which two DNA polymerases bind in an asymmetric fashion. Likewise, the associated DNA primase domains of gp4 are arranged asymmetrically (Fig. 2B). Two molecules of gp5/trx are bound to the hexameric ring of gp4 at the back of the replisome relative to the direction of dsDNA unwinding by the helicase. Although DNA is not visible, the unique structures obtained suggest strongly its presence (Discussion).

Fig. 2.
Fig. 2.

Architecture of the replisome. (A) The structure reveals the presence of the hexameric ring of DNA helicase (blue), five covalently linked DNA primase domains (orange), and two gp5/trx complexes. The polymerase domain (green) and the exonuclease domain (purple) of gp5 as well as thioredoxin (yellow) are indicated. The two molecules of DNA polymerase bind to gp4 in an asymmetric fashion, reflecting their roles in leading- and lagging-strand synthesis (gp5/trxlead and gp5/trxlag, respectively). (B) An asymmetric arrangement of RNA primases at the replisome.

Asymmetric Positioning of DNA Polymerases at the Replisome.

The cryo-EM structure provides evidence for the asymmetric positioning of the two DNA polymerases at the replisome. The two molecules of gp5/trx can be implicated in leading- and lagging-strand synthesis [gp5/trxlead and gp5/trxlag, respectively (Fig. 2B)]. The interaction between gp5/trx and gp4 differs on the lagging and leading strand. We previously reconstituted the gp4–gp5/trxlead (5) and gp4–gp5/trxlag (18) complexes and characterized both interactions. The biochemical analysis is consistent with the replisome structure, and thus we can determine the location of gp5/trxlead and gp5/trxlag (Fig. 2). A low-resolution structure of the gp4–gp5/trxlead complex has been determined by small-angle X-ray scattering (SAXS) (5). The position of gp4-gp5/trxlead in the SAXS structure is consistent with the position of one of the two gp5/trx molecules from the cryo-EM structure, confirming the assignment of gp5/trxlead.

Interaction of the Leading-Strand DNA Polymerase with DNA Primase-Helicase.

Gp5/trxlead makes contacts with the two neighboring subunits of gp4 through the fingers and the exonuclease domains. The α-helix P and the region containing a loop connecting the α-helix P with the β-strand 10 (residues 566–586) from the fingers domain interact with a region of DNA helicase in close proximity to the α-helix (residues 402–416) (Fig. 3A). Residues 566–586 are missing in the crystal structure of the gp5/trx primer template (10), suggesting flexibility in this region. Interestingly, the loading patch of gp5 containing the basic amino acids K587, K589, R590, and R591 is located in close proximity to this region (SI Appendix, Fig. S10). We previously showed that this basic loading patch of gp5 interacts with the C-terminal acidic tail of gp4 and is essential for loading gp5 at the replication fork (22). The second interaction involves the primase domain from the adjacent subunit of gp4 and the region containing the α-helix E* from the exonuclease domain of gp5/trxlead. The α-helix E* is unique for gp5/trx in the Pol I family of DNA polymerases.

Fig. 3.
Fig. 3.

Intermolecular interactions at the replisome. (A) Interaction of gp5/trxlead and gp4. (B) The fingers of gp5/trxlead interact with the loop connecting α-helices H1 and H2 in the thumb domain. (C) Gp5/trxlead adopts a closed conformation of the polymerase active site, in which the thumb and fingers are rotated inward. (D) In contrast, the polymearse active site of gp5/trxlag adopts an open conformation. In this open conformation, the fingers rotate outward away from the thumb. (E) Interaction between gp5/trxlag and gp4. (F) Trx binds to a unique 76-amino-acid loop, TBD, inserted in the thumb subdomain of gp5/trxlag. (G) Binding of gp5/trxlead and gp5/trxlag. The specific interactions are described in the text.

In the crystal structure of the gp5/trx primer template, trx binds to a unique 76-amino-acid loop, the thioredoxin-binding domain (TBD), inserted in the thumb subdomain. The cryo-EM structure reveals that trx contacts the TBD in an extended conformation. The electron density is slightly larger than a contour level of trx (Fig. 2A), suggesting flexibility in this region, consistent with SAXS analysis of the trx–TBD interaction (23). A loop connecting α-helices H1 and H2 in the thumb domain makes contacts with the tip of the fingers, an interaction not observed in the crystal structure (Fig. 3B). In summary, gp5/trxlead makes numerous interactions with gp4. The thumb and fingers of gp5/trxlead adopt a conformation similar to a closed conformation of the polymerase active site observed upon binding of gp5/trx to a primer template in the presence of an incoming nucleotide (Fig. 3 B and C) (10).

