Solution X-ray scattering combined with computationalmodeling reveals multiple conformations ofcovalently bound ubiquitin on PCNASusan E. Tsutakawaa, Adam W. Van Wynsbergheb, Bret D. Freudenthalc, Christopher P. Weinachtd, Lokesh Gakhare,M. Todd Washingtonc, Zhihao Zhuangd, John A. Tainera,f,g,1, and Ivaylo Ivanovh,1

aLife Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720; bDepartment of Chemistry, Hamilton College, Clinton, NY 13323;cDepartment of Biochemistry, University of Iowa College of Medicine, Iowa City, IA 52242; dDepartment of Chemistry and Biochemistry, University ofDelaware, Newark, DE 19716; eProtein Crystallography Facility, University of Iowa College of Medicine, Iowa City, IA 52242; fDepartment of MolecularBiology, The Scripps Research Institute, La Jolla, CA 92037; gThe Skaggs Institute for Chemical Biology, La Jolla, CA 92037; and hDepartment of Chemistry,Georgia State University, Atlanta, GA 30302-4098

Edited by John Kuriyan, University of California, Berkeley, CA, and approved September 9, 2011 (received for review June 28, 2011)

PCNA ubiquitination in response to DNA damage leads to therecruitment of specialized translesion polymerases to the damagelocus. This constitutes one of the initial steps in translesion synth-esis (TLS)—a critical pathway for cell survival and for maintenanceof genome stability. The recent crystal structure of ubiquitinatedPCNA (Ub–PCNA) sheds light on the mode of association betweenthe two proteins but also revealed that paradoxically, the ubiquitinsurface engaged in PCNA interactions was the same as the surfaceimplicated in translesion polymerase binding. This finding implieda degree of flexibility inherent in the Ub–PCNA complex that wouldallow it to transition into a conformation competent to bind theTLS polymerase. To address the issue of segmental flexibility, wecombined multiscale computational modeling and small angleX-ray scattering. This combined strategy revealed alternative posi-tions for ubiquitin to reside on the surface of the PCNA homotri-mer, distinct from the position identified in the crystal structure.Two mutations originally identified in genetic screens and knownto interfere with TLS are positioned directly beneath the boundubiquitin in the alternative models. These computationally derivedpositions, in an ensemble with the crystallographic and flexiblepositions, provided the best fit to the solution scattering, indicatingthat ubiquitin dynamically associated with PCNA and is capable oftransitioning between a few discrete sites on the PCNA surface.The finding of new docking sites and the positional equilibriumof PCNA–Ub occurring in solution provide unexpected insight intopreviously unexplained biological observations.

SAXS ∣ DNA replication ∣ DNA repair ∣ mutagenesis

Classical DNA polymerases—those involved in normal DNAreplication and repair—cannot accommodate many types

of DNA damage in the template strand. As a result, replicationforks stall when encountering DNA damage, and this is one of themajor sources of genome instability. One of the principal path-ways for overcoming replication blocks is translesion synthesis(TLS) by one of a variety of specialized TLS polymerases (1–4).These TLS synthesis polymerases are recruited to stalled replica-tion forks where they replace the classical polymerase and carryout DNA synthesis across from the DNA damage. DNA polymer-ase eta (pol η), for example, replicates efficiently and accuratelythrough thymine dimers and 8-oxoguanine lesions (5, 6). Signifi-cant insights into how TLS polymerases accommodate DNAdamage have come from structural and kinetic studies of theseenzymes over the last decade (2, 4).

Arguably the least understood steps of translesion synthesisare the recruitment of the TLS polymerase to stalled replicationforks and the subsequent polymerase switch between the stalledclassical polymerase and the TLS polymerase. These steps aregoverned by the monoubiquitylation of proliferating cell nuclearantigen (PCNA). PCNA is monoubiquitylated on Lys-164 by the

Rad6-Rad18 complex in response to DNA damage (7, 8). Theattachment of ubiquitin provides an additional binding surfacefor TLS polymerase, most of which possess ubiquitin-bindingmotifs (9). In addition, studies using in vitro reconstituted reac-tions have shown that the polymerase switching event requiresthe ubiquitylation of PCNA (10).

Some insights into the structural and mechanistic basis of TLSpolymerase recruitment have come from a recent X-ray crystalstructure of PCNA–Ub (11). The ubiquitin moiety interactedwith PCNA on the back face of PCNA, in a previously unde-scribed position for PCNA-binding proteins. Generally, mostPCNA-binding proteins are located on the front face, the sidewhere replication would be occurring. The attachment of ubi-quitin to the PCNA did not significantly alter the structure ofPCNA, arguing against allosteric models of TLS polymerase re-cruitment. Instead, it suggests that ubiquitylation of PCNA simplyprovides an additional binding surface for the TLS polymerases.The X-ray crystal structure, however, poses a problem regardingthe mechanism of TLS polymerase recruitment. The surface ofthe ubiquitin that interacts with ubiquitin-binding motif of TLSpolymerases and centered on Leu-8, Ile-44, and Val-70 (12) isburied at the ubiquitin–PCNA interface in the X-ray structure ofPCNA–Ub (11). Consequently the binding of the TLS polymer-ase requires that the conformation of PCNA–Ub change toexpose this surface.

