Analysis of the role of PCNA-DNA contacts during clamp loading
© McNally et al; licensee BioMed Central Ltd. 2010
Received: 8 October 2009
Accepted: 30 January 2010
Published: 30 January 2010
Skip to main content
© McNally et al; licensee BioMed Central Ltd. 2010
Received: 8 October 2009
Accepted: 30 January 2010
Published: 30 January 2010
Sliding clamps, such as Proliferating Cell Nuclear Antigen (PCNA) in eukaryotes, are ring-shaped protein complexes that encircle DNA and enable highly processive DNA replication by serving as docking sites for DNA polymerases. In an ATP-dependent reaction, clamp loader complexes, such as the Replication Factor-C (RFC) complex in eukaryotes, open the clamp and load it around primer-template DNA.
We built a model of RFC bound to PCNA and DNA based on existing crystal structures of clamp loaders. This model suggests that DNA would enter the clamp at an angle during clamp loading, thereby interacting with positively charged residues in the center of PCNA. We show that simultaneous mutation of Lys 20, Lys 77, Arg 80, and Arg 149, which interact with DNA in the RFC-PCNA-DNA model, compromises the ability of yeast PCNA to stimulate the DNA-dependent ATPase activity of RFC when the DNA is long enough to extend through the clamp. Fluorescence anisotropy binding experiments show that the inability of the mutant clamp proteins to stimulate RFC ATPase activity is likely caused by reduction in the affinity of the RFC-PCNA complex for DNA. We obtained several crystal forms of yeast PCNA-DNA complexes, measuring X-ray diffraction data to 3.0 Å resolution for one such complex. The resulting electron density maps show that DNA is bound in a tilted orientation relative to PCNA, but makes different contacts than those implicated in clamp loading. Because of apparent partial disorder in the DNA, we restricted refinement of the DNA to a rigid body model. This result contrasts with previous analysis of a bacterial clamp bound to DNA, where the DNA was well resolved.
Mutational analysis of PCNA suggests that positively charged residues in the center of the clamp create a binding surface that makes contact with DNA. Disruption of this positive surface, which had not previously been implicated in clamp loading function, reduces RFC ATPase activity in the presence of DNA, most likely by reducing the affinity of RFC and PCNA for DNA. The interaction of DNA is not, however, restricted to one orientation, as indicated by analysis of the PCNA-DNA co-crystals.
The faithful and efficient copying of chromosomal DNA is performed by chromosomal replicases, which generate double-stranded DNA from primed, single-stranded DNA [1, 2]. These replicases achieve great speed and processivity by tethering to ring-shaped protein complexes called sliding clamps, which encircle DNA and have the ability to slide freely along the DNA strand [3, 4]. The DNA polymerase subunits of these replicases maintain their association with the template strand by remaining attached to the clamp, which moves along with the polymerases during DNA synthesis. The E. coli replicase (DNA polymerase III) is one of the best-studied chromosomal replicases, and it has been shown that in the absence of the sliding clamp, the replicative polymerase cannot synthesize more than a few bases without falling off. In contrast, the presence of the sliding clamp allows thousands of nucleotides to be incorporated at speeds up to 1000 base pairs per second [5, 6]. Similar conclusions have been reached for the replicase of T4 bacteriophage [7, 8].
The general architecture of sliding clamps is conserved throughout all domains of life [9–11]. In E. coli, the sliding clamp is the β subunit of the pol III holoenzyme (β clamp), and the eukaryotic sliding clamp is called Proliferating Cell Nuclear Antigen (PCNA). The sliding clamps share a domain architecture that has a six-fold pseudosymmetric rotation axis running through the center of the clamp and a central pore lined with α-helices through which DNA is threaded [10, 11]. The clamps have two unique faces, one of which (referred to here as the proximal face) is used for interaction with binding partners, such as the polymerases and clamp loader complexes [10–13]. The other face (referred to here as the distal face) has not been implicated in functional interactions.
To bind, open, and load the clamp on DNA, the clamp loader executes a cycle of ATP binding and hydrolysis. ATP binding allows for the formation of a stable complex between the clamp loader and an open clamp [18, 19]. Upon binding primer-template DNA , ATP hydrolysis is stimulated, and results in the release of the clamp on DNA (Figure 1B) [21–23].