Interaction of the Lagging-Strand DNA Polymerase with DNA Primase-Helicase.

Gp5/trxlag makes extensive contacts with two primase domains from adjacent subunits in gp4 through the fingers and the exonuclease domain (Fig. 3D). However, in contrast to gp5/trxlead, the polymearse active site of gp5/trxlag adopts an open conformation (Fig. 3E). In this open conformation, the fingers rotate outward away from the polymerase active site. The observed open conformation is consistent with the biochemical data indicating that DNA primers ≥21 nt are necessary for closure of the fingers (23) and supports the observation that the presence of the primase is required for stabilization of shorter primers in the polymerase active site of gp5 (18). Interestingly, a loop connecting the α-helix P with the β-strand 10 and α-helices E* and F from the exonuclease domain is also located in this interface, indicating that gp5/trxlag uses similar structural motifs to contact gp4 as does gp5/trxlead. The TBD of gp5/trxlag displays an extended conformation. As in gp5/trxlead, a slightly larger electron density than a contour level of trx indicates flexibility in this region (Fig. 3F).

The Interaction Between the Leading- and Lagging-Strand DNA Polymerase.

Gp5/trxlead and gp5/trxlag contact each other at the replisome. The protein–protein interface is formed by a region containing the α-helices G, G1, and R from the palm domain of gp5/trxlead and by the tip of the fingers domain of gp5/trxlag (Fig. 3G). Taken together, the 3D map of the ∼650-kDa replisome permits an unambiguous fitting of the atomic coordinates of gp4, gp5/trxlead, and gp5/trxlag.

Discussion

DNA in the Replisome Structure.

Although we do not directly observe the density representing DNA in the map of the replisome, the conformation of the replication proteins (gp4-E343Q, gp5/trxlead) is compatible with the one observed in the DNA-bound state (5⇓–7, 9, 10).

To visualize DNA in the replisome samples, we labeled a biotinylated fork DNA with quantum dots (QD) and visualized protein–DNA complexes by negative-stain EM. A biotin is attached to the tetra-loop sequence in the DNA template (Fig. 1A). The DNA-QD conjugates were purified using magnetic beads coated with diethylaminoethyl cellulose (SI Appendix, Fig. S4A). A micrograph presented in SI Appendix, Fig. S4B shows QD-DNA conjugates that were negatively stained with uranyl formate in the absence of the replication proteins. Mixing of gp4 and gp5/trx with QD-DNA results in formation of protein complexes that colocalize with QDs, indicative of DNA binding (SI Appendix, Fig. S4 C and D). This method is not quantitative, and QD-DNA is also observed distant from the replisome.

DNA binding by the replisome is also supported by the following observations: (i) Gp4 in the structure is hexameric. Gp4 exists in multiple oligomeric forms, but only hexamers bind to ssDNA (5⇓–7, 9, 21). In the absence of ssDNA, gp4 is predominantly heptameric (6). (ii) No binding between gp4 and gp5/trx can be detected by negative-stain EM in the absence of DNA. (iii) We demonstrated DNA binding by the replisome using fluorescence anisotropy measurements with a fluorescein-labeled DNA (Fig. 1D), filter-binding experiments with P32-labeled DNA (SI Appendix, Fig. S1), and a pull-down assay with a biotinylated fork DNA (SI Appendix, Fig. S2). In addition, the cryo-EM map of the replisome reveals a density located in the channel protruding through the center of the hexameric ring of DNA helicase that may represent DNA (Fig. 2A). Previous EM and biochemical data indicate that ssDNA binds to gp4 in this channel (5, 8, 14). The lack of the density representing DNA implies conformational heterogeneity in the sample. For example, it is likely that ssDNA can be bound within the central channel of gp4 by different subsets of the six subunits of the helicase. In support of this notion, it is difficult to visualize the helicase–DNA complexes by cryo-EM even at high resolution (24).

