Article

Structurally Distinct Ubiquitin- and Sumo-Modified

PCNA: Implications for Their Distinct Roles in the
DNA Damage Response
Graphical Abstract

Authors
Susan E. Tsutakawa, Chunli Yan, ...,
John A. Tainer, Ivaylo Ivanov

Correspondence
iivanov@gsu.edu (I.I.),
jatainer@lbl.gov (J.A.T.)

In Brief
Tsutakawa et al. combine computational
modeling and small-angle X-ray
scattering to show that SUMOylated and
ubiquitinated PCNA have strikingly
different conformations in solution due to
opposite electrostatics of the modifiers.
The distinct conformations underlie
distinct roles for PCNA-Ub and PCNASUMO in DNA damage responses.

Highlights
d

Ubiquitinated and SUMOylated PCNA complexes are

structurally distinct in solution

Ubiquitin has segmental flexibility and occupies discrete

positions on PCNA

SUMO associates by simple tethering and adopts extended

flexible conformations

Distinct PTM-PCNA conformations underlie distinct roles in

DNA damage response

Tsutakawa et al., 2015, Structure 23, 110

April 7, 2015 2015 Elsevier Ltd All rights reserved
http://dx.doi.org/10.1016/j.str.2015.02.008

Please cite this article in press as: Tsutakawa et al., Structurally Distinct Ubiquitin- and Sumo-Modified PCNA: Implications for Their Distinct Roles in
the DNA Damage Response, Structure (2015), http://dx.doi.org/10.1016/j.str.2015.02.008

Structure

Article
Structurally Distinct Ubiquitin- and Sumo-Modified
PCNA: Implications for Their Distinct
Roles in the DNA Damage Response
Susan E. Tsutakawa,1,7 Chunli Yan,2,7 Xiaojun Xu,2 Christopher P. Weinacht,3 Bret D. Freudenthal,4,8 Kun Yang,3
Zhihao Zhuang,3 M. Todd Washington,4 John A. Tainer,1,5,6,* and Ivaylo Ivanov2,*
1Life

Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA

of Chemistry, Center for Diagnostics and Therapeutics, Georgia State University, PO Box 3965, Atlanta, GA 30302, USA
3Department of Chemistry and Biochemistry, University of Delaware, Newark, DE 19716, USA
4Department of Biochemistry, University of Iowa College of Medicine, Iowa City, IA 52242, USA
5Department of Molecular Biology, The Scripps Research Institute, La Jolla, CA 92037, USA
6Skaggs Institute for Chemical Biology, La Jolla, CA 92037, USA
7Co-first author
8Present address: National Institute of Environmental Health Sciences, 111 T.W. Alexander Drive, Research Triangle Park, NC 27709, USA
*Correspondence: iivanov@gsu.edu (I.I.), jatainer@lbl.gov (J.A.T.)
http://dx.doi.org/10.1016/j.str.2015.02.008
2Department

SUMMARY

Proliferating cell nuclear antigen (PCNA) is a pivotal

replication protein, which also controls cellular responses to DNA damage. Posttranslational modification of PCNA by SUMO and ubiquitin modulate
these responses. How the modifiers alter PCNAdependent DNA repair and damage tolerance pathways is largely unknown. We used hybrid methods
to identify atomic models of PCNAK107-Ub and
PCNAK164-SUMO consistent with small-angle X-ray
scattering data of these complexes in solution. We
show that SUMO and ubiquitin have distinct modes
of association to PCNA. Ubiquitin adopts discrete
docked binding positions. By contrast, SUMO associates by simple tethering and adopts extended flexible conformations. These structural differences are
the result of the opposite electrostatic potentials of
SUMO and Ub. The unexpected contrast in conformational behavior of Ub-PCNA and SUMO-PCNA
has implications for interactions with partner proteins, interacting surfaces accessibility, and access
points for pathway regulation.

INTRODUCTION
Dynamic assembly, coordinated access to DNA, and conformational switching are critical aspects of DNA replication and
repair, essential processes upon which life depends. In both processes, the sliding clamp proliferating cell nuclear antigen
(PCNA) (Ivanov et al., 2006; Kelman, 1997; Krishna et al., 1994;
Moldovan et al., 2007) acts as a master coordinator of multiple
pathways controlling replication and DNA damage responses
(DDR). The toroidal shape of PCNA allows it to topologically
encircle DNA while also binding to core replisomal constituents
and numerous repair and cell-cycle control proteins (Jonsson

and Hubscher, 1997; Maga and Hubscher, 2003; Moldovan

et al., 2007; Stolimenov and Helleday, 2009). In this capacity,
the sliding clamp acts as a platform for the assembly of the replication machinery on DNA. Emerging evidence suggests that
posttranslational modifications (PTMs) of PCNA play a critical
role in coordinating DDR to suppress genome instability. This
added level of regulation is realized through the interplay of
PTMs and partner protein recruitment to PCNA. Specifically,
reversible covalent attachment of ubiquitin or ubiquitin-like
(UBL) proteins (Bienko et al., 2005; Hicke, 2001; Hoege et al.,
2002; Perry et al., 2008; Ulrich, 2005, 2012; Ulrich and Walden,
2010) (e.g. SUMO) selects among alternative damage response
pathways such as homologous recombination or translesion
synthesis (TLS). The structural basis for such selection has so
far remained elusive. Ubiquitin and SUMO both attach to their
PCNA target through a glycine residue from the end of their flexible carboxy termini. This glycine forms an isopeptide linkage
with a specific lysine side chain on the target. Despite sharing
a common b-GRASP fold, SUMO and ubiquitin elicit distinct
functional outcomes, and characteristically have been portrayed
as having antagonistic roles (Perry et al., 2008). However, recent
studies (Stelter and Ulrich, 2003) have significantly attenuated
this picture by showing that the two modifiers could act in concert to affect cellular fates.
UV irradiation or exposure to DNA-damaging agents
commonly results in lesions that cause replication fork stalling.
The cellular response is monoubiquitination of PCNA at a
conserved lysine K164, which in turn triggers TLS, an essential
damage tolerance pathway (Hoege et al., 2002; Stelter and Ulrich, 2003). In TLS the replicative polymerase is transiently
exchanged with a specialized TLS polymerase capable of progressing past lesions in the template strand (Yang et al., 2013).
Ubiquitin attachment provides an additional binding surface for
TLS polymerases with a ubiquitin-binding motif (Bienko et al.,
2005). By contrast, PCNA SUMOylation at positions K164 or
K127 leads to suppression of homologous recombination
through the recruitment of the antirecombinogenic helicase
Srs2 (Pfander et al., 2005). Srs2 interacts with SUMO-PCNA
through its carboxy-terminal domain, which harbors tandem