The central hole within the sliding clamp is lined with α-helices that present a number of conserved, positively charged residues [10, 11], creating a surface that complements the negatively charged DNA backbone. Because a strong interaction between the inner surface of the sliding clamp and DNA might hinder movement of the polymerase, tight binding of DNA to the clamp in a specific orientation is not expected. Molecular dynamics simulations have predicted that while DNA has a preference to be oriented within the hole in PCNA at an angle of about 20° relative to the axis of PCNA, the DNA is also highly mobile and is not strictly restricted to any particular position .
Given the transient nature of the interaction between DNA and the clamp that is expected, the crystal structure of the E. coli β clamp bound to primer-template DNA was a surprise . This structure showed that DNA was well localized in the center of the clamp, making an angle of 22° with the axis of the clamp. The β clamp-DNA structure also revealed that the single-stranded overhang of the DNA binds to the polymerase binding site on the proximal face of the clamp, and that mutation of residues at that interface causes deficiencies in clamp loading . Thus, at least for the bacterial replicase, a specific interaction between primer-template junctions and the clamp might hold the clamp in position prior to arrival of the polymerase . Two residues in the β clamp that make contact with the double-stranded region of the DNA in the structure (Arg 24 and Gln 149) were shown to be necessary for efficient clamp loading  and have also been implicated in the ability of the clamp loader to discriminate between a 3' or 5' primer-template junction . The interaction of a sliding clamp with a tilted DNA was also observed in an electron microscopic reconstruction of DNA ligase bound to PCNA and DNA .
Several of the positively charged residues in the central helices of PCNA have been shown to be essential for the stimulation by PCNA of the activity of eukaryotic DNA polymerase δ , as well as resulting in deficient mismatch repair when mutated . In contrast, none of the residues in the center of PCNA have been implicated so far in clamp loading. The residues in the β clamp-DNA structure (Arg 24 and Gln 149) that make contact with double-stranded DNA and have been implicated in clamp loading reside in loops on the proximal face of the clamp  rather than in the central hole.
This work presents mutational analysis that shows that interactions between DNA and several positively charged residues in the center of S. cerevisiae PCNA play a role in the clamp loading process. In addition, we present a model of PCNA bound to DNA that is derived from X-ray diffraction data, which shows interactions between DNA and an additional set of positively charged residues in the center of PCNA.
To determine if PCNA-DNA interactions have relevance for clamp loading, a RFC-PCNA-DNA model was generated to visualize the path of the DNA through PCNA during the clamp loading process. In order to construct this RFC-PCNA-DNA model, we combined features of two different crystal structures, the E. coli clamp loader (γ-complex) bound to primer-template DNA (Figure 1A) , and the S. cerevisiae clamp loader (RFC) bound to closed PCNA . We aligned the RFC-PCNA structure onto the γ-complex structure using the B subunits of the two complexes. This alignment allows us to transfer bound DNA from the γ-complex structure to that of RFC, and is justified because of the similarity between the general organization of the ATPase domains in the two structures . Each ATP-coordinating subunit-subunit transition in the γ-complex-DNA structure maintains a rotation and translation that aligns along the axis of the bound DNA, and this transition is maintained roughly in the RFC-PCNA structure.
Because the PCNA in the RFC-PCNA-DNA model is closed, the step in the clamp loader cycle it represents would likely be after the clamp has been loaded but before release by the clamp loader, or perhaps a transient closed state sampled by the clamp during the process of recognizing DNA. To determine if an open PCNA during clamp loading could contact DNA in a manner analogous to that modeled for a closed clamp, a model of an open PCNA was aligned onto the PCNA in the RFC-PCNA-DNA model. The open PCNA model was generated using molecular dynamics (MD) simulations in an earlier work  and was aligned on the PCNA subunit that contacts the A subunit of RFC. This results in the open clamp following the spiral formed by the RFC subunits [32, 33]. The alignment shows that while part of the open PCNA is pulled away from the DNA in an out-of-plane manner, interactions with RFC allow the PCNA subunit that makes much of the contact (gray subunit, Figure 2C) with DNA to remain in the same position whether PCNA is open or closed.
To determine if the center of the clamp interacts with DNA during clamp loading as suggested by the RFC-PCNA-DNA model, all nine of the arginine and lysine residues that reside on the helices lining the center of S. cerevisiae PCNA (Figure 2B) were mutated individually and their effect on DNA-dependent ATPase stimulation of RFC was assayed. Because ATP hydrolysis occurs as the clamp is released onto DNA, ATPase activity serves as a useful readout of clamp loading activity.