While our manuscript was under review, an article was published describing a low-resolution X-ray structure (4.8 Å) of the gp4–gp5/trx complex in the absence of DNA (25). Consistent with our data, gp4 in this structure is heptameric. Three molecules of gp5/trx are associated with gp4 through the C-terminal tails of DNA helicase; the interaction described earlier used single-molecule methods (26, 27) and SAXS (5).

Architecture of the Replication Fork.

Upon analysis of the T7 replisome structure, in combination with the extensive biochemical and mutational studies reviewed in refs. 2 and 4, we propose a functional model of the replication fork (Fig. 4A). DNA helicase contacts dsDNA and separates dsDNA into ssDNA. The two ssDNAs are bound by gp5/trxlead and gp5/trxlag. Both DNA polymerases are positioned behind gp4 relative to the direction of dsDNA unwinding.

Fig. 4.
Fig. 4.

Architecture of the replication fork and a model for formation of the lagging-strand replication loop. (A) A model of the replication fork before formation of the replication loop on the lagging strand. (B) Polymerization of nucleotides by gp5/trxlag combined with the movement of ssDNA through the central cavity of gp4 accounts for formation of the replication loop on the lagging strand.

The lagging-strand ssDNA passes through the central channel in the hexameric ring of DNA helicase (5, 8). The crystal structure of the hexameric gp4 reveals the presence of three loops protruding into the central cavity (14). These loops contain a number of basic residues involved in DNA binding (16). Hydrolysis of dTTP by gp4 is coupled with a sequential transfer of ssDNA by the DNA-binding loops from one subunit to the adjacent subunit within the central channel (14, 16). ssDNA exiting from the channel has the correct polarity for binding by the primase and the DNA-binding cleft of gp5/trxlag.

Displacement of the leading-strand ssDNA by the C-terminal acidic surface of gp4 allows for binding of ssDNA by the positively charged active site of gp5/trxlead. In the model, a primer template bound to gp5/trxlead (PDB #1T7P) is located such that the extension of the 5′-end of the template would position it in close proximity to the subunit interface β-hairpin containing residue F523 of gp4 (SI Appendix, Fig. S11). We showed earlier that F523 interacts directly with the extruded DNA leading strand (21). The orientation of gp5/trxlead at the replisome helps to stabilize partially unwound dsDNA at the junction between ssDNA and dsDNA, thus helping to separate dsDNA. Additionally, gp5/trxlead binding to gp4 can change the equilibrium between the forward and the backward Brownian sliding of DNA helicase along ssDNA (28, 29), favoring the movement of gp4 in the 5′-3′ direction or sliding of the lagging-strand ssDNA through the central channel of gp4 in the 3′-5′ direction.

The surface of the replisome is negatively charged with the exception of regions implicated in DNA binding (SI Appendix, Fig. S10). These regions have strong electropositive potential, namely, the central channel in the hexameric ring of gp4, gp5/trx active sites, loading and exchange patches located on gp5/trx, the primase domains, and trx. All these regions make contacts with DNA in the model.

Our model differs from the recently proposed model for the Saccharomyces cerevisiae replication fork based on a low-resolution cryo-EM map (30) of the complex between CMG helicase and the leading-strand DNA polymerase epsilon (Pol ε). The model proposed by Sun et al. (30) suggests that Pol ε is positioned ahead of the replisome, possibly functioning as the prow for unwinding of dsDNA. In contrast, in our model gp5/trxlead is positioned behind the replisome relative to the direction of dsDNA unwinding. Inconsistencies in the two models may reflect differences between prokaryotic and eukaryotic replication systems.

Formation of the Lagging-Strand Replication Loop.

The loop on the lagging strand implies changes in the orientation of gp5/trxlag. In this new orientation, both leading- and lagging-strand synthesis occur at a single point and proceed in the same direction. How the replication loop is formed is not known. The geometry of binding between gp5/trxlag and gp4 in the cryo-EM structure, however, does suggest a mechanism. The interaction of gp5/trxlag with the primase domains of gp4 positions gp5/trxlag such that its polymerization activity pushes the newly synthesized dsDNA toward the central channel of gp4 (Fig. 4B). Consequently, both the polymerization activity of gp5/trxlag combined with the movement of ssDNA through the central cavity of the helicase account for formation of the replication loop on the lagging strand.