Structure 23, 110, April 7, 2015 2015 Elsevier Ltd All rights reserved 1

Please cite this article in press as: Tsutakawa et al., Structurally Distinct Ubiquitin- and Sumo-Modified PCNA: Implications for Their Distinct Roles in
the DNA Damage Response, Structure (2015), http://dx.doi.org/10.1016/j.str.2015.02.008

receptor motifs: a SUMO interaction motif (SIM) and a noncanonical PIP box (Armstrong et al., 2012; Freudenthal et al.,
2011). Both motifs are required for specific recognition of
SUMO-PCNA. One possible mechanism for Srs2 to suppress
homologous recombination is through disruption of Rad51 filaments (Krejci et al., 2003; Veaute et al., 2003), thereby assisting
the ubiquitin-dependent TLS pathway. A novel alternative mechanism has been proposed wherein Srs2 binding to SUMO-PCNA
dissociates the replicative and TLS polymerases from the repair
synthesis machinery and, thus, prevents the synthesis dependent extension of recombination intermediates (Burkovics
et al., 2013). This latter mechanism requires only Srs2 recruitment through the SIM motif but neither its translocase
activity nor interaction with Rad51. PCNA ubiquitination and
SUMOylation offer striking examples of crosstalk between pathways controlled by distinct PTMs (Stelter and Ulrich, 2003; Ulrich, 2005). While attachment of a single ubiquitin is the dominant
DDR response in mammalian cells, yeast PCNA can also undergo polyubiquitination through the non-proteasomal K63 linkage (Hoege et al., 2002). Polyubiquitination of PCNA at K164
channels DDR to yet another pathway, error-free damage
bypass by template switching. In addition, in yeast PCNA monoubiquitination at position K107 helps cells to overcome defects in
DNA ligation through an S-phase checkpoint activation (DasBradoo et al., 2010; Nguyen et al., 2013). Damage-related ubiquitination at other PCNA positions has been recently identified:
PCNA was ubiquitinated at K248 after UV irradiation (Povlsen
et al., 2012) and ubiquitinated at K168 after colchicine treatment
(Xu et al., 2010). How could PTMs introduced at different positions on PCNA result in such vastly divergent functional outcomes? Furthermore, how do posttranslational modifications
of PCNA facilitate recruitment of subsequent effector proteins
in these pathways?
To answer these questions, we modeled PCNA covalently
modified by ubiquitin at residue K107 and SUMO at K164
using a multiscale computational protocol. Models consistent
with solution X-ray scattering data were identified (Hura et al.,
2009; Putnam et al., 2007; Rambo and Tainer, 2010, 2011),
which allowed us to assess the structural differences of PTMPCNA complexes (PCNAK107-Ub and PCNAK164-SUMO) and
compare these with PCNAK164-Ub from a previous study. Here
we show that SUMO and Ub have distinct modes of interaction
with PCNA and that the position of ubiquitin attachment, 107
versus 164, alters conformation. Ubiquitin bound to PCNA can
dynamically adopt multiple discrete docked conformations. By
contrast, SUMO is flexibly tethered, with no substantial docking
interactions with PCNA. Our hybrid structural analysis reveals
the biologically relevant conformations of modified PCNA in
solution that effector proteins would first encounter and interact
with.
RESULTS
Distinct Architectures of Modified PCNA Complexes
Uncovered by SAXS Data
We conducted small-angle X-ray scattering (SAXS) experiments
to probe the overall architecture and flexibility of Saccharomyces cerevisiae PCNAK107-Ub and PCNAK164-SUMO complexes in solution. Covalent attachment in the PCNA-Ub com-

plex was achieved by chemical crosslinking with a K107C

PCNA mutant (Chen et al., 2010). The PCNA-SUMO complex
was formed by split-fusion (Freudenthal et al., 2010, 2011).
The split-fusion and chemical ligation methods yield linkages
close to the enzymatically produced isopeptide bond (Chen
et al., 2010; Tsutakawa et al., 2011). SDS-PAGE analysis
showed that there were three modifications per PCNA
homotrimer.
SAXS data were collected for PCNAK107-Ub and PCNAK164SUMO complexes, and compared with the previously obtained
scattering profile of PCNAK164-Ub and simulated profiles from
available crystal structures (split-fusion PCNAK164-Ub [Freudenthal et al., 2010] and PCNAK164-SUMO [Armstrong et al., 2012];
PDB ID 3L10 and 3V60, respectively). All three experimental
scattering profiles (Figure 1) are distinct, revealing that the three
PTM complexes adopt conformations with different levels of
compactness in solution. PCNAK164-Ub is slightly less compact
than PCNAK107-Ub with the corresponding P(r) distribution
tailing to higher values of r (Figure 1B). SUMOylated PCNA presents an I(q) profile (Figure 1A) with overall shape similar to the
SAXS profile of the 3V60 crystal structure, but considerably
more extended (Figure 1B). Fitting to the SAXS data, the crystal
structures 3L10 and 3V60 produced high cfree values consistent
with significant discrepancies between the conformations adopted in the crystal structures and the solution phase ensembles.
Guinier plots (Figure 1C) for PCNAK107-Ub and PCNAK164SUMO were linear, indicating that the samples were not aggregated. The radius of gyration, Rg, is accurately defined by
SAXS regardless of sample concentration and, therefore, provides an important constraint for possible solution structures
(Rambo and Tainer, 2011). The Rg values (Table S1) based on
the Guinier analyses were 35.5 and 35.6 A for PCNAK107-Ub
and PCNAK164-Ub, respectively. These values are comparable
with the Rg of 36.05 A calculated from the 3L10 crystal structure.
The PCNAK164-SUMO complex features a significantly larger Rg
of 47.1 A, suggesting that SUMO may occupy more extended
positions on PCNA compared with ubiquitin. To determine the
compactness of the solution conformation, we compared the
Porod-Debye constant, as determined from a linear regression
analysis (I versus q) of the top of the first peak in the Porod-Debye plot (q4 3 I(q) versus q4). The Porod-Debye constant reflects
the protein density and volume compared with the Rg (Rambo
and Tainer, 2011). Well-folded proteins have a constant
of 4, while globular domains connected by linkers, such as
Mre11-Rad50 (Deshpande et al., 2014), can be as low
as 3. PCNAK164-SUMO, PCNAK107-Ub, and PCNAK164-Ub had
Porod-Debye constants of 3.5, 4, and 3.7, respectively. We
previously showed that PCNAK164-Ub adopted both docked
and flexible conformations, in agreement with the Porod-Debye
constant of 3.7. The Porod-Debye constant for PCNAK164SUMO suggests a more flexible structure, while that for
PCNAK107-Ub is consistent with a compact structure. To further
examine the compactness of the different modified PCNA complexes, we used a volume of correlation (Vc)-based dimensionless Kratky plot (Figure 1D). (Durand et al., 2010; Reyes et al.,
2014) The three modified PCNAs show a ranking in order of
compactness of PCNAK107-Ub, PCNAK164-Ub, and PCNAK164SUMO (at 1.7 on the abscissa), consistent with the Porod-Debye
constants.

2 Structure 23, 110, April 7, 2015 2015 Elsevier Ltd All rights reserved

Please cite this article in press as: Tsutakawa et al., Structurally Distinct Ubiquitin- and Sumo-Modified PCNA: Implications for Their Distinct Roles in
the DNA Damage Response, Structure (2015), http://dx.doi.org/10.1016/j.str.2015.02.008

Figure 1. Distinct Architectures of PCNAUb and PCNA-SUMO Complexes from

SAXS Analysis
(A) SAXS profiles of PCNAK164-SUMO (blue),
PCNAK107-Ub (green), and PCNAK164-Ub (red).
(B) P(r) functions for PCNAK164-SUMO (blue),
PCNAK107-Ub (green), and PCNAK164-Ub (red).
(C) Guinier analyses of SAXS data for PCNAK107Ub (green) and PCNAK164-SUMO (blue) showing
relative linearity of the samples in the Guinier region, indicating lack of aggregation.
(D) A dimensionless Kratky plot indicates
PCNAK164-SUMO being significantly less compact
than PCNAK164-Ub, which is slightly less compact
than PCNAK107-Ub (green).
All P(r) distributions were normalized by dividing
the values by the peak height.