Using the RFC-PCNA-DNA ATPase level as a benchmark, the ATPase activity of RFC with DNA and the PCNA mutants was measured (Figure 3B). The DNA construct used in the assay (called DNA-30) has a 30-base pair duplex region with a 10-base 5' overhang. Each of the PCNA mutants showed either a modest decrease in DNA-dependent ATPase stimulation of RFC or no significant effect. The largest effects were observed for PCNA R80A and PCNA R149A, but these corresponded to only a 20-25% reduction in ATPase activity compared to wild-type PCNA. Indeed, a previous study did not detect an effect on DNA-dependent RFC ATPase activity from individually mutating each of the positive residues in the center of human PCNA . The minimal effects of these mutations might reflect the fact that each point mutation removes only one-ninth of the positively charged residues in the center of PCNA.
Because the effects of single mutations were so small, it was necessary to mutate several positive residues in the center of PCNA in combination to produce a more robust effect. A number of combinations of triple mutants and quadruple mutants were studied. The triple mutant PCNA K20A/K77A/R80A was generated to test the effect of mutating residues on the α-helices of domain 1 (Figure 2B) of PCNA, while PCNA K146A/R149A/K210A tested the effects of mutating residues of domain 2. PCNA K20A/K77A/R80A/R149A combines the point mutations that individually had the four largest effects on DNA-dependent RFC ATPase activity as measured above.
The results (Figure 3C) indicate that the domain 2 triple mutant, PCNA K146A/R149A/K210A, shows no decrease in RFC ATPase activity in the presence of DNA. In contrast, a significant decrease in ATPase activity was seen for the domain 1 triple mutant PCNA K20A/K77A/R80A and the quadruple mutant PCNA K20A/K77A/R80A/R149A (which are the same except for an additional mutation in the latter). Strikingly, the ATPase activity with PCNA K20A/K77A/R80A/R149A present was the same as that observed in the absence of PCNA.
The drastic effects on PCNA-stimulated and DNA-dependent RFC ATPase activity for the PCNA K20A/K77A/R80A and PCNA K20A/K77A/R80A/R149A mutants suggest that interaction of the mutated residues with DNA plays an important role during the loading process. To verify whether contact with DNA by these residues in the center of PCNA is a necessary step for the stimulation of RFC ATPase activity, ATPase assays were performed with PCNA K20A/K77A/R80A/R149A and DNA constructs of various lengths, the shortest of which are too short to extend beyond the RFC complex and into the clamp. The RFC-PCNA-DNA model predicts that 25 base pairs in the duplex region of DNA would extend completely through the clamp, that 20 base pairs would extend partially into the clamp, and that 15 base pairs would terminate near the opening of the clamp (Figure 2A). DNA constructs shorter than 15 base pairs are predicted to be too short to contact residues in the center of PCNA when bound to RFC. Each DNA construct used contained 10-base 5' overhangs and was named according to the length of its duplex region (DNA-25 has 25 base pairs in the duplex region, etc.).
The results show that the length of the DNA constructs did not affect their ability to stimulate RFC ATPase activity when no PCNA is present (Figure 3D, orange columns 1-5). Also, for each of the DNA constructs, wild-type PCNA stimulated RFC ATPase activity to approximately the same level (Figure 3D, red columns 1-5). In the presence of DNA constructs that were predicted to enter or pass through the clamp during loading (DNA-30, DNA-25, and DNA-20), PCNA K20A/K77A/R80A/R149A failed to stimulate RFC ATPase activity (Figure 3D, blue columns 1-3). In the presence of the DNA construct predicted to extend near the opening of the clamp, DNA-15, the mutant PCNA stimulated ATPase activity to an intermediate level (Figure 3D, blue column 4). Finally, for the shortest DNA construct DNA-13, the mutant PCNA approaches wild-type PCNA in ability to stimulate RFC ATPase activity (Figure 3D, blue column 5). The DNA length dependence of these results suggests that the lack of contact between certain positively charged residues in the center of PCNA and DNA passing through the clamp negatively affects the clamp loading process. The ability of the mutant PCNA to stimulate RFC in the presence of DNA-13 indicates that the mutations did not impair the stability of the clamp.