To understand the dynamics of PCNA–Ub in solution, wecombined small angle X-ray scattering (SAXS) with multiscalemolecular modeling. We identified previously undescribed posi-tions for ubiquitin on the side of the PCNA ring. We show thata model of PCNA–Ub that is in dynamic equilibrium betweenside, back, and flexibly extended positions best matches theexperimental SAXS data. We argue that this dynamic range ofpositions is important for PCNA–Ub function in binding andpositioning of TLS polymerases for lesion bypass.

ResultsTo resolve questions arising from the crystal structures of yeastPCNA–Ub regarding how Pol η might access the ubiquitin inter-face, we used both SAXS and computational analyses to charac-terize the structure of PCNA–Ub in solution.

Author contributions: S.E.T., M.T.W., Z.Z., J.A.T., and I.I. designed research; S.E.T., A.W.V.W.,L.G., and I.I. performed research; B.D.F. and C.P.W. contributed new reagents/analytic tools;S.E.T. and I.I. analyzed data; and S.E.T., M.T.W., Z.Z., J.A.T., and I.I. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence may be addressed. E-mail: JATainer@lbl.gov or iivanov@gsu.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1110480108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1110480108 PNAS Early Edition ∣ 1 of 6

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

July

21,

202

1

Small Angle X-ray Scattering in Solution. To experimentally testthe architecture and flexibility of PCNA–Ub, we collected SAXSdata on yeast PCNA–Ub in solution (Fig. 1). SAXS combinedwith crystal structure restraints and conformational analyses canaccurately define flexible conformations and ensembles in solu-tion (13–17). We analyzed PCNA–Ub complexes formed byeither split-fusion (11) or by chemical cross-linking with PCNAmutant K164C (18) to ensure that the structural results are con-sistent regardless of the type of linkage between Ub and PCNA.The experimental scattering curves were practically identical forboth the split-fusion and chemically cross-linked complexes andfollowed the overall shape of a scattering curve calculated fromthe crystal structure of yeast PCNA–Ub (3L10.pdb) (11). Guinierplots were linear, indicating that the samples were not aggre-gated. The radius of gyration, Rg, is accurately defined by theSAXS experiment independent of sample concentration and con-trast and so provides an important constraint for possible solutionstructures (13). The Rg values based on the Guinier analyses were35.6 and 36.8 Å, respectively, resembling the 36.05 Å Rg calcu-lated from the crystal structure.

Ab initio shape predictions from the experimental scatteringcurves of the split-fusion and cross-linked PCNA–Ub complexesrevealed a toroidal shape readily identifiable as PCNA. However,only one protrusion corresponding to the crystallographic posi-tion for ubiquitin was determined from the ab initio shape recon-struction of the split-fusion PCNA–Ub (Fig. 1C). The shapeprediction for the cross-linked PCNA–Ub had a second protru-sion that was located in a position not corresponding to the one inthe crystal structure. The χ2 fit of the SAXS model to the PCNA–

Ub crystal structure produced relatively high values, consistentwith significant discrepancies between the experimental and com-puted scattering profiles (Fig. 1A). In fact, the fit was significantlybetter for a model of PCNA with only one or two ubiquitins pertrimer (χ2 ¼ 12.32 for split-fusion and χ2 ¼ 5.32 for cross-linked).However, a PCNA homotrimer with a single ubiquitin modifica-tion was inconsistent with the PAGE analysis showing that theUb–PCNA linkages were present in 1∶1 ratio for all three PCNAsubunits for both the split and cross-linked protein preparations.Thus we postulated that the ubiquitin must occupy positions dif-ferent from the crystallographically observed position, either at adifferent position docked against PCNA or in flexible confor-mation.

We first tested if a model that allowed flexibility in the linkerbetween PCNA and ubiquitin would better fit the split-fusion

PCNA–Ub SAXS data. To objectively examine the implicationsof the SAXS results, we used the CHARMM program implemen-ted in BILBOMD to generate 2,200 models where the threeubiquitins were allowed to move in solution (19). A minimal en-semble search (MES), a program that enables analysis of flexibleproteins, was used to identify three conformations that as anensemble best fits the scattering data. The fit to the split-fusionPCNA–Ub SAXS data was significantly better with the χ2 fitdecreasing from 23.8 to 7.53. Interestingly, some of the positionswere near the PCNA ring, suggesting that there may be alterna-tive positions for ubiquitin to bind to PCNA.