Ubiquitin-PCNA and SUMO-PCNA Conformations from

Conjugated Protein Docking
To further explore the substantial differences between ubiquitin
and SUMO positioning in the modified PCNA complexes, we
constructed computational models of ubiquitin (SUMO) ligated
to PCNA. We used a recently developed protocol for conjugated
protein docking in Rosetta 3.4. The method samples the conformational space available to ligated proteins (Baker et al., 2013;
Leaver-Fay et al., 2011; Saha et al., 2011) using a standard Rosetta Metropolis Monte Carlo random sampling tool (Rohl et al.,
2004). The protocol involved docking a single Ub (SUMO) moiety
to S. cerevisiae PCNA conjugated at the K107 or K164 position,
respectively. The short linker between the modifier and the
PCNA attachment point limited accessible conformations. This
allowed us to locally saturate the PCNA surface with decoys
and achieve extensive sampling. Scoring results from Rosetta
are shown in Figure 2. For PCNAK107-Ub, the all-atom score of
the docking decoys is plotted as a function of the Ca rootmean-square deviation (RMSD) from the lowest-scoring
PCNA-Ub structure. For PCNAK164-SUMO, the crystal structure
3V60 (Armstrong et al., 2012) was used as a reference to
compute the Ca RMSD. Figure 2 reveals a clear distribution
pattern for the decoys, which clustered into bands around specific RMSD values. For PCNA-Ub the lowest-scoring decoys
formed score funnels indicating several preferred positions
for Ub to reside on the surface of PCNA. The majority of
PCNAK107-Ub decoys had Rosetta energy scores ranging
from 1360 to 1100. For PCNAK164-SUMO most scores were
found in a much narrower range between 1030 and 920.
Collectively, the higher Rosetta scores, the narrow score range,
and the absence of low-scoring outliers for the PCNA-SUMO
complex implied significantly weaker association between
PCNA and SUMO in comparison with ubiquitin. Next, we

selected the lowest-scoring structurally

distinct models from the Rosetta output
(Figure 2). These docking positions (one
chosen from each score funnel) were
refined using all-atom explicit solvent molecular dynamics (MD) with the program
NAMD (Kale et al., 1999; Phillips et al.,
2005). The MD simulation and analysis allowed us to eliminate conformations where the Ub or SUMO moiety interacted primarily through polar contacts and consequently
moved away from the PCNA surface during MD refinement. Nine
unique docked positions for each system were selected to build
triplet structures of modified PCNA. These structures after the
MD refinements (sorted by Rosetta score) are presented in
Figure S2.
Optimal Positions of Ubiquitin/SUMO on PCNA from
Minimal Ensemble Search against Experimental
SAXS Data
As the experimentally tested PCNA would have three modifications per homotrimer, we generated triplet models with three
ubiquitin or three SUMO moieties linked to homotrimeric PCNA
for comparison with the experimental SAXS data. Thirteen positions for PCNAK107-Ub (including the 3L10 crystal structure and
three detached flexible Ub positions identified by averaging from
the MD ensemble) and 13 positions for PCNAK164-SUMO
(including the 3V60 crystal structure and three detached flexible
SUMO positions) were used to build triplet structures for the
modified complexes. All possible combinations of positions
were generated with three Ub or SUMO moieties per PCNA
homotrimer.
SAXS data encode all interatomic distance information even
from low-populated flexible conformations (Hammel, 2012;
Putnam et al., 2007; Schneidman-Duhovny et al., 2010; Tsutakawa et al., 2011). The experimental data can be compared
with scattering predicted from atomic models or ensembles of
atomic models, allowing elimination of models not consistent
with the experimental data. To implement this strategy, we
computed theoretical SAXS profiles for all triplet models of the
modified complexes using the program FoXS and determined
the fit of these profiles to the experimental scattering data. The

Structure 23, 110, April 7, 2015 2015 Elsevier Ltd All rights reserved 3

Please cite this article in press as: Tsutakawa et al., Structurally Distinct Ubiquitin- and Sumo-Modified PCNA: Implications for Their Distinct Roles in
the DNA Damage Response, Structure (2015), http://dx.doi.org/10.1016/j.str.2015.02.008

Figure 2. Rosetta Score versus RMSD Plots for Ubiquitinated and

SUMOylated PCNA
(A) Decoys from PCNAK107-Ub docking are shown in gray. Lowest-scoring
structurally distinct models (selected for building triplets) are shown in red. One
model (blue dot) departed from its binding position during molecular dynamics
(MD) and was not considered for minimal ensemble search (MES) analysis.
(B) Decoys from PCNAK164-SUMO docking are shown in gray. Lowest-scoring
models (selected for building triplets) are shown in red. Four models (blue dots)
departed during MD and were not considered for MES analysis. One partially
flexible position (purple) was subsequently included in MES analysis.

c and cfree (Rambo and Tainer, 2013) values were used to measure goodness of fit of the computed profiles to the experimental
scattering data. Computed c values for PCNAK107-Ub and
PCNAK164-SUMO are plotted in Figures 3A and 3B as a function
of Ca RMSD for each conformation. The single best fit to the
experimental data was c = 0.78 (cfree = 0.82) for PCNAK107-Ub
and c = 2.55 (cfree = 2.97) for PCNAK164-SUMO, respectively.
However, in a flexible system multiple conformations will
contribute to the SAXS profile. Therefore, a minimal ensemble
search (MES) (Pelikan et al., 2009) was used to identify a small
subset of conformations that optimally represent the scattering
data. In the case of PCNAK107-Ub, an excellent agreement to
the experimental curve was obtained (c = 0.74; cfree = 0.78) (Figure 3C; Figure S3A) by a set of two distinct conformers with
ensemble contributions of 59% and 41% (Figure 3G). In these
conformers the Ub occupied primarily docked positions, closely
associating with the PCNA surface. In the case of PCNAK164SUMO, however, the ensemble was dominated by two flexible,
extended conformations (45% and 31% occupancy) and a third
flexible but more compact conformation (24% occupancy) (Figure 3H). MES notably improved the goodness of fit to the PCNA-

SUMO SAXS data (c = 2.16, cfree = 2.41, Figure 3D; Figure S3B).
Conversely, large discrepancies between the crystal structure
3V60 and the solution SAXS profile were observed (c = 6.70,
cfree = 7.65, Figure 3B), indicating that the conformation found
in the crystal is not prevalent in the solution phase ensemble.
In all models selected by MES, we calculated the occupancy
of docked versus extended flexible positions for PCNAK107-Ub
and PCNAK164-SUMO. In the PCNAK107-Ub complex, the ubiquitins were all positioned on the back side of PCNA. Approximately
half of the Ub population was docked along the P loop, similar to
the position observed in the crystal structure. One-third was at
the subunit interface. The remainder (20%) was located near
the central cavity. All three positions allow dsDNA to pass
through PCNA and, intriguingly, would place Ub in proximity to
the dsDNA backbone passing through PCNA (Figure S4). The
attachment position at K107 is on the back face of the clamp
close to the subunit interface. Thus, switching the point of
attachment from K164 to K107 leads to an increased probability
to locate the Ub moiety on the back face of PCNA. Unlike the
PCNAK164-Ub study, the MES did not identify any flexible conformations. The goodness of the fit of these back positions to the
experimental SAXS data was consistent with this observation.
In the case of PCNA-SUMO, extended positions were identified in the MES calculation. As shown in Figure 3D, MES analysis
identifies SUMO with 84% occupancy in two extended flexible
positions. Interestingly, MES did not pick any positions docked
to the PCNA surface despite inclusion of the 3V60 crystallographic position in the MES optimization. This outcome indicates
that SUMO is largely flexible in solution. It is notable that the
docked positions observed by crystallography (Armstrong
et al., 2012; Freudenthal et al., 2011) (3PGE and 3V60) have
significant crystallographic contacts with few intramolecular
contacts.
Detailed Interactions in the MES Identified Models
To identify specific structural features that might drive the closer
association of Ub to PCNA, we inspected the two most populated MES triplet models. The primary position near the P loop
was originally observed in the crystal structure and is based on
a hydrophobic contact of the major protein interacting face
on Ub, Ile44 Ub and Val70Ub with Met188PCNA and Val186PCNA.
The Ub position above the subunit interface was stabilized
through several salt bridges: Arg42Ub-Glu7PCNA, Lys48UbGlu59PCNA, Glu51Ub-Lys5PCNA, and Glu58Ub-Arg61PCNA. The
position near the central cavity had two salt bridges, Asp39UbK107PCNA and Asp58Ub-Lys242PCNA. A stacking of aliphatic
chains between Arg72Ub-Asp109PCNA is also observed at the
interface between PCNA and Ub. This Ub position could also
be explained by a general electrostatic attraction between the
acidic face of ubiquitin moving toward the basic central cavity.
It is notable in the last two positions that Ile44Ub and Val70Ub
are exposed, perhaps promoting an interface for protein-protein
interaction with Ub.
In contrast to the ubiquitin-modified PCNA, our MES results
with SUMOylated PCNA suggest that covalently attached
SUMO actually makes few contacts to PCNA in solution and
adopts extended flexible positions. The crystal structure of
PCNAK164-SUMO revealed that the interface with the modifier
comprises loop regions of PCNA and SUMO, which involve