Coordination of Leading- and Lagging-Strand Synthesis.

Coordination of leading- and lagging-strand synthesis is critical for high processivity of the replisome (5, 31⇓⇓⇓⇓–36). However, the enzymatic events on the lagging and leading strands are not equivalent. The lagging-strand DNA polymerase repeatedly dissociates from DNA and recycles to a new primer upon completion of each Okazaki fragment. Furthermore, one of the slowest events on the lagging strand is the recognition of the primase recognition sequence, synthesis of oligoribonucleotides by the primase, and transfer to gp5/trxlag, which, in the time required for these events, would lead to accumulation of approximately 1,000 nucleotides of leading-strand DNA. Why does the synthesis of the leading strand not outpace the synthesis of the lagging strand?

Two models have been proposed to achieve coordination of leading- and lagging-strand synthesis. In the first model, leading-strand synthesis transiently stops allowing the completion of the slow enzymatic events on the lagging strand. The 12-s pause allows for the relatively slow process of primer synthesis (31, 32). In the second model, coordination is achieved by different synthesis rates of the leading and the lagging strands. Contrary to the results described by Lee et al. (32), the studies of the T7 replisome by Pandey et al. (37) indicated that lagging-strand DNA synthesis is 38% faster than leading-strand DNA synthesis. In addition, no pausing of replisome movement was detected during primer synthesis.

The different binding modes of gp5/trxlag and gp5/trxlead to gp4 are compatible with the two polymerases synthesizing DNA at different rates. The geometry of binding between gp5/trxlead and gp4 indicates that the rate of leading-strand synthesis is limited by the rate of dsDNA unwinding by the helicase (28). DNA synthesis by gp5/trxlag is not limited by its binding to gp4, but rather by the availability of ssDNA from the replication loop (Fig. 4B).

A direct interaction between gp5/trxlag and gp5/trxlead at the replisome provides a mechanism for coordination of the leading- and lagging-strand synthesis. Conformational changes in a structure of DNA polymerase (i.e., transitions between open and closed conformations of the polymerase active site) resulting from dissociation/association to the primer template offer a sensing mechanism, in which changes on one strand are transmitted to the other strand via a protein–protein interaction. For example, an inward rotation of the fingers in gp5/trxlag resulting from the engagement of the primer by gp5/trxlag, and a subsequent extension of the primer to an Okazaki fragment, would lead to disruption of the contact between gp5/trxlag and gp5/trxlead. However, upon completion of the synthesis of an Okazaki fragment, gp5/trxlag disengages from the primer, an event manifested by the outward rotation of the fingers. The fingers of gp5/trxlag are then in contact with gp5/trxlead, providing an interaction that could modulate leading-strand synthesis (Fig. 3G). The geometry of binding between gp5/trxlead and gp5/trxlag excludes a possibility to form a similar interaction with a third copy of gp5/trx at the replication fork, explaining why only two DNA polymerases can be actively engaged in a coordinated DNA synthesis. Interestingly, T4 DNA polymerase forms dimers. The fingers of T4 DNA polymerase are also involved in mediating the protein–protein interaction (38), suggesting an analogous mechanism for coordination of DNA synthesis in the phage T4 replication system.

Although absent in the replisome structure, the ssDNA-binding protein (gp2.5) encoded by T7 is essential for coordination of the synthesis of the two strands (39). In its absence the resulting Okazaki fragments are considerably shorter than those synthesized in its presence (39). Gp2.5 stimulates activities of gp4 and gp5 during coordinated synthesis of DNA (39). Gp2.5 increases the frequency of initiation of lagging-strand synthesis greater than 10-fold (4). The acidic 26-residue C-terminal tail of gp2.5 interacts with the basic loading patch of gp5/trxlag (26), the region where the acidic C-terminal tail of DNA helicase also binds. A competition between these two overlapping interactions during coordinated synthesis provides a switching mechanism to release gp5/trxlead from the replisome. Gp2.5 binding to the basic loading patch of gp5/trxlag, located in the hinge region connecting the palm domain and the fingers, may affect the inward and outward rotation of the fingers. Consequently, gp2.5 can modulate binding of the fingers of gp5/trxlag with the palm domain of gp5/trxlead, the interaction important for coordination of leading- and lagging-strand synthesis. Because the basic loading patch of gp5/trxlag is located next to the region involved in stabilization of gp5/trxlag on the replisome (SI Appendix, Fig. S10), gp2.5 binding could also regulate release of the replication loop.