Computational Docking of Ubiquitin on PCNA. The MES analysisoutcome suggested that it may be worthwhile to examine theavailable conformational space for the Ub moiety on PCNA inmore systematic ways to identify the discrete binding positions.Fortunately, a computational modeling study of the humanPCNA–Ub complex, which was already in progress, provided al-ternative models for the PCNA–Ub complex and these computa-tional results could be directly incorporated into the experimentalSAXS investigation. The computational modeling involved acombination of tethered Brownian dynamics (TBD) (20),protein–protein docking with RosettaDock (a component of theROSETTA++ software package) (21–25), flexible loop model-ing with ModLoop (26, 27), and molecular dynamics (Fig. S1).

This multiscale approach is similar to the relaxed complexscheme (28), with the TBD simulations serving to identify pro-tein–protein conformations as likely candidates for the nativelybound complex based on electrostatic and shape complemen-tarity. At the first stage of our protocol, TBD simulations wereemployed to generate a large ensemble of 6,837 electrostaticallyand geometrically favorable configurations for one ubiquitin toone PCNA homotrimer (Fig. 2 and Fig. S1).The advantage ofTBD is that it allows very extensive sampling of the conforma-tional space for Ub–PCNA in the course of a 34-μs simulation.Subsequently we clustered the conformations using a previouslyreported clustering algorithm (29) implemented in GROMACS(g_cluster) (30) and calculated centroids for 90 clusters. Becausein TBD, PCNA and ubiquitin interact as rigid bodies, it wasnecessary to apply local docking with RosettaDock to allow forpacking of the interacting side chains of Ub and PCNA.

With Rosetta Dock, we input in the 90 cluster centroids fromTBD and developed 510 refined models of the PCNA/monoubi-quitin complex from a local Monte Carlo search using the Rosetta

Fig. 1. SAXS analysis of PCNA–Ub in solution suggests that ubiquitin is not exclusively oriented in the position determined by crystallography. (A) SAXScurves of split-fusion and cross-linked PCNA–Ub overlaid on the calculated curve for PCNA–Ub trimer from the crystal structure 3L10.pdb. (B) Guinieranalyses of SAXS data showing relative linearity of sample in Guinier region, indicating lack of aggregation in sample. (C) Ab initio shape predictions calculatedfrom experimental scattering curves of split-fusion or cross-linked PCNA–Ub suggests only one or two ubiquitin positions are extending away from thePCNA ring.

2 of 6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1110480108 Tsutakawa et al.

Dow

nloa

ded

by g

uest

on

July

21,

202

1

potential energy function. The function includes van der Waalsinteractions, ee1 implicit solvation model, hydrogen bondingand Coulombic interactions. We examined the largest 90 clusters.One of the clusters corresponded to the one observed in thePCNA–Ub crystal structure (Fig. S2). From the 90 clusters, wetook the top three Rosetta-scoring models, which were locatedalong the plane of the ring in the cavity formed between PCNAsubunits. We modeled in the linker (Ub residues 72–76) into thestructures obtained from clustering the RosettaDock output (29)using Modloop. We then refined the three models using all-atomexplicit solvent molecular dynamics with the program NAMD(31, 32) and the AMBER Parm99SB force field (33).

The three models were positioned distinctly from the ubiquitinin the crystal structure and are nestled in a groove of the PCNAring directly above the PCNA subunit interface (Fig. 3). Themodels were well-packed and exhibit a high degree of electro-static and geometric complementarity, well-known factors indetermining productive complexation. Approximately half of theinteracting residues are conserved between human and yeast,indicating that the MD-identified positions would be similar inthe yeast PCNA–Ub. The buried surface area between PCNAand ubiquitin ranged from 730 Å to 1,038 Å, an area similarin size to the 866 Å buried in the yeast PCNA–Ub structure3L10.pdb. Importantly, in the low resolution SAXS ab initioshape prediction, the ubiquitin occupying these positions wouldbe hidden along the ring.

Minimal Ensemble Search to Determine Positions of Ubiquitin on PCNAin Solution. Because the binding sites identified by computationalmethods were conserved between human and yeast, we used thepositions identified for human PCNA–Ub to reexamine the yeastPCNA–Ub solution data with theMES program. First, we made aset of 30 unique models for every permutation of yeast PCNA–

Ub with three ubiquitins per PCNA homotrimer (Fig. 4). In eachof the three positions, ubiquitin was placed in the crystallographicposition (x, 3L10.pdb) or one of the three computationally iden-tified positions ða;b;cÞ. To allow for potential flexibility in thesystem, we next generated all possible models where 1, 2, or 3ubiquitins per PCNA homotrimer were allowed to move flexiblyusing the CHARMM molecular dynamics program implementedin BILBOMD (19). We then added the models from each BIL-BOMD run that best fit the SAXS data to the 30 models wherethe ubiquitin is only in the discrete positions bound to PCNA togenerate a final set of 130 models.