4 Structure 23, 110, April 7, 2015 2015 Elsevier Ltd All rights reserved

Please cite this article in press as: Tsutakawa et al., Structurally Distinct Ubiquitin- and Sumo-Modified PCNA: Implications for Their Distinct Roles in
the DNA Damage Response, Structure (2015), http://dx.doi.org/10.1016/j.str.2015.02.008

Figure 3. Ub Primarily Adopts Docked Positions in PCNAK107-Ub while SUMO Occupies Extended Positions in PCNAK164-SUMO
An MES produces the best fit to the experimental SAXS data for PCNAK107-Ub and PCNAK164-SUMO.
(A) c values for the triplet PCNAK107-Ub structures plotted against RMSD. Conformations selected by MES are highlighted in blue and magenta, respectively.
(B) c values for the triplet PCNAK164-SUMO structures plotted against RMSD. Conformations selected by MES are highlighted in blue, magenta, and red,
respectively.
(C) Overlaid experimental SAXS profile for PCNAK107-Ub (green), computed profile for 3L10 crystal structure (red dashed line), and computed profile from MES
model (black).
(D) Overlaid experimental SAXS profile for PCNAK164-SUMO (blue), computed profile for 3V60 crystal structure (red dashed line), and computed profile from MES
model (black).
(E) P(r) functions for PCNAK107-Ub (green), 3L10 crystal structure (red dashed line), and MES model (black).
(F) P(r) functions for PCNAK164-SUMO (blue), 3V60 crystal structure (red dashed line), and MES model (black).
(G) The two most populated atomic structures from MES analysis of PCNAK107-Ub in surface representation.
(H) The three most populated atomic structures from MES analysis of PCNAK164-SUMO in surface representation. The P loop is shown in purple; the K107 and
K164 attachment points are depicted in red. PCNA, Ub, and SUMO are shown in gray, green, and blue, respectively. The MES occupancies for the three
conformations are labeled in blue, magenta and red, respectively. All P(r) distributions were normalized by dividing the values by the peak height.

Structure 23, 110, April 7, 2015 2015 Elsevier Ltd All rights reserved 5

Please cite this article in press as: Tsutakawa et al., Structurally Distinct Ubiquitin- and Sumo-Modified PCNA: Implications for Their Distinct Roles in
the DNA Damage Response, Structure (2015), http://dx.doi.org/10.1016/j.str.2015.02.008

Figure 4. Structural Differences in PCNAK107Ub and PCNAK164-SUMO Complexes from

the Anticorrelated Electrostatic Potential of
Ub and SUMO
(A) Ub docked onto PCNA in the most populated
MES positions (P-loop, subunit interface, and
central cavity positions) and SUMO position from
the 3V60 crystal structure. Ub, SUMO, and PCNA
are colored green, blue, and gray, respectively.
The interdomain connector loops (IDCL) and P
loops on PCNA are shown in orange and purple,
respectively. The attachment positions, K107 and
K164 residues, are displayed as red spheres.
(B) Electrostatic potential surfaces corresponding
to the bound positions of Ub and SUMO on PCNA
(P loop, subunit interface, and central cavity, from
left to right for PCNAK107-Ub) and 3V60 structure
for PCNAK164-SUMO. NaCl (100 mM) was introduced to screen the electrostatics mimicking
physiological conditions. The potential varies from
5KBT/e to +5KBT/e and is depicted from red to
blue, respectively.

acidic residues from both sides. Furthermore, the outer surface

of PCNA is overwhelmingly negatively charged. We verified
that these interfacial contacts from the crystal structure could
not be maintained in an MD simulation. The distinct behavior
by Ub and SUMO becomes apparent upon inspection of the surface electrostatic potentials for Ub, SUMO, and PCNA (Figure 4).
For all three dominant docked conformations of PCNAK107-Ub
we observe electrostatic complementarity at the Ub/PCNA interface. By contrast, there is clear repulsion for the PCNAK164SUMO crystal structure where both surfaces are negatively
charged. Thus, it is likely that the docked conformation identified
in the X-ray structure is stabilized by the crystal environment and
is not highly populated in solution.
DISCUSSION
PCNA monoubiquitination and SUMOylation are reversible
dynamic modifications that orchestrate cellular events in
response to DNA damage. How Ub and SUMO associate
with PCNA and how this association may facilitate subsequent
partner interactions has not been fully established. Three general models can be envisioned: (1) simple tethering; (2) structured interface formation; or (3) allosteric response. Under the
simple tethering model, the only function of the modifier (ubiquitin or SUMO) is to provide an additional interaction interface
to the protein partner carrying ubiquitin- or SUMO-interacting
domain. Under the structured interface model, the modifier
and effector protein can form a continuous binding interface
with PCNA. In the allosteric model, Ub or SUMO conjugation
induces conformational change in PCNA, which in turn facilitates recruitment of downstream effectors in the respective
pathways.
This allosteric hypothesis is exemplified by the SUMOinduced ordering of thymine DNA glycosylase (Baba et al.,

2005; Steinacher and Schar, 2005). Our multiple-trajectory MD

simulation results show little change in PCNA structure, and
argue against an allosteric model for the mechanism of PCNA
ubiquitination or SUMOylation in effecting distinct functional responses to DNA damage.
In ubiquitinated PCNA, there appears to be formation of
structured interfaces. Ubiquitin is stabilized against the PCNA
surface and is able to adopt several discrete positions relative
to the PCNA ring (Figure 5). PCNAK107-Ub and PCNAK164-Ub
have distinct distributions of conformations. The P-loop position
observed in the crystal structure is found in both, but the predominant position for the PCNAK164-Ub complex, which is along
the side of the PCNA ring, is not found for PCNAK107-Ub. The
side position would not be easily accessible when the Ub
is linked via K107. The newly identified positions found for
PCNAK107-Ub, on the back side of PCNA, are interesting, as
they are close to where the DNA would pass though the
PCNA ring and as the major protein-protein interaction interface
for Ub is exposed. To describe this situation where covalently
bound ubiquitin occupies docked positions against the
PCNA surface, we use the concept of segmental flexibility
and define it as the stabilization of functionally relevant positions in otherwise flexible systems. Both PCNAK107-Ub and
PCNAK164-Ub exhibit such segmental flexibility. Ubiquitin in
docked positions could control the orientation of effector proteins upon initial encounter of PCNA-Ub and determine the
side from which partner proteins approach PCNA. Thus, our
findings highlight the importance of spatial constraints for ubiquitin, which could allow Ub to be recognized by TLS polymerases in combination with binding to PCNA. Furthermore,
segmental flexibility is likely a common characteristic of eukaryotic ubiquitin regulatory systems. As other lysines on the edge
of the PCNA ring are also uniquely modified (Povlsen et al.,
2012; Xu et al., 2010), we postulate that these distinct covalent