In summary, the structure of the replisome reveals multiple interactions between DNA polymerase, DNA helicase, and RNA primase. Spatial topology of binding between the individual replication proteins provides a structural framework that brings together multiple enzymatic activities that orchestrate coordinated DNA synthesis.

Materials and Methods

A more detailed description of materials and methods used in this work can be found in SI Appendix, Supplementary Materials and Methods.

Recombinant Proteins.

Proteins were overproduced and purified as described previously: wild-type gp4 (40), gp4-E343Q (16), gp5/trx complex in which two amino acids in the exonuclease domain of gp5 are replaced with alanine: D5A and E7A (10).

cryo-EM.

Samples of the replisome were adsorbed onto freshly glow-discharged holey carbon grids (Quantifoil R1.2/1.3, Cu 400 mesh) and flash-frozen in liquid ethane using a Vitrobot with a controlled temperature and humidity. Data were acquired with a Titan Krios cryo-transmission electron microscope (FEI) at Aarhus University, operating at 300 kV and a dose of <20 e− Å-2 using the program for the automated data collection Leginon. Images were collected using a GATAN 4k × 4k pixel CCD camera. A total of 3,995 images were recorded at a defocus of 2–3 µm and a magnification of 50,000 times corresponding to 1.19-Å pixel size. Contrast transfer function parameters were estimated with ctffind4. All subsequent data-processing steps were performed using single- and double-precision floating-point compilations of Relion-1.4 (41).

Acknowledgments

We thank Aarhus University (Aarhus, Denmark) for access to the EM facilities; Suavek Rucinski for assistance with Appion; and Steven Moskowitz for assistance in preparation of figures. This work was supported by Grant MCB150101 from Extreme Science and Engineering Discovery Environment (XSEDE) (to A.W.K.); Grant DFF Mobilex 4093-00238B (to A.M.); Grant R01CA163647 from the National Cancer Institute (to P.S.); and Grant MCB150136 from XSEDE (to A.W.K., P.S., and C.C.R.). This work used the facilities of the XSEDE, which is supported by National Science Foundation Grant ACI-1053575. Some of the work presented here was conducted at the National Resource for Automated Molecular Microscopy at the Scripps Institute, San Diego, which is supported by grants from the National Center for Research Resources (2P41RR017573) and the National Institute of General Medical Sciences (9 P41 GM103310) of the National Institutes of Health.

Footnotes

  • ↵1To whom correspondence may be addressed. Email: arek{at}hms.harvard.edu or ccr{at}hms.harvard.edu.

  • ↵2Present address: Department of Structural Biology, Max Planck Institute of Biophysics, 60538 Frankfurt, Germany.

  • Author contributions: A.W.K. and C.C.R. designed research; A.W.K. performed research; A.W.K., A.M., P.M., and P.S. contributed new reagents/analytic tools; A.W.K. and C.C.R. analyzed data; and A.W.K. and C.C.R. wrote the paper.

  • Reviewers: J.M.B., Johns Hopkins University School of Medicine; and N.E.D., University of Wollongong.

  • The authors declare no conflict of interest.

  • Data deposition: The cryo-EM map of the replisome was deposited in the EM Data Bank, www.emdatabank.org (accession no. EMD-8565).

  • This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1701252114/-/DCSupplemental.