The fit to the experimental data improved significantly (Fig. 4).For the split-fusion PCNA–Ub, the χ2 fit went from 23.8 to 4.35for the best ensemble of three conformations. For the cross-linkPCNA–Ub, χ2 fit went from 6.8 to 1.89. There is one deviation ofthe MES curve from the experimental curve at high angle; com-parison of the electron pair distribution or PðrÞ plots show thatthis difference is negligible for the global shape and is likely dueto more high resolution differences with the pdb models, such asordered water packing, which is difficult to model. Based on therepresentation of each conformation in solution, we calculatedthe frequency in which the ubiquitin would have populated thecrystallographic, computational, or flexible position. In both thesplit-fusion and the cross-linked PCNA–Ub SAXS studies, theubiquitin was about 25 to 30% in the crystallographic position,about 40 to 50% in the computationally determined positions,and about 25% to 30% flexible in solution. As a negative controlwe included in the MES analyses models where ubiquitin wasplaced in other positions or unmodified PCNA. These othermodels were not picked up by the MES program as significantcontributors to the SAXS data. Although we cannot exclude thepossibility that other discrete conformations are also present,SAXS does provide a picture of the range of motions possiblefor PCNA–Ub. The SAXS analysis supported a model of segmen-tal flexibility, where ubiquitin can occupy, relative to PCNA, bothdiscrete and flexible positions moving from one to the other indynamic equilibrium.

DiscussionMechanisms controlling PCNA interactions with translesionpolymerases including the role of posttranslational modificationshave general and broad-based implications for cell biology. Theattachment of ubiquitin to Lys-164 of PCNA is essential for TLS

Fig. 2. Tethered Brownian dynamics simulation shows the range of cova-lently bound Ub positions on the surface of PCNA. The positions of Ub heavyatoms (C,N,S, and O) in 6,837 frames from a 34-μs TBD simulation werebinned and displayed relative to PCNA as a 3D histogram.

Fig. 3. Three positions of ubiquitin derived frommultiscale refinement. The three computationally derived positions (blue, orange, red) are shown relative toPCNA (gray) and to the ubiquitin position determined in the crystallographic studies (black). The computationally derived positions cover up biologicallyimportant positions (G178 and E113) identified in yeast mutation studies and the J and P loops that are thought to be important structurally for ubiquitinand sumo positioning.

Tsutakawa et al. PNAS Early Edition ∣ 3 of 6

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

July

21,

202

1

(7, 8), and this posttranslational modification is believed tobe necessary for the recruitment of TLS polymerases to stalledreplication forks (9, 34). Structural characterization of the con-formation(s) of PCNA–Ub is an important step to understandingthe structural and mechanistic basis for TLS polymerase re-cruitment.

Our analysis of PCNA–Ub in solution expands the recentcrystallographic analysis of split-fusion PCNA–Ub that showeda single PCNA-binding position for ubiquitin on the back faceof the PCNA ring (11). Combining SAXS with multiscale com-putational modeling we revealed a previously undescribed modelfor PCNA–Ub, where the position of the ubiquitin moiety isdynamic and can adopt a variety of positions relative to thePCNA ring. Approximately 25 to 30% of the time the ubiquitinmoieties are positioned on the back face of the ring in the posi-tion indicated by the X-ray crystal structure. About 25 to 30% ofthe time the ubiquitin moieties are in a flexible position in whichthe ubiquitin is attached to Lys-164 of PCNA but is not otherwiseinteracting with the PCNA. About 40 to 50% of the time the ubi-quitin moieties are interacting with the side of the PCNA ring atthe subunit–subunit interface, as predicted by the Brownian dy-namics, docking and molecular dynamics simulations. This modelwas obtained using two different analogs of PCNA–Ub: a chemi-cally cross-linked one and a split-fusion one. Thus the dynamicsand the orientation of the ubiquitin are not impacted by thechemical nature of the ubiquitin–PCNA linkage. Notably, noneof these discrete positions are on the front face of the PCNA ringwhere most PCNA-binding proteins are located near the interdo-main linker (35, 36), and thus all of these positions may allow

PCNA–Ub to function as a “tool belt” (11, 37). This means thatthe TLS polymerase can be recruited to the back face or side ofthe PCNA ring without disrupting ongoing activity on the frontface of the PCNA ring.

