Vanderbilt University
Vanderbilt Institute of Chemical Biology

home research discoveries core facilities training & research opportunities seminars & events news giving contact


Discoveries Featured

A New Structural Framework for Replication Protein A (RPA)

By: Carol A. Rouzer, VICB Communications
Published: January 28, 2013

Small angle X-ray scattering provides insight into the structure and function of RPA, a potential new target for anticancer therapy.

Replication protein A (RPA) is a heterotrimeric multidomain protein that binds single-stranded DNA (ssDNA) and serves as a scaffold for the assembly of a large number of other proteins involved in DNA replication, damage response, and repair. RPA protects ssDNA from the action of endonucleases and prevents it from forming nonproductive secondary structures. As the primary ssDNA-binding protein in eukaryotic cells, RPA is required for nearly all DNA processing events. Consequently, understanding how RPA links the binding of ssDNA to its role as a scaffold for DNA processing proteins is of critical importance. Now, Vanderbilt Institute of Chemical Biology (VICB) and Center for Structural Biology member Walter Chazin and his laboratory provide new insight into this question, [C. A. Brosey, et al. (2013) Nucleic Acids Research, published online January 8, DOI:10.1093/nar/gks1332].

The RPA heterotrimer comprises three subunits, designated RPA70, RPA32, and RPA14. RPA70 and RPA32 are multidomain proteins, including the 70N, 70A, 70B, and 70C domains for RPA70 and the 32N, 32D, and 32C domains for RPA32 (Figure 1A). Association between RPA14, the 32D domain of RPA32, and the 70C domain of RPA70 forms the RPA trimer core. The remaining subunit domains are connected to this core via flexible linkers, yielding a complex modular structure capable of assuming a wide range of architectures (Figure 1B). This complexity and flexibility make RPA a poor subject for structural elucidation by most current methods (Figure 2). Structures of the domains have been determined by X-ray crystallography and NMR, but these data provide only limited information regarding the structure and movements of the entire protein as well as interactions with ssDNA and other proteins.

Figure 1. Structure of RPA. (A) RPA is a heterotrimeric protein comprising the RPA70, RPA32, and RPA14 subunits. RPA70 contains 4 domains designated 70N, 70A, 70B, and 70C. RPA32 contains the 32D and 32C domains. (B) Assembly of the complete RPA trimer requires interaction between RPA14, the 32D domain of RPA32 and the 70C domain of RPA70. The remaining domains of each subunit are tethered to the protein via flexible linker regions. (C) To simplify the data analysis, only the portion of RPA that directly binds ssDNA (RPA-DBC) was used for SAXS studies. This includes the trimer core, the 70A and 70B domains of RPA70, and the linker regions, BC and AB.


Figure 2.Ribbon model of the structure of RPA domains showing the portions of the molecule for which crystal structure data are available. The dotted lines connecting these domains indicate flexible linkers that are the source of the dynamic architecture of the complete protein. Image kindly provided by Walter Chazin.

To better understand RPA’s interaction with ssDNA in solution, the Chazin lab used small angle X-ray scattering (SAXS) combined with small angle neutron scattering and computational modeling. Scattering techniques provides low resolution (10 Å to 50 Å) information on folding, aggregation, conformation, and assembly of proteins in solution. When a highly collimated beam of X-rays passes through the protein sample, the X-rays are deflected due to their interactions with pairs of electrons in the protein molecule (Figure 3). Analysis of the pattern of scattered X-rays reveals information about the molecule, such as:

  • The radius of gyration (Rg), which is equal to the square root of the averaged squared distance of each scattering electron from the protein’s center.
  • The maximum distance (Dmax) between two scattering electrons in the particle.
  • The distance distribution function (P(r)), which describes the distribution of distances between pairs of electrons within the protein.


Figure 3. Basic approach for acquisition of SAXS data. A highly collimated X-ray beam is passed through the sample containing the protein in solution. The X-rays, which are scattered due to their interactions with electrons in the protein, are recorded on a detector.

Prior studies of the interaction of RPA with ssDNA had suggested three distinct binding modes. The first was thought to involve an interaction between the 70A and 70B domains and a stretch of ssDNA about 10 nucleotides (nt) long. An intermediate mode, requiring a 12 to 23 nt stretch of ssDNA, was assumed to engage 70C along with 70A and 70B. The third binding mode, which requires 24 to 30 nt of ssDNA, was believed to incorporate the 32D domain of the trimer core in the complex. To further explore these multimodal binding interactions, the Chazin lab prepared samples of the RPA DNA-binding core (RPA-DBC), including the 70A, 70B, 70C, 32D, and RPA14 domains (Figure 1C), and used SAXS to evaluate the structure of this simplified complex in the absence and presence of ssDNA of various lengths.