References

  1. ↵
    1. Sinha NK,
    2. Morris CF,
    3. Alberts BM
    (1980) Efficient in vitro replication of double-stranded DNA templates by a purified T4 bacteriophage replication system. J Biol Chem 255(9):4290–4293.
    .
    OpenUrlAbstract/FREE Full Text
  2. ↵
    1. Kulczyk AW,
    2. Richardson CC
    (2016) The replication system of bacteriophage T7. Enzymes 39:89–136.
    .
    OpenUrl
  3. ↵
    1. Ciesielski GL,
    2. Oliveira MT,
    3. Kaguni LS
    (2016) Animal mitochondrial DNA replication. Enzymes 39:255–292.
    .
    OpenUrl
  4. ↵
    1. Hamdan SM,
    2. Richardson CC
    (2009) Motors, switches, and contacts in the replisome. Annu Rev Biochem 78:205–243.
    .
    OpenUrlCrossRefPubMed
  5. ↵
    1. Kulczyk AW, et al.
    (2012) An interaction between DNA polymerase and helicase is essential for the high processivity of the bacteriophage T7 replisome. J Biol Chem 287(46):39050–39060.
    .
    OpenUrlAbstract/FREE Full Text
  6. ↵
    1. Crampton DJ,
    2. Ohi M,
    3. Qimron U,
    4. Walz T,
    5. Richardson CC
    (2006) Oligomeric states of bacteriophage T7 gene 4 primase/helicase. J Mol Biol 360(3):667–677.
    .
    OpenUrlCrossRefPubMed
  7. ↵
    1. Egelman EH,
    2. Yu X,
    3. Wild R,
    4. Hingorani MM,
    5. Patel SS
    (1995) Bacteriophage T7 helicase/primase proteins form rings around single-stranded DNA that suggest a general structure for hexameric helicases. Proc Natl Acad Sci USA 92(9):3869–3873.
    .
    OpenUrlAbstract/FREE Full Text
  8. ↵
    1. Yu X,
    2. Hingorani MM,
    3. Patel SS,
    4. Egelman EH
    (1996) DNA is bound within the central hole to one or two of the six subunits of the T7 DNA helicase. Nat Struct Biol 3(9):740–743.
    .
    OpenUrlCrossRefPubMed
  9. ↵
    1. Patel SS,
    2. Hingorani MM
    (1993) Oligomeric structure of bacteriophage T7 DNA primase/helicase proteins. J Biol Chem 268(14):10668–10675.
    .
    OpenUrlAbstract/FREE Full Text
  10. ↵
    1. Doublié S,
    2. Tabor S,
    3. Long AM,
    4. Richardson CC,
    5. Ellenberger T
    (1998) Crystal structure of a bacteriophage T7 DNA replication complex at 2.2 A resolution. Nature 391(6664):251–258.
    .
    OpenUrlCrossRefPubMed
  11. ↵
    1. Hollis T,
    2. Stattel JM,
    3. Walther DS,
    4. Richardson CC,
    5. Ellenberger T
    (2001) Structure of the gene 2.5 protein, a single-stranded DNA binding protein encoded by bacteriophage T7. Proc Natl Acad Sci USA 98(17):9557–9562.
    .
    OpenUrlAbstract/FREE Full Text
  12. ↵
    1. Kato M,
    2. Ito T,
    3. Wagner G,
    4. Richardson CC,
    5. Ellenberger T
    (2003) Modular architecture of the bacteriophage T7 primase couples RNA primer synthesis to DNA synthesis. Mol Cell 11(5):1349–1360.
    .
    OpenUrlCrossRefPubMed
  13. ↵
    1. Sawaya MR,
    2. Guo S,
    3. Tabor S,
    4. Richardson CC,
    5. Ellenberger T
    (1999) Crystal structure of the helicase domain from the replicative helicase-primase of bacteriophage T7. Cell 99(2):167–177.
    .
    OpenUrlCrossRefPubMed
  14. ↵
    1. Singleton MR,
    2. Sawaya MR,
    3. Ellenberger T,
    4. Wigley DB
    (2000) Crystal structure of T7 gene 4 ring helicase indicates a mechanism for sequential hydrolysis of nucleotides. Cell 101(6):589–600.
    .
    OpenUrlCrossRefPubMed
  15. ↵
    1. Toth EA,
    2. Li Y,
    3. Sawaya MR,
    4. Cheng Y,
    5. Ellenberger T
    (2003) The crystal structure of the bifunctional primase-helicase of bacteriophage T7. Mol Cell 12(5):1113–1123.
    .
    OpenUrlCrossRefPubMed
  16. ↵
    1. Crampton DJ,
    2. Mukherjee S,
    3. Richardson CC