Flexibility of the ubiquitin moiety on PCNA–Ub is supportedby a recent finding that PCNA chemically monoubiquitylatedat Lys-164, Lys-127, Lys-107, or Arg-44 were indistinguishablein recruiting TLS polymerase pol η to exchange with classicalpol δ in an in vitro polymerase exchange reaction (18). Thisobservation suggests that ubiquitin moieties attached at distinctpositions can achieve whatever conformations are necessary toform functional complexes with pol η and that conformationsare not determined by the linker or linker position.

SAXS support of a discrete position of ubiquitin on the backface of PCNA was expected given the X-ray crystal structureof PCNA–Ub. Finding a second set of discrete positions of ubi-quitin on the side of the PCNA ring, however, was surprising.The importance of the side position for TLS is supported by ex-perimental results on PCNAmutant proteins. E113G and G178S,two key amino acid substitutions in PCNA that disrupt TLS,are located at the subunit-subunit interface and would directlycontact the ubiquitin moiety in the side position (Fig. 3). TheE113G blocks the functional interactions between PCNA andTLS polymerases but does not block PCNA ubiquitination inthe presence of DNA damage (38). Similarly, the G178S substi-tution, which is directly across the subunit–subunit interface fromE113G, also blocks functional interactions with TLS polymerases(39, 40). Neither amino acid substitution interferes with normalDNA replication and repair. Moreover, the position of ubiquitin

Fig. 4. An MES ensemble of both discrete and flexible positions of ubiquitin relative to PCNA best fit the experimental SAXS data for split-fusion (green) andcross-linked (blue) PCNA–Ub. (A) Schematic showing MES methodology. One hundred thirty PDB models were generated where the three ubiquitins per PCNAhomotrimer were placed at the crystallographic (x) position, the MD-identified positions ða;b;cÞ, or the BILBOMD-generated flexible (f ). Ensembles of threemodels were then compared to the experimental SAXS data in FOXS. (B) The scattering curve of the best MES ensemble fits the experimental scattering databetter than the crystal structure 3L10.pdb. (C) PðrÞ plots showing the good fit of the MES ensemble to the experimental data. (D) The three models that as anensemble best fit the experimental scattering curve are shown in ribbon models. Relative proportion of each position in x, a∕b∕c, or f shows that ubiquitinadopts both the crystallographic and computationally determined discrete positions as well as being flexible in solution.

4 of 6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1110480108 Tsutakawa et al.

Dow

nloa

ded

by g

uest

on

July

21,

202

1

on the side of the PCNA ring is reminiscent of the position ofthe C-terminal domain (the PAD or “little finger” domain) of theEscherichia coli TLS polymerase pol IV bound to the beta slidingclamp (41). Pertinent to the role of PCNA–Ub in other pathwaysas well, the side positions are adjacent to conserved patches inPCNA, suggesting sites where interacting proteins could bindto both PCNA and ubiquitin (Fig. S3).

In general, the spatial organization of proteins on PCNA andDNA needed for efficient replication and repair is a critical as-pect of cell biology. Spatial control is essential for managing ofDNA ends while avoiding off pathway activities that would other-wise cause toxicity and mutations. In this context, these resultsprovide insights into how PCNA posttranslational modificationprovides both specificity and flexibility during translesion synth-esis in response to DNA damage. Each of the three PCNA–Ubinteraction sites identified here (flexibly, back, side) likely plays adifferent role in TLS (Fig. 5). In the flexible position, for example,the hydrophobic surface of the ubiquitin is free to bind to theTLS polymerase. As a result, this position likely plays a key rolein initially binding to TLS polymerase. Once the polymerase isbound, the ubiquitin moiety could remain in the flexible positionor move to either the back or the side position. Of course, whenthe polymerase is bound, the ubiquitin in the back or side posi-tions would likely have to reorient or rotate slightly to accommo-date the polymerase–ubiquitin interaction. Besides the canonicalhydrophobic patch of ubiquitin centered on Ile44, TLS polymer-ase such as Pol η likely binds ubiquitin at other surface patch(es)(42). This mode of interaction may allow Pol η to access the ubi-quitin moiety without fully dislodging the ubiquitin from the backor side positions on PCNA. The back position could be the posi-tion of the ubiquitin when the PCNA–Ub is functioning as a toolbelt. In this position, the TLS polymerase could be held in reserveaway from ongoing activity on the front face of the PCNA ringuntil needed. The side position could be the position of the ubi-quitin when the TLS polymerase is engaged in DNA synthesis onthe primer template in the front of the PCNA ring. Further un-derstanding of these issues awaits structures of PCNA–Ub boundto TLS polymerases.

On a more general level, this work illustrates how multidomainor covalently modified proteins can dynamically adopt multiplediscrete conformations and have segmental flexibility. Thesesystems are difficult to study structurally and often depend onthe ability to capture discrete conformations in different crystals.Here, our study that used multiscale computational analysis toidentify discrete positions and SAXS for structural analysis insolution has revealed the range of motions possible for ubiquitinlinked to PCNA and a resulting segmental flexibility for thePCNA–Ub complex suitable to regulate recruitment and coordi-nated actions of TLS polymerase.