To examine whether single-stranded DNA passing through the clamp might also affect clamp loading, a DNA construct that had a 13-base duplex region with an overhanging 17-base 3' strand (DNA-13/overhang) was tested. PCNA K20A/K77A/R80A/R149A showed a large deficiency in stimulating RFC ATPase activity in the presence of this DNA construct, in marked contrast to the full activation seen in the presence of DNA-13, which is a similar DNA construct but lacks the 3' overhang (Figure 3D, blue columns 5 & 6). This suggests that interaction between single-stranded DNA and the clamp could take place during clamp loading.
PCNA K20A/K77A/R80A/R149A is not deficient in stimulating RFC ATPase activity in the presence of a short DNA construct that does not extend through the center of PCNA (Figure 3D). This suggests that while residues at the center of PCNA are not necessary to stabilize a conformation of RFC capable of ATPase activation and DNA binding when the DNA is too short to contact them, they are necessary when the DNA is long enough to reach them. The eukaryotic primase, pol α, generates a RNA/DNA primer strand longer than 30 bases, which is easily long enough to thread through the clamp during loading, emphasizing the importance of the interactions with PCNA [1, 35].
There are two possible reasons that the loss of PCNA residues Lys 20, Lys 77, Arg 80, and Arg 149 interrupts the ATPase activity of RFC in the presence of DNA. First, the PCNA-DNA interaction may become unfavorable, which would destabilize the position of the DNA within the RFC-PCNA-DNA complex or even disrupt formation of the ternary complex itself. Second, following formation of the RFC-PCNA-DNA complex, an unfavorable interaction between PCNA and DNA may hinder clamp closing and release of the clamp by RFC, interfering with the typical RFC ATPase cycle. Fluorescence anisotropy binding experiments were performed to measure the affinity of RFC for TAMRA-labeled DNA in the presence of wild-type PCNA or PCNA K20A/K77A/R80A/R149A. RFC was titrated into a mixture of primer-template DNA labeled with a TAMRA fluorophore on the 5' end of the template strand in the presence of PCNA (or mutant) and the non-hydrolyzable ATP analog, ATP-γ-S.
The anisotropy binding data allows us to determine which step of the three-part clamp loader cycle is affected by the mutant clamp in the presence of a long DNA construct. In step 1 of the cycle, ATP-loaded RFC binds and opens PCNA; in step 2, RFC and PCNA bind to DNA; and in step 3, ATP is hydrolyzed and the clamp is released around the DNA (Figure 4C). The mutations do not affect step 1; because the mutant PCNA performed at wild-type level with the shortest DNA in both ATPase stimulation and DNA binding affinity, it is clear that the mutant clamp is capable of binding RFC. However, the anisotropy binding data using the long DNA suggests that the formation of the RFC-PCNA-DNA ternary complex in step 2 is hindered when contact between DNA and the surface of the PCNA hole is rendered incompatible by mutations. This effect may account for much of the deficiency in ATPase activity when the mutant PCNA is used for longer DNA constructs. Finally, it is possible, perhaps likely, that step 3 is also affected by the mutant clamp in the presence of long DNA. It is reasonable to expect that loss of DNA-contacting residues would hinder closing of the clamp and its release around DNA.
The ATPase and anisotropy binding experiments presented above show that DNA interacts with the center of PCNA during clamp loading. To structurally characterize interactions between the center of PCNA and DNA, we set out to solve the X-ray crystal structure of a PCNA-DNA complex. A key to the successful determination of the structure of the β clamp-DNA complex  was the use of DNA labeled with a Cy5 chromophore that colored the DNA-containing crystals blue. Our crystallization of S. cerevisiae PCNA in complex with DNA used a similar approach. Blue crystals of wild-type PCNA were obtained readily, in the presence of primer-template DNA of varying lengths labeled with a Cy5 molecule on the 5' end of the primer strand. Electron density maps calculated after placing PCNA in the unit cell by molecular replacement showed no features corresponding to DNA. As for crystals of S. cerevisiae PCNA alone , these newly obtained blue crystals belong to a cubic space group that features a crystallographic three-fold rotation axis running through the center of PCNA. This results in three-fold rotational averaging of electron density in the center of the PCNA ring, thereby obscuring features arising due to DNA.
