2088 | Chem. Commun., 2015, 51, 2088--2090 This journal is©The Royal Society of Chemistry 2015

Cite this:Chem. Commun., 2015,

51, 2088

Following the aggregation of human prion proteinon Au(111) surface in real-time†

Bin Wang,a Cunlan Guo,ab Zhichao Louac and Bingqian Xu*a

Aggregations of human prion protein (23–231) were monitored by

atomic force microscopy in real-time under pH 4. Prion dimers and

trimers were determined as the basic units by AFM images and simulated

structures. Aggregates aligned with the herringbone structures of an

Au(111) reconstructed surface via Au–S bonds as the first layer, while the

second layer was formed by non-covalent interactions.

Prion protein (PrP) is an intrinsically disordered protein (IDP) and isinvolved in some fatal neurodegenerative diseases of humans andanimals, such as Creutzfeldt–Jakob disease and bovine spongiformencephalopathy (mad cow disease).1 It has been widely accepted thatthe self-assembly and aggregation of IDPs will lead to amyloid fibrilsin brain tissues and kill neural cells.2–4 Therefore, it is critical tostudy the self-assembly behaviours of IDPs and their relationships todiseases. Moreover, the self-assembly of proteins on inorganic solidsis of great interest for biological and biomedical engineering.5,6

Proteins with specific chemical and physical properties such as IDPscan be used as building blocks for ‘‘bottom-up’’ fabrication offunctional nano-devices, such as biosensors and lab-on-a-chipdevices.7 These hybrid biomaterials and nanostructures havepotential applications in the diagnosis and treatment of humandiseases, such as cancers and neurodegenerative disorders. The goldsurface has been extensively used in both biosensor fabrications andfundamental studies of protein structure–function relationships,and the cysteine residue from proteins can bind to the gold surface

and form gold–sulphur bonds, which is a well-studied self-assemblymechanism.8 The PrP molecule has one internal disulfide bond, andthe two cysteine residues can bind to a gold surface and immobilizethe protein. Therefore, the self-assembly of PrP molecules onan Au(111) surface becomes a convenient approach to studythis protein.

The PrP molecules have been studied with many techniquessuch as X-ray crystallography, optical microscopy, and transmissionelectron microscopy (TEM).9,10 However, the challenge in currentPrP studies is to elucidate the molecular basis of PrP aggregation atits initial stage and at the single-molecule level, for which theaggregation mechanism is still unknown.10 The mechanistic studiesof PrP aggregation at the single-molecule level will provide criticalinformation for understanding how native PrP (PrPC) convertsto misfolded PrP (PrPSc) and how this self-propagation processeventually leads to large amyloid fibrils.10 Atomic force microscopy(AFM) has been used to show large PrP aggregates and fibers.11 TheAFM technique can also provide single-molecule resolution forstructural studies of the PrP small oligomers and monitor the initialPrP aggregation process on Au(111) or biomembrane surfaces.Therefore, we used AFM topography imaging to monitor the aggre-gation process of PrP oligomers on an Au(111) surface in real-timefor almost 4 hours. The influence of solution pH on the PrP oligomerconformations and aggregation processes on Au(111) surface werealso investigated. With the help of molecular dynamics and dockingsimulations, an overall mechanism was developed for the initialstage of PrP aggregation on the Au(111) surface. The results of thisstudy could be useful for the development of biosensor or diagnosismethods for prion proteins and prion diseases.

The AFM images showed clear parallel patterns of PrP moleculeson the Au(111) surface (Fig. 1A to D) in pH 4 buffer. The anglesamong those patterns are around 1201, which agrees with the anglevalue of the reconstructed Au(111) surface herringbone structure(Fig. 1E). Therefore, the growing directions of these patterns weredetermined by the reconstructed Au(111) surface. The human prionprotein monomer has one disulfide bond connecting Cys179 andCys214.12 The sulfur atoms from these two residues can form twoAu–S bonds with the Au(111) surface and immobilize PrP monomer

a Single Molecule Study Laboratory, Faculty of Engineering and Nanoscale Science

and Engineering Center, University of Georgia, Athens, GA 30602, USA.

E-mail: bxu@engr.uga.edu; Fax: +1-706-542-3804; Tel: +1-706-542-0502b Departments of Materials and Interfaces, Weizmann Institute of Science, POB 26,

Rehovot, 76100, Israelc College of Materials Science and Technology, Nanjing University of Aeronautics

and Astronautics, Nanjing, 210016, P. R. China

† Electronic supplementary information (ESI) available: AFM images of PrP on Au(111)surface under pH 7, molecular dynamics simulations of PrP monomer under pH 4 andpH 7, docking simulations for PrP long filaments on the second layer, combinations ofthree basic units for PrP filaments on the second layer, different combinations of PrPthree-unit aggregates on the second layer and comparison with simulations, dockingsimulations of the interface residues and hydrogen bonds formed on the interfacesbetween the first layer and the second layer in Fig. 4. See DOI: 10.1039/c4cc09209k

Received 18th November 2014,Accepted 8th December 2014

DOI: 10.1039/c4cc09209k

www.rsc.org/chemcomm

ChemComm

COMMUNICATION

Publ

ishe

d on

08

Dec

embe

r 20

14. D

ownl

oade

d by

Uni

vers

ity o

f G

eorg

ia o

n 1/

17/2

019

6:50

:03

PM.

View Article OnlineView Journal | View Issue

This journal is©The Royal Society of Chemistry 2015 Chem. Commun., 2015, 51, 2088--2090 | 2089

on the substrate. Moreover, some studies have reported experi-mental observations and theoretical hypotheses suggesting thatthe mechanism of PrP aggregation to long nanofilaments isthrough PrP oligomers (dimers or trimmers).13,14 Therefore, thepatterns of PrP shown on reconstructed the Au(111) surface areassumed to be generated by the Au–S bonds at the interfacebetween PrP dimers (or trimers) and the Au surface.