Materials and MethodsSAXS Analysis of Split-Fusion and Cross-Linked PCNA–Ub. Split-fusion and cross-linked yeast PCNA–Ub was purified as before (11, 18). SAXS data of PCNA–Ubwere collected at the SIBYLS 12.3.1 beamline at the Advanced Light Source,LBNL (17, 43). Scattering measurements were performed on 20-μL samples at15 °C loaded into a helium-purged sample chamber, 1.5 m from the Mar165detector. Prior to data collection, PCNA–Ub were purified by gel filtration.Split-fusion PCNA–Ub was purified on a 24 mL Superdex200 column equili-brated in 20 mM Tris pH 7.5, 50 mM KGlutamate, 5 mMDTT, and 5% glycerol.Cross-linked PCNA–Ubwas purified on a 24mL Superose6 column (GE Health-care) equilibrated in 20mM Tris pH 7.5, 150 mMNaCl, 5% glycerol. Data werecollected on both the original gel filtration fractions and samples concen-trated approximately 2 × –8× from individual fractions. Fractions prior tothe void volume and concentrator eluates were used for buffer subtraction.Sequential exposures (0.5, 0.5, 5, and 0.5 s) were taken at 12 keV. Althoughscattering of the split-fusion PCNA–Ub showed no differences betweensequential exposures, the scattering of the cross-linked PCNA–Ub showeda decrease in slope, indicative that the molecules in solution were becomingsmaller. It is likely due to X-ray irradiation breaking the disulfide bond. Thefirst and second 0.5 s exposures overlaid, so the first exposure was assumedto have only minimal damage. The best data, based on signal-to-noise andGuinier, were collected on split-fusion PCNA–Ub (4.1 mg∕mL, 5 s exposure)and cross-linked PCNA–Ub (9.5 mg∕mL, 0.5 s exposure). Data was analyzedusing PRIMUS. The Porod exponent was determined from a linear regressionanalysis (I vs Q) of the top of the first peak in the Porod–Debye plot (q4 � IðqÞvs. q4) of the scattering data, implemented in OPTIMUS. The FOXS/MESserver was used for rapid comparison of experimental scattering curveswith individual or ensembles of pdb models (23). We applied BILBOMD (19)to generate models of PCNA–Ub where one, two, or three ubiquitins weremoved as rigid bodies relative to the remainder of PCNA–Ub. The other fixedubiquitins were either in the crystal structure position 1 or one of three posi-tions identified in the computational modeling. The range of Rg was set be-tween 30 and 50 Å. For each set of positions, 2,200 models were generated.

Computational Analyses. To model the human PCNA–Ub complex we em-ployed a multiscale computational strategy concisely summarized in Fig. S1.Details of the analysis are given in Supplemental Methods. Themodeling pro-tocol involved the following consecutive stages: (i) TBD (20); (ii) protein–protein docking with RosettaDock (21–25) (a component of the ROSETTA++software package); (iii) flexible loop modeling with ModLoop (26, 27); and(iv) molecular dynamics. The results of each intermediate stage were clus-tered based on pairwise distance rmsd (29, 30) with the centroids of the clus-ters subsequently used as input for the next stage. The clustering results fromTBD (29) were used as a starting point for extensive local docking searcheswith RosettaDock. The major advantage of TBD is the exhaustive sampling ofdifferent orientations of Ub on PCNA allowable by the length of the flexiblelinker connecting the two proteins (Ub residues 72–76). This advantage ispartially offset by the limitation that both Ub and PCNA are modeled as rigidbodies. The limitation is partially relaxed at the second stage of our protocol(local docking) by allowing side chain repacking and optimization. Theprotein backbone is still kept fixed in order to reduce the computationaldemands as compared to a fully flexible search. The six-residue linker be-tween Ub and PCNA was omitted at the docking stage. Therefore, in stagethree it had to be reintroduced through a loop modeling procedure (Mod-

Fig. 5. Possible biological significance of the different ubiquitin positions relative to PCNA. The observation that covalently attached ubiquitin can dyna-mically occupy different positions on PCNA allows a tool belt model for TLS polymerase binding and functions.

Tsutakawa et al. PNAS Early Edition ∣ 5 of 6

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

July

21,

202

1

loop) (25–27). Finally, in the fourth stage (all-atom explicit solvent moleculardynamics), all of the above-mentioned restrictions were removed and afully flexible conformational search is carried out to produce final refinedmodels from each cluster. These multiple trajectory MD runs (each with ap-proximately 25-ns duration) required the use of extensive supercomputingresources provided by the Teragrid and the Oak Ridge Leadership ComputingFacility (OLCF).