    (2006) DNA-induced switch from independent to sequential dTTP hydrolysis in the bacteriophage T7 DNA helicase. Mol Cell 21(2):165–174.
    .
    OpenUrlCrossRefPubMed
  17. ↵
    1. Kulczyk AW,
    2. Yang JC,
    3. Neuhaus D
    (2004) Solution structure and DNA binding of the zinc-finger domain from DNA ligase IIIalpha. J Mol Biol 341(3):723–738.
    .
    OpenUrlCrossRefPubMed
  18. ↵
    1. Kulczyk AW,
    2. Richardson CC
    (2012) Molecular interactions in the priming complex of bacteriophage T7. Proc Natl Acad Sci USA 109(24):9408–9413.
    .
    OpenUrlAbstract/FREE Full Text
  19. ↵
    1. Morin A, et al.
    (2013) Collaboration gets the most out of software. eLife 2:e01456.
    .
    OpenUrlAbstract/FREE Full Text
  20. ↵
    1. Towns JCT, et al.
    (2014) XSEDE: Accelerating scientific discovery. Comput Sci Eng 16(5):62–74.
    .
    OpenUrlCrossRef
  21. ↵
    1. Satapathy AK,
    2. Kulczyk AW,
    3. Ghosh S,
    4. van Oijen AM,
    5. Richardson CC
    (2011) Coupling dTTP hydrolysis with DNA unwinding by the DNA helicase of bacteriophage T7. J Biol Chem 286(39):34468–34478.
    .
    OpenUrlAbstract/FREE Full Text
  22. ↵
    1. Zhang H, et al.
    (2011) Helicase-DNA polymerase interaction is critical to initiate leading-strand DNA synthesis. Proc Natl Acad Sci USA 108(23):9372–9377.
    .
    OpenUrlAbstract/FREE Full Text
  23. ↵
    1. Akabayov B, et al.
    (2010) Conformational dynamics of bacteriophage T7 DNA polymerase and its processivity factor, Escherichia coli thioredoxin. Proc Natl Acad Sci USA 107(34):15033–15038.
    .
    OpenUrlAbstract/FREE Full Text
  24. ↵
    1. Abid Ali F, et al.
    (2016) Cryo-EM structures of the eukaryotic replicative helicase bound to a translocation substrate. Nat Commun 7:10708.
    .
    OpenUrlCrossRefPubMed
  25. ↵
    1. Wallen JR, et al.
    (2017) Hybrid methods reveal multiple flexibly linked DNA polymerases within the bacteriophage T7 replisome. Structure 25(1):157–166.
    .
    OpenUrl
  26. ↵
    1. Hamdan SM, et al.
    (2007) Dynamic DNA helicase-DNA polymerase interactions assure processive replication fork movement. Mol Cell 27(4):539–549.
    .
    OpenUrlCrossRefPubMed
  27. ↵
    1. Loparo JJ,
    2. Kulczyk AW,
    3. Richardson CC,
    4. van Oijen AM
    (2011) Simultaneous single-molecule measurements of phage T7 replisome composition and function reveal the mechanism of polymerase exchange. Proc Natl Acad Sci USA 108(9):3584–3589.
    .
    OpenUrlAbstract/FREE Full Text
  28. ↵
    1. Stano NM, et al.
    (2005) DNA synthesis provides the driving force to accelerate DNA unwinding by a helicase. Nature 435(7040):370–373.
    .
    OpenUrlCrossRefPubMed
  29. ↵
    1. Pandey M,
    2. Patel SS
    (2014) Helicase and polymerase move together close to the fork junction and copy DNA in one-nucleotide steps. Cell Reports 6(6):1129–1138.
    .
    OpenUrl
  30. ↵
    1. Sun J, et al.
    (2015) The architecture of a eukaryotic replisome. Nat Struct Mol Biol 22(12):976–982.
    .
    OpenUrlCrossRefPubMed
  31. ↵
    1. Hamdan SM,
    2. Loparo JJ,
    3. Takahashi M,
    4. Richardson CC,
    5. van Oijen AM
    (2009) Dynamics of DNA replication loops reveal temporal control of lagging-strand synthesis. Nature 457(7227):336–339.
    .
    OpenUrlCrossRefPubMed
  32. ↵
    1. Lee JB, et al.
    (2006) DNA primase acts as a molecular brake in DNA replication. Nature 439(7076):621–624.
    .
    OpenUrlCrossRefPubMed
  33. ↵
    1. Hernandez AJ,
    2. Lee SJ,
    3. Richardson CC
    (2016) Primer release is the rate-limiting event in lagging-strand synthesis mediated by the T7 replisome. Proc Natl Acad Sci USA 113(21):5916–5921.
    .