Crystallographic Data and Refinement Statistics
83.9 93.7 136.9
Number of atoms
r.m.s. deviation, bond lengths (Å)
r.m.s. deviation, bond angles (°)
In an attempt to crystallographically trap an open clamp, the single-chain PCNA contained mutations R110S and Y114S in the first subunit [36, 37]. The PCNA in the model is very well resolved and, despite these mutations, is closed and does not significantly deviate from wild-type PCNA. The inter-subunit linkers engineered into the single-chain PCNA were not resolved in the electron density, making it difficult to determine where the DNA lies in relation to the one unconstrained PCNA interface. The four-base DNA overhang was also not visible in the experimental maps. As a result, it is unknown which end of the DNA in the model is blunt and which is the primer-template junction. Because the strongest DNA density lies towards the distal face of PCNA (the face that does not bind polymerases) and density corresponding to the overhang is not visible there, the identities of the DNA bases are assigned as if this is the blunt end. The DNA strand labeled as the primer strand in the model will be referred to as DNA strand 1 and the template strand will be referred to as DNA strand 2 (Figures 5B &5C).
To examine whether the PCNA-DNA interactions in the X-ray model participate in clamp loading, residues from the PCNA-DNA model that contact DNA strand 1 were simultaneously mutated (PCNA K13A/R14A/K146A/R149A) and assayed for ability to stimulate RFC ATPase activity in the presence of DNA. This mutant had no effect on DNA-dependent RFC ATPase activity (Figure 3C, blue column 4). The effects of mutations in the two proposed DNA-contacting loops of PCNA, the His 190 loop and the Asn 84 loop, were also tested (Figure 3C, blue columns 5 & 6). Mutation of the two polar residues together in the His 190 loop, His 190 and Thr 193, to alanine had no significant effect on DNA-dependent ATPase activity of RFC. For the Asn 84 loop, a triple mutant was produced in which the three polar residues Asn 83, Asn 84, and Thr 85 were mutated to alanine simultaneously; this mutant also displayed only a small decrease in the ability to stimulate DNA-dependent RFC ATPase activity.
The positively charged residues in the α-helices lining the hole of PCNA that interact with DNA in our PCNA-DNA model (Lys 13, Arg 14, Lys 146, and Arg 149) do not affect RFC ATPase activity when mutated, and largely do not overlap with residues that were found to have the greatest effect on ATPase activity during loading (Lys 20, Lys 77, Arg 80, and Arg 149). In addition, the DNA angle relative to PCNA in the PCNA-DNA model is ~26° greater than that in the RFC-PCNA-DNA model (Figure 6D), and the DNA in the PCNA-DNA model would collide with PCNA if it was lengthened towards the proximal face. Thus, the DNA angle seen in the PCNA-DNA model likely does not occur during clamp loading. However, it may be used for functions that are yet to be identified, and raises the possibility that PCNA presents different sets of positively charged residues to produce tilted PCNA-DNA interactions of various angles for different functions.
Our PCNA-DNA model adds to a growing body of evidence that DNA has a propensity to lie in the center of the clamp at an angle [24, 25, 27]. To compare the orientations of DNA bound to the β clamp and to PCNA, the β clamp-DNA structure and the PCNA-DNA model were aligned on the clamps. The overlay demonstrates the sharper ~40° tilt angle of the DNA in the PCNA-DNA model compared to the 22° tilt of the DNA bound to the β clamp (Figure 6E). In both structures, the tilt of the DNA results in interactions between DNA and the α-helices that line the central holes in the clamps . Thus, the observed tilts of the DNA bound to both the β clamp and to PCNA appear to be partly a consequence of the optimization of contact between the negatively charged DNA backbone and the positively charged center of the clamp.
Another PCNA-DNA interaction has been reported in which the DNA adopts a tilt inside the ring; a single-particle EM reconstruction of P. furiosis DNA ligase in complex with PCNA and DNA displays a DNA molecule threaded through the clamp at an angle . Docking of our PCNA-DNA model into the ligase-PCNA-DNA volume shows that the angle formed by the DNA with the PCNA axis in the X-ray derived model is greater than that seen in the EM volume.