ACKNOWLEDGMENTS. We thank Andrew MacCammon for his advice on thetheoretical calculations. We thank Greg Hura, Michal Hammel, Robert

Rambo, and Ivan Rodic for help with SAXS analysis methods developing at12.3.1. SAXS data was collected at the SIBYLS beamline 12.3.1 (ALS, ContractDE-AC02-05CH11231). Computational resources were provided in part by aNational Science Foundation (NSF) Teragrid allocation (CHE110042) andthrough an allocation from the Innovative and Novel Computational Impacton Theory and Experiment (INCITE) program to I.I. at the Oak RidgeLeadership Computing Facility (BIP007). Work on PCNA–Ub is supportedby Georgia State University (to I.I.), a Cleon C. Arrington Research initiationgrant (to I.I.), National Cancer Institute Grants P01 CA092584 and R01CA081967 (to J.A.T), National Institute of General Medical Sciences GrantR01GM081433 (to T.W.), and NSF Grant MCB0953764 (to Z.Z.).

1. Goodman MF (2002) Error-prone repair DNA polymerases in prokaryotes and eukar-yotes. Annu Rev Biochem 71:17–50.

2. Prakash S, Johnson RE, Prakash L (2005) Eukaryotic translesion synthesis DNA poly-merases: Specificity of structure and function. Annu Rev Biochem 74:317–353.

3. Prakash S, Prakash L (2002) Translesion DNA synthesis in eukaryotes: A one- or two-polymerase affair. Genes Dev 16:1872–1883.

4. Washington MT, Carlson KD, Freudenthal BD, Pryor JM (2010) Variations on a theme:Eukaryotic Y-family DNA polymerases. Biochim Biophys Acta 1804:1113–1123.

5. Johnson RE, Prakash S, Prakash L (1999) Efficient bypass of a thymine-thymine dimerby yeast DNA polymerase, Pol eta. Science 283:1001–1004.

6. Haracska L, Yu SL, Johnson RE, Prakash L, Prakash S (2000) Efficient and accuratereplication in the presence of 7,8-dihydro-8-oxoguanine by DNA polymerase eta.Nat Genet 25:458–461.

7. Hoege C, Pfander B, Moldovan GL, Pyrowolakis G, Jentsch S (2002) RAD6-dependentDNA repair is linked to modification of PCNA by ubiquitin and SUMO. Nature419:135–141.

8. Stelter P, Ulrich HD (2003) Control of spontaneous and damage-induced mutagenesisby SUMO and ubiquitin conjugation. Nature 425:188–191.

9. Bienko M, et al. (2005) Ubiquitin-binding domains in Y-family polymerases regulatetranslesion synthesis. Science 310:1821–1824.

10. Zhuang ZH, et al. (2008) Regulation of polymerase exchange between Pol eta andPol delta by monoubiquitination of PCNA and the movement of DNA polymeraseholoenzyme. Proc Natl Acad Sci USA 105:5361–5366.

11. Freudenthal BD, Gakhar L, Ramaswamy S, Washington MT (2010) Structure of mono-ubiquitinated PCNA and implications for translesion synthesis and DNA polymeraseexchange. Nat Struct Mol Biol 17:479–U123.

12. Bomar MG, Pai MT, Tzeng SR, Li SSC, Zhou P (2007) Structure of the ubiquitin-bindingzinc finger domain of human DNA Y-polymerase eta. EMBO Rep 8:247–251.

13. Rambo RP, Tainer JA (2011) Characterizing flexible and intrinsically unstructured bio-logical macromolecules by SAS using the Porod-Debye law. Biopolymers 95:559–571.

14. Rambo RP, Tainer JA (2010) Bridging the solution divide: Comprehensive structuralanalyses of dynamic RNA, DNA, and protein assemblies by small-angle X-ray scatter-ing. Curr Opin Struct Biol 20:128–137.

15. Putnam CD, Hammel M, Hura GL, Tainer JA (2007) X-ray solution scattering (SAXS)combined with crystallography and computation: Defining accurate macromolecularstructures, conformations and assemblies in solution. Q Rev Biophys 40:191–285.

16. Tsutakawa SE, Hura GL, Frankel KA, Cooper PK, Tainer JA (2007) Structural analysis offlexible proteins in solution by small angle X-ray scattering combined with crystallo-graphy. J Struct Biol 158:214–223.

17. Hura GL, et al. (2009) Robust, high-throughput solution structural analyses by smallangle X-ray scattering (SAXS). Nat Methods 6:606–612.

18. Chen J, Ai Y, Wang J, Haracska L, Zhuang Z (2010) Chemically ubiquitylated PCNA as aprobe for eukaryotic translesion DNA synthesis. Nat Chem Biol 6:270–272.