The SAXS data revealed that free RPA-DBC was a dynamic protein containing multiple globular domains. However, the Rg of 38.8 Å was substantially smaller than the value of 53 Å that would be expected from a fully extended architecture, although the Dmax of 171 Å suggested the presence of a small population of such highly extended species. The peak at 34 Å in the P(r) distribution could be attributed to paired electron distances within the folded domains of the protein, while shoulders at 70 Å, 100 Å, and 130 Å corresponded to distances between the trimer core and 70B, the AB linker, and 70A, respectively (Figure 4A). Low resolution ab initio structural models predicted a bilobed organization that agreed well with the SAXS data. The crystal structure of the trimer core fit well into the ab initio model, although that of the 70AB unit did not, likely because there are dynamic fluctuations of the 70A and 70B domains as well as between the 70AB unit and the trimer core. Molecular dynamics simulations generated 70,000 models that were used to calculate predicted SAXS scattering data. These predictions, which were in good agreement with the experimental SAXS results, led the researchers to conclude that RPA-DBC is dynamic, and no single conformation describes its architecture in solution. Chazin and coworkers concluded that the 70A and 70B domains and trimer core are uncoupled and move independently. Nevertheless, the protein favors organizations that are somewhat condensed, likely due to transient helix formation of the BC linker and its interactions with the 70B and 70C domains (Figure 5).

Figure 4. Diagrammatic representations of (A) ssDNA-free RPA-DBC, (B) RPA-DBC in complex with a 10 nt stretch of ssDNA, and (C) RPA-DPC in complex with a 30 nt stretch of ssDNA.


Figure 5. Diagrammatic and SAXS-derived bead models of RPA-DBC (left). Variations in architecture observed in molecular dynamics simulations (right) showing the range of flexibility of the RPA-DBC structure. Image kindly provided by Walter Chazin.

Binding of a 10 nt ssDNA resulted in a substantial compaction of RPA-DBC architecture as indicated by a 2.5 Å and 47 Å decrease in Rg and Dmax, respectively. Only one shoulder was now visible on the P(r) curve, corresponding to a 70 Å distance between the trimer core and the combined 70A and 70B domains, which had associated upon binding the ssDNA (Figure 4B). Low resolution ab initio models of this complex were now in good agreement with both the SAXS data and the X-ray crystal structures of the trimer core and 70AB. Models generated by molecular dynamics simulations were also supportive of the conclusions drawn from the SAXS data.

Association of RPA-DBC with a 20 nt ssDNA resulted in a further reduction of Rg and Dmax of 1.3 Å and 10 Å, respectively, consistent with further compaction of the protein. These results suggested that the trimer core was now interacting with the DNA, although molecular dynamics simulations starting from complexes that included ssDNA interactions with 70C alone or 70C plus 32D did not produce models in good agreement with the SAXS data. The 20 nt binding mode seems not to be a stable state, even though it was well accepted in previous reports. The investigators were surprised to find that increasing the ssDNA length up to 30 nt resulted in no additional substantial change in RPA-DBC architecture. However, both ab initio low resolution modeling and molecular dynamics simulations of the 30 nt ssDNA complex provided results in good agreement with the SAXS data. The models indicated ssDNA interactions with 70A, 70B, 70C, and 32D (Figure 4C).

Together, the results suggest that RPA-DBC has only two binding modes with ssDNA, rather than the three modes previously proposed (Figure 6). These include a complex in which short strands of ssDNA interact only with 70A and 70B and a complex in which these domains plus 70C and 32D of the trimer core are also engaged. Although the latter complex is compact, a substantial degree of flexibility is retained. The results also provide the foundation for a model that explains how RPA may coordinate ssDNA binding with its role as a scaffold for DNA modification complexes. The investigators propose that RPA-ssDNA complexes normally exist in an equilibrium between a fully engaged form, in which all domains are bound, and a partially dissociated form, in which ssDNA is not bound to the trimer core (Figure 7). This equilibrium normally favors the fully bound complex; however, binding of other proteins involved in DNA modification can cause a shift to the partial dissociation of ssDNA. This frees a portion of the ssDNA molecule for modification by the other proteins that are now bound to RPA.



Figure 6. Molecular surfaces with inserted ribbon model of RPA-DBC demonstrating its transition from the ssDNA-free protein (left) to the first (center) and the second (right) ssDNA-binding modes. The 70A, 70B, and 70C domains are light, medium, and dark blue, respectively. The 32D domain is in green, and RPA14 is in red. ssDNA is shown in gold. Image kindly provided by Walter Chazin.


Figure 7. Proposed mechanism for the coordination of ssDNA binding by RPA to its role as a scaffold for DNA modification proteins. The complex naturally exists in an equilibrium between states in which different lengths of ssDNA are or are not associated with the trimer core. Association with DNA modification proteins shifts that equilibrium to the 32D unbound state. This frees a portion of the ssDNA for interactions with the DNA modification proteins, making it available for replication, damage response, and repair or other DNA transactions.

Clearly, further work is necessary to test this novel hypothesis of RPA function. If true, this new insight provides an important framework for ongoing efforts to explore the potential of RPA as a target for cancer chemotherapy, which is currently the focus of a collaborative effort between the Chazin laboratory and VICB investigators Stephen Fesik and David Cortez.












Follow us on Twitter

Vanderbilt University School of Medicine | Vanderbilt University Medical Center | Vanderbilt University | Eskind Biomedical Library

The Vanderbilt Institute of Chemical Biology 896 Preston Building, Nashville, TN 37232-6304 866.303 VICB (8422) fax 615 936 3884
Vanderbilt University is committed to principles of equal opportunity and affirmative action. Copyright © 2013 by Vanderbilt University Medical Center