X-ray crystallography has provided snapshots of several steps of the clamp loader cycle, including the apo-γ-complex , nucleotide-bound γ-complex , RFC bound to closed PCNA , and DNA-bound γ-complex . These structures, along with other biochemical studies, have established that loops facing the interior of the clamp loader recognize DNA by presenting conserved positively charged and polar residues that are positioned to track the negatively charged phosphate backbone of a bound DNA molecule [31, 40, 41], and that DNA binding helps create a geometry of clamp loader subunits that is activated for ATPase catalysis . There is presently no high-resolution structure depicting a clamp loader-clamp-DNA ternary complex, and so less is understood about the role of DNA binding to the clamp during loading. A DNA-binding role during loading for residues located near the proximal face of the β clamp has been established ; in this work, however, an additional set of residues is proposed to function during clamp loading. Mutational analysis suggests that positively charged residues in the center of PCNA present a binding surface to DNA that is used during clamp loading, expanding the role of the clamp during loading beyond what is currently understood.
Mutation of Lys 20, Lys 77, Arg 80, and Arg 149 in combination had a large effect on RFC ATPase activity while triple and quadruple mutation of other positively charged residues in the center of PCNA did not. Interestingly, the contacts observed in the PCNA-DNA model derived from crystallographic data do not appear to play a role in the clamp loading process. Thus, as predicted previously by molecular dynamics simulations , the interaction of DNA with PCNA is relatively unrestricted, with a tilted orientation of the DNA being preferred because it maximizes electrostatic complementarity with the clamp. That one combination of mutations showed effects while others did not suggests that the placement of the mutated residues relative to the DNA passing through the clamp, and not simply their charge, is a significant factor in affecting clamp loading. Further structural studies are necessary to elucidate the exact path of the DNA through the clamp during loading and which residues of the clamp it interacts with most closely.
S. cerevisiae PCNA was cloned into a modified pET-28 vector (Novagen) so that it was preceded by an N-terminal 6-His tag cleavable by PreScission™ protease. E. coli BL21(DE3) cells were transformed with the vector and grown at 37°C to an optical density of about 1.00, and the protein was expressed at 18°C for 12-16 hours after induction with the addition of 1 mM IPTG. The harvested cells were resuspended in Buffer A (30 mM Hepes pH 7.6, 20 mM imidazole, 500 mM NaCl, 10% glycerol, and 5 mM β-mercaptoethanol). The cells were lysed by several passes through a French pressure cell (EmulsiFlex-C5; Avestin) and then centrifuged at 40,000 rpm for 40 minutes to remove cellular debris. After the lysate was filtered through a 0.45 μm filter disk (Nalgene), it was applied to a 5 mL HisTrap FF column (GE Healthcare) and eluted with a step of 50% Buffer B (same as Buffer A but with 500 mM imidazole). The selected fractions were dialyzed against Buffer A while incubating with PreScission™ protease, and again applied to a 5 mL HisTrap FF column. The selected fractions were dialyzed against Buffer C (30 mM Tris pH 7.6, 50 mM NaCl, 10% glycerol, 2 mM DTT) and applied to a 5 mL HiTrap Q HP column (GE Healthcare) and eluted with a gradient of 0 to 50% Buffer D (same as Buffer C but with 2 M NaCl) over 20 column volumes. The selected fractions were dialyzed against Buffer E (30 mM Tris pH 7.5, 100 mM NaCl, 2 mM DTT), concentrated, and stored at -80°C.
The single-chain PCNA was cloned into the same modified pET-28 vector and was preceded by an N-terminal 6-His tag cleavable by PreScission™ protease. The C-terminal glutamate residues of the first and second PCNA subunits in the chain were eliminated, as were the N-terminal methionine residues of the second and third subunits. The linker sequence between the first and second subunits was GSNSQSNGSGA; the linker sequence between the second and third subunits was GSNSQASNSGA. The single-chain PCNA contained mutations R110S and Y114S in the first subunit [36, 37]. Single-chain PCNA was expressed and purified in an identical manner to that of wild-type PCNA. PCNA mutants were expressed and purified as wild-type. S. cerevisiae RFC was expressed and purified as described elsewhere  except that after the last purification step, the sample was buffer exchanged into 30 mM Hepes pH 7.6, 200 mM NaCl, 2 mM DTT, 1 mM ATP, and 10 mM MgCl2, then concentrated and stored at -80°C.