19. Pelikan M, Hura GL, Hammel M (2009) Structure and flexibility within proteins asidentified through small angle X-ray scattering. Gen Physiol Biophys 28:174–189.

20. Gabdoulline RR, Wade RC (1998) Brownian dynamics simulation of protein-proteindiffusional encounter. Methods 14:329–341.

21. Gray JJ, et al. (2003) Protein-protein docking with simultaneous optimization ofrigid-body displacement and side-chain conformations. J Mol Biol 331:281–299.

22. Gray JJ, et al. (2003) Protein-protein docking predictions for the CAPRI experiment.Proteins 52:118–122.

23. Schneidman-Duhovny D, Hammel M, Sali A (2010) FoXS: A web server for rapidcomputation and fitting of SAXS profiles. Nucleic Acids Res 38:W540–544.

24. Schueler-Furman O, Wang C, Baker D (2005) Progress in protein-protein docking:Atomic resolution predictions in the CAPRI experiment using RosettaDock with animproved treatment of side-chain flexibility. Proteins 60:187–194.

25. Wang C, Schueler-Furman O, Baker D (2005) Improved side-chain modeling forprotein-protein docking. Protein Sci 14:1328–1339.

26. Fiser A, Do RK, Sali A (2000) Modeling of loops in protein structures. Protein Sci9:1753–1773.

27. Fiser A, Sali A (2003) ModLoop: Automated modeling of loops in protein structures.Bioinformatics 19:2500–2501.

28. Lin JH, Perryman AL, Schames JR, McCammon JA (2002) Computational drug designaccommodating receptor flexibility: The relaxed complex scheme. J Am Chem Soc124:5632–5633.

29. Daura X, Jaun B, Seebach D, van Gunsteren WF, Mark AE (1998) Reversible peptidefolding in solution by molecular dynamics simulation. J Mol Biol 280:925–932.

30. Lindahl E, Hess B, van der Spoel D (2001) GROMACS 3.0: A package for molecularsimulation and trajectory analysis. J Mol Model 7:306–317.

31. Phillips JC, et al. (2005) Scalable molecular dynamics with NAMD. J Comput Chem26:1781–1802.

32. KalÈ L, et al. (1999) NAMD2: Greater scalability for parallel molecular dynamics. J Com-put Phys 151:283–312.

33. Hornak V, et al. (2006) Comparison of multiple amber force fields and development ofimproved protein backbone parameters. Proteins 65:712–725.

34. Kannouche PL, Wing J, Lehmann AR (2004) Interaction of human DNA polymerase etawith monoubiquitinated PCNA: A possible mechanism for the polymerase switch inresponse to DNA damage. Mol Cell 14:491–500.

35. Maga G, Hubscher U (2003) Proliferating cell nuclear antigen (PCNA): a dancer withmany partners. J Cell Sci 116:3051–3060.

36. Bruning JB, Shamoo Y (2004) Structural and thermodynamic analysis of human PCNAwith peptides derived from DNA polymerase-delta p66 subunit and flap endonu-clease-1. Structure 12:2209–2219.

37. Freudenthal BD, Brogie JE, Gakhar L, Kondratick CM, Washington MT (2011) Crystalstructure of SUMO-modified proliferating cell nuclear antigen. J Mol Biol 406:9–17.

38. Northam MR, Garg P, Baitin DM, Burgers PMJ, Shcherbakova PV (2006) A novel func-tion of DNA polymerase zeta regulated by PCNA. EMBO J 25:4316–4325.

39. Zhang HS, Gibbs PEM, Lawrence CW (2006) The Saccharomyces cerevisiae rev6-1mutation, which inhibits both the lesion bypass and the recombination mode of DNAdamage tolerance, is an allele of POL30, encoding proliferating cell nuclear antigen.Genetics 173:1983–1989.

40. Freudenthal BD, Ramaswamy S, Hingorani MM, Washington MT (2008) Structure ofa mutant form of proliferating cell nuclear antigen that blocks translesion DNA synth-esis. Biochemistry 47:13354–13361.

41. Bunting KA, Roe SM, Pearl LH (2003) Structural basis for recruitment of translesionDNA polymerase Pol IV/DinB to the beta-clamp. EMBO J 22:5883–5892.

42. Ai Y, et al. (2011) A novel ubiquitin binding mode in the S. cerevisiae translesion synth-esis DNA polymerase eta. Mol Biosyst 7:1874–1882.

43. Classen S, et al. (2010) Software for the high-throughput collection of SAXS data usingan enhanced Blu-Ice/DCS control system. J Synchrotron Radiat 17:774–781.

6 of 6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1110480108 Tsutakawa et al.

Dow

nloa

ded

by g

uest

on

July

21,

202

1