6 Structure 23, 110, April 7, 2015 2015 Elsevier Ltd All rights reserved

Please cite this article in press as: Tsutakawa et al., Structurally Distinct Ubiquitin- and Sumo-Modified PCNA: Implications for Their Distinct Roles in
the DNA Damage Response, Structure (2015), http://dx.doi.org/10.1016/j.str.2015.02.008

Figure 5. Distinct Types of Interfaces Are

Exposed Depending on the Different Modes
of Association of Ubiquitin and SUMO to
PCNA
The PCNA trimer is shown in gray, the Ub modifier
in green, and SUMO in blue; the attachment positions are indicated by red dots; curved arrows
indicate flexible attachment. Approximate occupancy (%) of the identified distinct positions of the
modifiers on PCNA is given below each model.

modification sites provide a specific conformational range for

the downstream effector proteins, such as we observed for
K107 and K164.
By contrast, the PCNAK164-SUMO complex adopted
extended flexible conformations wherein SUMO was covalently
bound to K164 but was not otherwise interacting with the surface
of PCNA (Figure 5). Therefore, for SUMOylated PCNA recruitment, our results support simple tethering, consistent with the
model of the Srs2 interaction with SUMOylated PCNA that adopted two positions (Armstrong et al., 2012). Flexible tethering of
SUMO is required for successful Srs2 recruitment and is consistent with the observation that Srs2 could not interact in cis with
both SUMO and PCNA in the conformation of PCNAK164SUMO crystal structure. In addition, the structural requirement
for tethered flexibility is supported by the fact that the K127
and K164 SUMOylation sites, more than 30 A apart from each
other, were genetically equivalent in their ability to interact with
Srs2 (Papouli et al., 2005; Pfander et al., 2005). This implies
that the crystallographic position is not required for Srs2
interaction.
The structurally distinct modes of association of Ub and
SUMO to the sliding clamp (Figure 5) present radically different
conformations and accessibility of the interacting surface for
partner proteins. PCNA-SUMO binds two structurally distinct
proteins: Srs2 and Rad18. Srs2 is an antirecombinase that
has been shown to possess two activities to prevent recombination. The first is the ability to disrupt Rad51 nucleoprotein
filaments (Krejci et al., 2003; Veaute et al., 2003), which could
be on the same partially duplex DNA passing through the
PCNA ring, on the partially duplex sister strand, or on another
partial duplex in the vicinity. The second is the ability to inhibit
DNA polymerases from extending recombination intermediates
(Burkovics et al., 2013). Again, these recombination intermediates could be on the same duplex of the PCNA-SUMO, on the
sister partial, or on another duplex in the vicinity of PCNA.
For both of these Srs2 activities, a greater degree of flexibility
is important for delivering the catalytic domain of Srs2 to the
appropriate location. Rad18 is a SUMO-directed E3 ubiquitin
ligase responsible for catalyzing the ubiquitylation of PCNA.
The SUMO moiety on PCNA-SUMO binds Rad18 and positions
it to allow ubiquitylation of K164 on other PCNA subunits of
the trimer (Parker and Ulrich, 2012). Since Rad18 is structurally

unrelated to Srs2, PCNA-SUMO must

bind different, structurally unrelated proteins and deliver their catalytic domains
to various positions around the PCNA
ring. Thus, an expanded spatial range
would be required to accommodate the interactions and activities of binding partners.
PCNA-Ub, by contrast, is only known to bind one class of
structurally related proteins, the Y-family translesion synthesis
polymerases. The basis of the interactions between PCNA-Ub
and the various Y-family polymerases is similar. The conserved
ubiquitin-binding motifs (UBMs) and ubiquitin-binding zinc finger
domains (UBZs) of these polymerases are located in their intrinsically disordered C-terminal regions. Moreover, the catalytic
domains of these polymerases only function on the partial duplex
on the front face of the PCNA ring. Thus, consistent with our
structural results, a more restricted spatial range and tighter
binding of the Ub modifier to PCNA is required to support the
function of these complexes. One indicator for the biological significance of the newly identified PCNA-Ub interfaces is slight
sensitivity to DNA damage, which is consistent with a partial or
complete defect in translesion synthesis. Substitutions of Ub-interacting residues in yeast PCNA all confer an increased sensitivity to DNA-damaging agents. For example, the ubiquitin at
the subunit interface forms a hydrogen bond with Lys5 of
PCNA and is close (within 4 A) to Arg61 of PCNA. A substitution
of either Lys5 or Arg61 increases the sensitivity to UV radiation,
hydroxyurea (HU), and methyl methanesulfonate (MMS) (Ayyagari et al., 1995). Similarly, the ubiquitin near the PCNA cavity
forms hydrogen bonds with Asp109 and is close to Glu189. A
substitution of Asp109 increases the sensitivity to UV radiation
and MMS, and a substitution of Glu189 increased the sensitivity
to UV radiation (Ayyagari et al., 1995). Lastly, the Ub position
along the PCNA ring, identified as the predominant position
for PCNAK164-Ub, includes hydrogen bonds with Ser177 and
Ser179 of PCNA. Although to our knowledge no substitutions
of these residues have been reported, a substitution in Gly178,
located in between these two residues, increases the sensitivity
to UV radiation and leads to a complete loss of error-prone
translesion synthesis (Zhang et al., 2006). In none of these substitutions is the growth of the cells affected.
More generally, our work underscores the power of hybrid
structural methods, and our results afford insights into how
PCNA posttranslational modifications by Ub and SUMO provide
both the necessary specificity and flexibility to regulate recruitment and coordinated actions of effector proteins to promote
genomic stability.

Structure 23, 110, April 7, 2015 2015 Elsevier Ltd All rights reserved 7

Please cite this article in press as: Tsutakawa et al., Structurally Distinct Ubiquitin- and Sumo-Modified PCNA: Implications for Their Distinct Roles in
the DNA Damage Response, Structure (2015), http://dx.doi.org/10.1016/j.str.2015.02.008