    OpenUrlAbstract/FREE Full Text
  34. ↵
    1. Duderstadt KE, et al.
    (2016) Simultaneous real-time imaging of leading and lagging strand synthesis reveals the coordination dynamics of single replisomes. Mol Cell 64(6):1035–1047.
    .
    OpenUrl
  35. ↵
    1. Pandey M, et al.
    (2015) Two mechanisms coordinate replication termination by the Escherichia coli Tus-Ter complex. Nucleic Acids Res 43(12):5924–5935.
    .
    OpenUrlAbstract/FREE Full Text
  36. ↵
    1. Geertsema HJ,
    2. Kulczyk AW,
    3. Richardson CC,
    4. van Oijen AM
    (2014) Single-molecule studies of polymerase dynamics and stoichiometry at the bacteriophage T7 replication machinery. Proc Natl Acad Sci USA 111(11):4073–4078.
    .
    OpenUrlAbstract/FREE Full Text
  37. ↵
    1. Pandey M, et al.
    (2009) Coordinating DNA replication by means of priming loop and differential synthesis rate. Nature 462(7275):940–943.
    .
    OpenUrlCrossRefPubMed
  38. ↵
    1. Salinas F,
    2. Benkovic SJ
    (2000) Characterization of bacteriophage T4-coordinated leading- and lagging-strand synthesis on a minicircle substrate. Proc Natl Acad Sci USA 97(13):7196–7201.
    .
    OpenUrlAbstract/FREE Full Text
  39. ↵
    1. Lee J,
    2. Chastain PD II,
    3. Kusakabe T,
    4. Griffith JD,
    5. Richardson CC
    (1998) Coordinated leading and lagging strand DNA synthesis on a minicircular template. Mol Cell 1(7):1001–1010.
    .
    OpenUrlCrossRefPubMed
  40. ↵
    1. Notarnicola SM,
    2. Richardson CC
    (1993) The nucleotide binding site of the helicase/primase of bacteriophage T7. Interaction of mutant and wild-type proteins. J Biol Chem 268(36):27198–27207.
    .
    OpenUrlAbstract/FREE Full Text
  41. ↵
    1. Scheres SH
    (2012) RELION: Implementation of a Bayesian approach to cryo-EM structure determination. J Struct Biol 180(3):519–530.
    .
    OpenUrlCrossRefPubMed
View Abstract
Back to top
Article Alerts
Email Article
Citation Tools
Request Permissions
Cryo-EM structure of the replisome
Arkadiusz W. Kulczyk, Arne Moeller, Peter Meyer, Piotr Sliz, Charles C. Richardson
Proceedings of the National Academy of Sciences Mar 2017, 114 (10) E1848-E1856; DOI: 10.1073/pnas.1701252114
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo
Proceedings of the National Academy of Sciences: 114 (10)
Table of Contents

Submit

Article Classifications

Jump to section

You May Also be Interested in

Thus far, researchers have mostly just monitored the body’s microbiome. Now they’re starting to modify it to treat disease. Image credit: Science Source/Paul Gunning.
News Feature: Editing the microbiome
Thus far, researchers have mostly just monitored the body’s microbiome. Now they’re starting to modify it to treat disease.
Image credit: Science Source/Paul Gunning.
Greenhouse gas emissions in the first decades of the 21st century are likely to affect global sea level for centuries and limiting long-term sea level rise might require stringent near-term emission reductions. Image courtesy of Lucy Reading-Ikkanda.
Sea-level rise and Paris Agreement pledges
Greenhouse gas emissions in the first decades of the 21st century are likely to affect global sea level for centuries, and limiting long-term sea-level rise might require stringent near-term emission reductions.
Image courtesy of Lucy Reading-Ikkanda (artist).
A single formulation combining two gene therapies treated an array of age-related disorders in mice. Image courtesy of Pixabay /Sipa.
Multidisease gene therapy in mice
A single formulation combining two gene therapies treated an array of age-related disorders in mice.
Image courtesy of Pixabay/Sipa.

Similar Articles