For crystallization of PCNA-DNA, single-chain PCNA was mixed with DNA at a 1.5 molar excess of DNA, producing a final PCNA concentration of 15 mg/mL. Crystals were grown using the hanging drop vapor diffusion method; an equal volume of well solution (100 mM Na acetate pH 4.6, 100 mM NaCl, 14% PEG 4K) was added to the PCNA-DNA mixture and equilibrated with well solution at 20°C. The oligonucleotide used in crystallization was purchased from Integrated DNA Technologies; the primer strand sequence was 5'-Cy5-CCCATCGTAT-3'. The template strand sequence was 5'-TTTTATACGATGGG-3'. The DNA was annealed by heating to >95°C and slowly cooling to room temperature. For cryoprotection, crystals were equilibrated in 100 mM Na acetate, 20% PEG 4K, 100 mM NaCl, 20% glycerol, and 200 mM DNA before looping and flash-freezing in liquid nitrogen.
X-ray data were collected on beamlines 8.2.2 and 8.2.1 (Advanced Light Source, Berkeley, CA). Reflections were processed and scaled using HKL2000 . Molecular replacement solutions using S. cerevisiae PCNA as a search model (RCSB Protein Data Bank code 1PLQ ) were obtained using Phaser . Structure refinement was performed using PHENIX  and model building was carried out using Coot . The DNA was refined as a rigid body, and the positions of the individual atoms within the DNA were not refined. The atomic displacement parameters of the DNA were refined with the DNA defined as one rigid (TLS) group. The inter-subunit linkers engineered into the single-chain PCNA were disordered, and not included in the final model. It was not clear from the experimental maps which PCNA subunit contained the mutated residues R110A and Y114A, and thus these mutations were also not included in the model. The atomic coordinates and structure factors have been deposited in the Protein Data Bank (3K4X).
ATPase activity was measured by continuously monitoring the pyruvate kinase and lactate dehydrogenase-coupled oxidation of NADH at a wavelength of 340 nm . With a reaction volume of 150 μL, 0.05 μM RFC, 0.10 μM PCNA or PCNA mutant, and 0.10 μM DNA were mixed with 30 mM Tris pH 7.5, 100 mM NaCl, 0.5 mM ATP, 1 mM phosphoenolpyruvate, 0.2 mg/mL NADH, 74 units/mL pyruvate kinase, 104 units/mL lactate dehydrogenase, and 10 mM MgCl2. The reaction was measured at 30°C, and reaction velocity was linearly fit to the slope of data recorded over a time period of 100-500 seconds. ATPase data was collected using a Molecular Devices SpectraMax Plus microtiter plate reader. The DNA-30 oligonucleotide had a double-stranded region 30 base pairs long and a 10-base pair overhang with a 3' recessed end. The template strand sequence was 5'-GAGCATTCAAGGACTTAGGTCTGTATTTATCTACCCACAA-3'; the primer strand sequence was 5'-TTGTGGGTAGATAAATACAGACCTAAGTCC-3'. The shorter DNA constructs had an identical sequence to DNA-30 but were truncated from the blunt end. The DNA was annealed by heating to >95°C and slowly cooling to room temperature. DNA constructs were obtained from Integrated DNA Technologies.
RFC was titrated into a reaction mixture of 30 mM Tris pH 7.5, 200 mM NaCl, 2 mM DTT, 1 mM ATP-γ-S, 10 mM MgCl2, and 100 nM DNA-30 or DNA-13 labeled with a TAMRA fluorophore on the 5' overhanging end (obtained from Integrated DNA Technologies). The reactions contained 1 μM PCNA or PCNA K20A/K77A/R80A/R149A. Anisotropy change upon addition of RFC was noted with an excitation wavelength of 550 nm and an emission wavelength of 580 nm using a Spex Jobin Yvon FluoroMax-3 fluorimeter with a Peltier-temperature controlled sample chamber at 25°C and Glan-Thompson calcite-prism polarizers. Data was fit using Prizm 5.0 (GraphPad Software) to the one-site binding equation Y = Bmax X/(Kd + X), where X is the RFC concentration, Y is the change in anisotropy, and Bmax is the maximum change in anisotropy.
We thank Xiaoxian Cao, Meindert Lamers, Kyle Simonetta, Brian Kelch, and Jodi Gureasko for valuable discussions and technical assistance, Steven Kazmirski for providing the model of the open clamp, and Kouta Mayanagi and Kosuke Morikawa for sharing the ligase-PCNA-DNA EM reconstruction. This work was supported in part by grants to J.K. from the NIH (GM045547) and NCI (CA092584). The Advanced Light Source is supported by the U.S. Department of Energy under contract DE-AC03-76SF00098 at the Lawrence Berkeley National Laboratory.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.