EXPERIMENTAL PROCEDURES
SAXS Analysis of PCNAK164-SUMO and PCNAK107-Ub
SUMOylated and crosslinked yeast PCNA were purified using established protocols (Chen et al., 2010; Freudenthal et al., 2010, 2011). SAXS data were
collected at the SIBYLS 12.3.1 beamline at the Advanced Light Source, Lawrence Berkeley National Laboratory (Classen et al., 2013; Classen et al., 2010;
Hura et al., 2009). Scattering measurements were performed on 20-ml samples
at 15 C (PCNAK164-SUMO) or 22 C (PCNAK107-Ub) loaded into a heliumpurged sample chamber, 1.5 m from the Mar165 detector. Prior to data collection, modified PCNA were purified by size-exclusion chromatography on a
24-ml Superose6 column equilibrated in 20 mM Tris (pH 7.5), 150 mM NaCl,
and 5% glycerol. Data were collected on both the original gel filtration fractions
and samples concentrated 23 from individual fractions (Table S2; Figure S1).
Fractions prior to the void volume and concentrator eluates were used
for buffer subtraction. Sequential exposures (0.5, 0.5, 5, and 0.5 s for
PCNAK164-SUMO and 0.5, 0,5, 2, 5, and 0.5 s for PCNAK107-Ub) were taken
at 12 keV to maximize signal to noise with visual checks for radiation-induced
damage to the protein. Although scattering of the split-fusion PCNAK164SUMO showed radiation-induced aggregation, the scattering of the crosslinked PCNAK107-Ub showed a decrease in slope, indicating that the
molecules in solution were becoming smaller, likely due to irradiation breaking
the disulfide bond. The first and second 0.5-s exposures overlapped, so the
first exposure was assumed to have only minimal damage. The best data,
based on signal-to-noise ratio and Guinier analysis, were collected on splitfusion PCNAK164-SUMO (2.7 mg/ml, 0.5-s exposure) and crosslinked
PCNAK107-Ub (1.1 mg/ml, 0.5-s exposure). Scattering data including c2free
(Rambo and Tainer, 2013) were analyzed using SCATTER available at beamline 12.3.1 (Classen et al., 2013; Rambo and Tainer, 2011). The Porod exponent was determined from a linear regression analysis (I versus q) of the top
of the first peak in the Porod-Debye plot (q4 3 I(q) versus q4) of the scattering
data. P(r) plots were calculated using Gnom implemented in the ATSAS
package.
Computational Models and Protocols
To model the PCNAK107-Ub and PCNAK164-SUMO complexes, we used a
chemically conjugated docking protocol written as a part of the Rosetta 3.4
suite (Baker et al., 2013; Leaver-Fay et al., 2011; Saha et al., 2011). We used
the standard Rosetta scoring function, score12, to rank and select the topscoring models (Kuhlman and Baker, 2000; Rohl et al., 2004). The protocol
was designed to search the conformational space available to ubiquitin
chemically conjugated via an isopeptide bond and sample rotations about
torsional angles in the vicinity of the isopeptide bond. To initiate the sampling
for PCNAK107-Ub, we used the 1UBQ structure for ubiquitin (Vijay-Kumar et al.,
1987) and 1PLQ structure for yeast PCNA (Krishna et al., 1994). For the
PCNAK164-SUMO complex, we used the structure of SUMOylated PCNA
(PDB ID: 3V60) (Armstrong et al., 2012). The initial PDB structures were minimized with the Rosetta relax protocol while using all-heavy-atom constraints
prior to conjugated docking. For the isopeptide linker, protocol UBQ_Gp_
LYX-cterm was used (Baker et al., 2013). Torsions allowed to change included
the c angles of Lys107 or Lys164 of PCNA, the isopeptide bond, and both F
and J angles for the Gly76, Gly75, and Arg74 of ubiquitin (Gly98, Gly97, and
Ile96 of SUMO). Sampling was performed with a standard Rosetta Metropolis
Monte Carlo search protocol (Rohl et al., 2004). Results were automatically
filtered according to the solvent accessible surface area (SASA) buried at
the protein interface (>500 A2) and total score of the docked complex (<0).
In total the sampling produced 4,791 decoys (i.e. Ub/SUMO docking poses)
for PCNA-Ub and 4,499 decoys for PCNA-SUMO using 20,000 Monte Carlo
cycles per trajectory. All Rosetta docking calculations were performed on
the Stampede supercomputer at the Texas Advanced Computing Center
(TACC).
Model Refinement
Outliers with best scores from Rosetta docking were refined using all-atom
explicit solvent MD. In setting up for MD hydrogen atoms, counterions (Na+)
and TIP3P solvent (Jorgensen et al., 1983) were introduced using the XLeap
module in AMBER 12 (Case et al., 2005; Cornell et al., 1995; Duan et al.,
2003). In addition, 100 mM NaCl was introduced to mimic physiological con-

ditions. The systems were then minimized for 5,000 steps with backbone
atoms fixed followed by 5,000 steps of minimization with harmonic restraints
to remove unfavorable contacts. The systems were then gradually brought
up to 300 K and run for 50 ps in the NVT ensemble while keeping the protein
backbone restrained. The equilibration was continued for another 2 ns in
the NPT ensemble and the harmonic restraints were gradually released. The
60-ns production simulations were performed in the NPT ensemble (1 atm
and 300 K) without constraints. A cutoff of 10 A was used for the short-range
non-bonded interactions with a switching function at 8.5 A. The long-range
electrostatic interactions were treated with a smooth particle mesh Ewald
method (Essmann et al., 1995). The r-RESPA multiple timestep method (Tuckerman et al., 1992) was adopted with a time step of 2 fs for bonded interactions, 2 fs for short-range non-bonded interactions, and 4 fs for long-range
electrostatic interactions. Bonds between hydrogen atoms and heavy atoms
of the protein were constrained with the SHAKE algorithm. All simulations
were performed with the NAMD 2.8 code (Kale et al., 1999; Phillips et al.,
2005) using the AMBER Parm99SB force field. Models that departed substantially from the initial Rosetta docking position during MD were eliminated from
further consideration.
MES
Thirteen positions for PCNAK107-Ub (including the 3L10 X-ray structure and
three detached flexible Ub positions identified by averaging from the MD trajectories) and 13 positions for PCNAK164-SUMO (including the 3V60 crystal
structure and three detached flexible SUMO positions) were used to build
triplet structures for the modified complexes. All possible combinations of
positions were generated with three Ub or SUMO moieties per PCNA homotrimer. Thus, we produced a final set of 862 PCNA-Ub triplet models and
1728 PCNA-SUMO triplet models. Computation of scattering profiles from
the models and comparison with the experimental data used the FoXS code
(Schneidman-Duhovny et al., 2010, 2013). The coexistence of different conformations that contribute to the experimental scattering curve had to be taken
into account by considering multiple Ub or SUMO positions. An algorithm
developed by Pelikan et al. (2009) was used to search for the minimal
ensemble of conformations from the pool of all Rosetta-generated triplet
models. The MES included only the minimum set of conformations necessary
to minimize cfree.
SUPPLEMENTAL INFORMATION
Supplemental Information includes two tables and four figures and can be
found with this article online at http://dx.doi.org/10.1016/j.str.2015.02.008.
AUTHOR CONTRIBUTIONS
S.E.T., M.T.W., Z.Z., J.A.T., and I.I. designed research; S.E.T., C.Y., X.X., and
I.I. performed research; B.D.F. and C.P.W., M.T.W., K.Y., and Z.Z. contributed
new reagents/analytic tools; S.E.T., C.Y., X.X., and I.I. analyzed data; and
S.E.T., C.Y., J.A.T, M.T.W., Z.Z., and I.I. wrote the paper.
ACKNOWLEDGMENTS
This work was supported by an NSF CAREER grant MCB-1149521 (to I.I.),
Georgia State University start-up funds (to I.I.), P01 CA092584 (NCI to
J.A.T.), R01 CA081967 (NCI to J.A.T.), NSF Grant MCB-0953764 (to Z.Z.),
and R01 GM108027 (to M.T.W.). Computational resources were provided in
part by a National Science Foundation XSEDE allocation (CHE110042) and
through an allocation at National Energy Research Scientific Computing Center (NERSC) supported by the U.S. Department of Energy Office of Science
(contract DE-AC02-05CH11231). SAXS data were collected at BL12.3.1 at
the Advanced Light Source (ALS), supported by the Integrated Diffraction
Analysis Technologies (IDAT) program (DOE/BER), by DOE contract DEAC02-05CH11231, and by NIH MINOS (R01GM105404).
Received: September 18, 2014
Revised: January 28, 2015
Accepted: February 9, 2015
Published: March 12, 2015

8 Structure 23, 110, April 7, 2015 2015 Elsevier Ltd All rights reserved

Please cite this article in press as: Tsutakawa et al., Structurally Distinct Ubiquitin- and Sumo-Modified PCNA: Implications for Their Distinct Roles in
the DNA Damage Response, Structure (2015), http://dx.doi.org/10.1016/j.str.2015.02.008

REFERENCES
Armstrong, A.A., Mohideen, F., and Lima, C.D. (2012). Recognition of SUMOmodified PCNA requires tandem receptor motifs in Srs2. Nature 483, 5963.
Ayyagari, R., Impellizzeri, K.J., Yoder, B.L., Gary, S.L., and Burgers, P.M.
(1995). A mutational analysis of the yeast proliferating cell nuclear antigen indicates distinct roles in DNA replication and DNA repair. Mol. Cell. Biol. 15,
44204429.
Baba, D., Maita, N., Jee, J.G., Uchimura, Y., Saitoh, H., Sugasawa, K.,
Hanaoka, F., Tochio, H., Hiroaki, H., and Shirakawa, M. (2005). Crystal structure of thymine DNA glycosylase conjugated to SUMO-1. Nature 435,
979982.
Baker, R., Lewis, S.M., Sasaki, A.T., Wilkerson, E.M., Locasale, J.W., Cantley,
L.C., Kuhlman, B., Dohlman, H.G., and Campbell, S.L. (2013). Site-specific
monoubiquitination activates Ras by impeding GTPase-activating protein
function. Nat. Struct. Mol. Biol. 20, 4652.
Bienko, M., Green, C.M., Crosetto, N., Rudolf, F., Zapart, G., Coull, B.,
Kannouche, P., Wider, G., Peter, M., Lehmann, A.R., et al. (2005). Ubiquitinbinding domains in Y-family polymerases regulate translesion synthesis.
Science 310, 18211824.
Burkovics, P., Sebesta, M., Sisakova, A., Plault, N., Szukacsov, V., Robert, T.,
Pinter, L., Marini, V., Kolesar, P., Haracska, L., et al. (2013). Srs2 mediates
PCNA-SUMO-dependent inhibition of DNA repair synthesis. EMBO J. 32,
742755.
Case, D.A., Cheatham, T.E., Darden, T., Gohlke, H., Luo, R., Merz, K.M.,
Onufriev, A., Simmerling, C., Wang, B., and Woods, R.J. (2005). The Amber
biomolecular simulation programs. J. Comput. Chem. 26, 16681688.
Chen, J.J., Ai, Y.X., Wang, J.L., Haracska, L., and Zhuang, Z.H. (2010).
Chemically ubiquitylated PCNA as a probe for eukaryotic translesion DNA
synthesis. Nat. Chem. Biol. 6, 270272.
Classen, S., Rodic, I., Holton, J., Hura, G.L., Hammel, M., and Tainer, J.A.
(2010). Software for the high-throughput collection of SAXS data using an
enhanced Blu-Ice/DCS control system. J. Synchrotron Radiat. 17, 774781.
Classen, S., Hura, G.L., Holton, J.M., Rambo, R.P., Rodic, I., McGuire, P.J.,
Dyer, K., Hammel, M., Meigs, G., Frankel, K.A., and Tainer, J.A. (2013).
Implementation and performance of SIBYLS: a dual endstation small-angle
X-ray scattering and macromolecular crystallography beamline at the
Advanced Light Source. J. Appl. Crystallogr. 46, 113.
Cornell, W.D., Cieplak, P., Bayly, C.I., Gould, I.R., Merz, K.M., Ferguson, D.M.,
Spellmeyer, D.C., Fox, T., Caldwell, J.W., and Kollman, P.A. (1995). A 2nd generation force-field for the simulation of proteins, nucleic-acids, and organicmolecules. J. Am. Chem. Soc. 117, 51795197.
Das-Bradoo, S., Nguyen, H.D., Wood, J.L., Ricke, R.M., Haworth, J.C., and
Bielinsky, A.K. (2010). Defects in DNA ligase I trigger PCNA ubiquitylation at
Lys 107. Nat. Cell Biol. 12, 7479.
Deshpande, R.A., Williams, G.J., Limbo, O., Williams, R.S., Kuhnlein, J., Lee,
J.H., Classen, S., Guenther, G., Russell, P., Tainer, J.A., and Paull, T.T.
(2014). ATP-driven Rad50 conformations regulate DNA tethering, end resection, and ATM checkpoint signaling. EMBO J. 33, 482500.
Duan, Y., Wu, C., Chowdhury, S., Lee, M.C., Xiong, G.M., Zhang, W., Yang, R.,
Cieplak, P., Luo, R., Lee, T., et al. (2003). A point-charge force field for molecular mechanics simulations of proteins based on condensed-phase quantum
mechanical calculations. J. Comput. Chem. 24, 19992012.
Durand, D., Vives, C., Cannella, D., Perez, J., Pebay-Peyroula, E., Vachette, P.,
and Fieschi, F. (2010). NADPH oxidase activator p67(phox) behaves in solution
as a multidomain protein with semi-flexible linkers. J. Struct. Biol. 169, 4553.
Essmann, U., Perera, L., Berkowitz, M.L., Darden, T., Lee, H., and Pedersen,
L.G. (1995). A smooth particle mesh Ewald method. J. Chem. Phys. 103,
85778593.
Freudenthal, B.D., Gakhar, L., Ramaswamy, S., and Washington, M.T. (2010).
Structure of monoubiquitinated PCNA and implications for translesion synthesis and DNA polymerase exchange. Nat. Struct. Mol. Biol. 17, 479484.

Freudenthal, B.D., Brogie, J.E., Gakhar, L., Kondratick, C.M., and Washington,
M.T. (2011). Crystal structure of SUMO-modified proliferating cell nuclear
antigen. J. Mol. Biol. 406, 917.
Hammel, M. (2012). Validation of macromolecular flexibility in solution by
small-angle X-ray scattering (SAXS). Eur. Biophys. J. 41, 789799.
Hicke, L. (2001). Protein regulation by monoubiquitin. Nat. Rev. Mol. Cell Biol.
2, 195201.
Hoege, C., Pfander, B., Moldovan, G.L., Pyrowolakis, G., and Jentsch, S.
(2002). RAD6-dependent DNA repair is linked to modification of PCNA by
ubiquitin and SUMO. Nature 419, 135141.
Hura, G.L., Menon, A.L., Hammel, M., Rambo, R.P., Poole, F.L., 2nd,
Tsutakawa, S.E., Jenney, F.E., Jr., Classen, S., Frankel, K.A., Hopkins, R.C.,
et al. (2009). Robust, high-throughput solution structural analyses by small
angle X-ray scattering (SAXS). Nat. Methods 6, 606612.
Ivanov, I., Chapados, B.R., McCammon, J.A., and Tainer, J.A. (2006).
Proliferating cell nuclear antigen loaded onto double-stranded DNA: dynamics, minor groove interactions and functional implications. Nucleic Acids
Res. 34, 60236033.
Jonsson, Z.O., and Hubscher, U. (1997). Proliferating cell nuclear antigen:
more than a clamp for DNA polymerases. Bioessays 19, 967975.
Jorgensen, W.L., Chandrasekhar, J., Madura, J.D., Impey, R.W., and Klein,
M.L. (1983). Comparison of simple potential functions for simulating liquid
water. J. Chem. Phys. 79, 926935.
Kale, L., Skeel, R., Bhandarkar, M., Brunner, R., Gursoy, A., Krawetz, N.,
Phillips, J., Shinozaki, A., Varadarajan, K., and Schulten, K. (1999). NAMD2:
greater scalability for parallel molecular dynamics. J. Comput. Physiol. 151,
283312.
Kelman, Z. (1997). PCNA: structure, functions and interactions. Oncogene 14,
629640.
Krejci, L., Van Komen, S., Li, Y., Villemain, J., Reddy, M.S., Klein, H.,
Ellenberger, T., and Sung, P. (2003). DNA helicase Srs2 disrupts the Rad51
presynaptic filament. Nature 423, 305309.
Krishna, T.S., Kong, X.P., Gary, S., Burgers, P.M., and Kuriyan, J. (1994).
Crystal structure of the eukaryotic DNA polymerase processivity factor
PCNA. Cell 79, 12331243.
Kuhlman, B., and Baker, D. (2000). Native protein sequences are close to
optimal for their structures. Proc. Natl. Acad. Sci. USA 97, 1038310388.
Leaver-Fay, A., Tyka, M., Lewis, S.M., Lange, O.F., Thompson, J., Jacak, R.,
Kaufman, K., Renfrew, P.D., Smith, C.A., Sheffler, W., et al. (2011).
ROSETTA3: an object-oriented software suite for the simulation and design
of macromolecules. Methods Enzymol. 487, 545574.
Maga, G., and Hubscher, U. (2003). Proliferating cell nuclear antigen (PCNA): a
dancer with many partners. J. Cell Sci. 116, 30513060.
Moldovan, G.L., Pfander, B., and Jentsch, S. (2007). PCNA, the maestro of the
replication fork. Cell 129, 665679.
Nguyen, H.D., Becker, J., Thu, Y.M., Costanzo, M., Koch, E.N., Smith, S.,
Myung, K., Myers, C.L., Boone, C., and Bielinsky, A.K. (2013). Unligated
Okazaki fragments induce PCNA ubiquitination and a requirement for
Rad59-dependent replication fork progression. PloS One 8, e66379.
Papouli, E., Chen, S.H., Davies, A.A., Huttner, D., Krejci, L., Sung, P., and
Ulrich, H.D. (2005). Crosstalk between SUMO and ubiquitin on PCNA is mediated by recruitment of the helicase Srs2p. Mol. Cell 19, 123133.
Parker, J.L., and Ulrich, H.D. (2012). A SUMO-interacting motif activates
budding yeast ubiquitin ligase Rad18 towards SUMO-modified PCNA.
Nucleic Acids Res. 40, 1138011388.
Pelikan, M., Hura, G.L., and Hammel, M. (2009). Structure and flexibility within
proteins as identified through small angle X-ray scattering. Gen. Physiol.
Biophys. 28, 174189.
Perry, J.J.P., Tainer, J.A., and Boddy, M.N. (2008). A SIM-ultaneous role for
SUMO and ubiquitin. Trends Biochem. Sci. 33, 201208.
Pfander, B., Moldovan, G.L., Sacher, M., Hoege, C., and Jentsch, S. (2005).
SUMO-modified PCNA recruits Srs2 to prevent recombination during S phase.
Nature 436, 428433.

Structure 23, 110, April 7, 2015 2015 Elsevier Ltd All rights reserved 9

Please cite this article in press as: Tsutakawa et al., Structurally Distinct Ubiquitin- and Sumo-Modified PCNA: Implications for Their Distinct Roles in
the DNA Damage Response, Structure (2015), http://dx.doi.org/10.1016/j.str.2015.02.008

Phillips, J.C., Braun, R., Wang, W., Gumbart, J., Tajkhorshid, E., Villa, E.,
Chipot, C., Skeel, R.D., Kale, L., and Schulten, K. (2005). Scalable molecular
dynamics with NAMD. J. Comput. Chem. 26, 17811802.

Steinacher, R., and Schar, P. (2005). Functionality of human thymine DNA

glycosylase requires SUMO-regulated changes in protein conformation.
Curr. Biol. 15, 616623.

Povlsen, L.K., Beli, P., Wagner, S.A., Poulsen, S.L., Sylvestersen, K.B.,
Poulsen, J.W., Nielsen, M.L., Bekker-Jensen, S., Mailand, N., and
Choudhary, C. (2012). Systems-wide analysis of ubiquitylation dynamics
reveals a key role for PAF15 ubiquitylation in DNA-damage bypass. Nat. Cell
Biol. 14, 10891098.

Stelter, P., and Ulrich, H.D. (2003). Control of spontaneous and damageinduced mutagenesis by SUMO and ubiquitin conjugation. Nature 425,
188191.

Putnam, C.D., Hammel, M., Hura, G.L., and Tainer, J.A. (2007). X-ray solution
scattering (SAXS) combined with crystallography and computation: defining
accurate macromolecular structures, conformations and assemblies in solution. Q. Rev. Biophys. 40, 191285.

Tsutakawa, S.E., Van Wynsberghe, A.W., Freudenthal, B.D., Weinacht, C.P.,

Gakhar, L., Washington, M.T., Zhuang, Z., Tainer, J.A., and Ivanov, I. (2011).
Solution X-ray scattering combined with computational modeling reveals
multiple conformations of covalently bound ubiquitin on PCNA. Proc. Natl.
Acad. Sci. USA 108, 1767217677.

Rambo, R.P., and Tainer, J.A. (2010). Bridging the solution divide: comprehensive structural analyses of dynamic RNA, DNA, and protein assemblies by
small-angle X-ray scattering. Curr. Opin. Struct. Biol. 20, 128137.
Rambo, R.P., and Tainer, J.A. (2011). Characterizing flexible and intrinsically
unstructured biological macromolecules by SAS using the Porod-Debye law.
Biopolymers 95, 559571.
Rambo, R.P., and Tainer, J.A. (2013). Accurate assessment of mass, models
and resolution by small-angle scattering. Nature 496, 477481.
Reyes, F.E., Schwartz, C.R., Tainer, J.A., and Rambo, R.P. (2014). Methods for
using new conceptual tools and parameters to assess RNA structure by smallangle X-ray scattering. Methods Enzymol. 549, 235263.
Rohl, C.A., Strauss, C.E., Misura, K.M., and Baker, D. (2004). Protein structure
prediction using Rosetta. Methods Enzymol. 383, 6693.

Stolimenov, I., and Helleday, T. (2009). PCNA on the crossroad of cancer.

Biochem. Soc. Trans. 37, 605613.

Tuckerman, M., Berne, B.J., and Martyna, G.J. (1992). Reversible multiple time
scale molecular-dynamics. J. Chem. Phys. 97, 19902001.
Ulrich, H.D. (2005). Mutual interactions between the SUMO and ubiquitin
systems: a plea of no contest. Trends Cell Biol. 15, 525532.
Ulrich, H.D. (2012). Ubiquitin and SUMO in DNA repair at a glance. J. Cell Sci.
125, 249254.
Ulrich, H.D., and Walden, H. (2010). Ubiquitin signalling in DNA replication and
repair. Nat. Rev. Mol. Cell Biol. 11, 479489.
Veaute, X., Jeusset, J., Soustelle, C., Kowalczykowski, S.C., Le Cam, E., and
Fabre, F. (2003). The Srs2 helicase prevents recombination by disrupting
Rad51 nucleoprotein filaments. Nature 423, 309312.
Vijay-Kumar, S., Bugg, C.E., and Cook, W.J. (1987). Structure of ubiquitin
refined at 1.8 A resolution. J. Mol. Biol. 194, 531544.

Saha, A., Lewis, S., Kleiger, G., Kuhlman, B., and Deshaies, R.J. (2011).
Essential role for ubiquitin-ubiquitin-conjugating enzyme interaction in ubiquitin discharge from Cdc34 to substrate. Mol. Cell 42, 7583.

Xu, G., Paige, J.S., and Jaffrey, S.R. (2010). Global analysis of lysine ubiquitination by ubiquitin remnant immunoaffinity profiling. Nat. Biotechnol. 28,
868873.

Schneidman-Duhovny, D., Hammel, M., and Sali, A. (2010). FoXS: a web

server for rapid computation and fitting of SAXS profiles. Nucleic Acids Res.
38, W540W544.

Yang, K., Weinacht, C.P., and Zhuang, Z. (2013). Regulatory role of ubiquitin in
eukaryotic DNA translesion synthesis. Biochemistry 52, 32173228.

Schneidman-Duhovny, D., Hammel, M., Tainer, J.A., and Sali, A. (2013).

Accurate SAXS profile computation and its assessment by contrast variation
experiments. Biophys. J. 105, 962974.

Zhang, H., Gibbs, P.E., and Lawrence, C.W. (2006). The Saccharomyces cerevisiae rev6-1 mutation, which inhibits both the lesion bypass and the recombination mode of DNA damage tolerance, is an allele of POL30, encoding
proliferating cell nuclear antigen. Genetics 173, 19831989.

10 Structure 23, 110, April 7, 2015 2015 Elsevier Ltd All rights reserved