Next, molecular dynamics and docking simulations have beenused to predict the conformations of PrP dimer (size around 5.4 nm,Fig. 2A) and trimer (size around 7.0 nm, Fig. 2B) under pH 4.15,16 ThePrP monomers use their b-sheet structures to form a hydrophobiccore, and the outer surfaces are formed by more hydrophilica-helices. The disulfide bond locates between two a-helices in eachmonomer. Therefore, the dimers and trimers can be stabilized onthe Au(111) surface with their outer surfaces bonded to the Ausubstrate and the hydrophobic cores connecting with each other.

The analysis of enlarged AFM images of PrP patternsrevealed that the basic units are a mixture of PrP dimers andtrimers. The sizes of single units (either dimers or trimers) aresimilar to the simulated values, as shown along the x-axis inFig. 2D. As the dimers and trimers attach to each other, themeasured sizes should be close to the actual values. In addition,the measured cross-sectional height values of the trimers anddimers indicate the relative heights of these two units, whichshow a clear difference. Therefore, both the simulated resultsand the images prove that dimers and trimers have differentsizes and can be easily distinguished. The enlarged 3D AFMimages showed these basic units, and their conformations andorientations are illustrated with the simulated dimer and trimerstructures in Fig. 2F. More representative dimer and trimerimages and simulations are shown in Fig. S1 and S2 (ESI†).The combination of AFM high-resolution images and molecularsimulations successfully distinguished PrP dimers and trimersin the first layer of PrP on the Au(111) surface.

The aggregation of PrP was also influenced by the solutionpH value. Fig. S3 (ESI†) shows the AFM topographic images ofPrP aggregates on the Au(111) surface under pH 7. No regularpattern was observed after 3 hours of continuous scanning.According to molecular dynamics simulations (Fig. S4, ESI†), thePrP monomer cannot have any stable b-sheet structure in 60 ns ofsimulation time period under pH 7. Therefore, PrP under pH 7 ismore flexible, which reduces the probability of the exposure of thedisulfide bond to the Au(111) surface. The three a-helices of PrP aremore hydrophilic, but the gold surface is more hydrophobic, whichmakes it difficult for PrP to form stable patterns.

In some areas of the Au surface, the AFM topographicimages showed the second layer of PrP above the first layer(Fig. 3A, topographic image and the cross-sectional profilemeasurement). The interactions happening between PrP oligomersfrom the two layers are not Au–S bonds, but non-covalent inter-actions such as hydrogen bonding and van der Waals forces. Basedon the AFM images obtained from continuous scanning, the PrPdimers or trimers are assumed to quickly form in solution and theninteract with the Au(111) surface to form the first layer with the helpof Au–S bonds and the reconstructed Au(111) herringbone geometry.Consequently, the second layer will start to form with weaker non-covalent interactions between the PrP in the first layer and thedimer–trimer in solution, and this process should be much slowerthan the one for the first layer.

The conformations of PrP patterns in the second layer arevery different from the ones in the first layer. In Fig. 3B, foursmall areas were selected to demonstrate the details of thefilament-type PrP patterns. Some conformations in the second layerwere compared with simulated aggregations of dimers and trimers,as shown in Fig. S5 and S6 (ESI†). These various aggregates havedifferent numbers of binding residues and hydrogen-bonds, whichprovide different degrees of stabilization (Tables S1 to S9, ESI†).Based on the experimental AFM images and structure simulations,

Fig. 1 Representative AFM topography images of PrP on an Au(111) surface.(A) to (D) are the topographic images obtained after 21 min, 47 min, 172 min,and 235 min of continuous scanning, respectively. Image size is 700 nm by700 nm. The red frame in each image includes the area size of 400 nm by400 nm. The highlighted areas are used in later analysis. (E) STM image ofreconstructed Au(111) surface. (F) AFM image of PrP patterns enlarged fromthe area inside the white frame in (A). The three angles marked for theherringbone structures in (E) and those for the PrP patterns in (F) all have thevalue of 1201. (E) and (F) have the same image size of 200 nm by 200 nm.

Fig. 2 PrP patterns formed by the mixture of PrP dimers and trimers.(A) PrP dimer conformation predicted by docking simulation. (B) PrP trimerconformation predicted by docking simulation. The height values labelledin (A) and (B) are in the unit of Å. (C) The area highlighted by the red framein Fig. 1A. Image size 400 nm by 400 nm. The cross-section profile of thewhite line is shown in (D). (E) Is the enlarged area from the red frame in (C),with the white dashed lines highlighting the dimers and trimers. Trimersare numbered as T1, T2, and T3, dimers D1 and D2. (F) Is the 3Dtopographic image of the same area in (E). Image size of (E) and (F) is40 nm � 40 nm. (G) and (H) are the enlarged 3D images for T1 and D1,respectively. The three monomers in T1 are numbered as M1, M2, and M3.M3 is located below M1 and M2. The two monomers in D1 are numberedas M1 and M2, and M2 is below M1.

Communication ChemComm

Publ

ishe

d on

08

Dec

embe

r 20

14. D

ownl

oade

d by

Uni

vers

ity o

f G

eorg

ia o

n 1/

17/2

019

6:50

:03

PM.

View Article Online

2090 | Chem. Commun., 2015, 51, 2088--2090 This journal is©The Royal Society of Chemistry 2015

the overall mechanism of PrP aggregation on the Au(111) surface isdivided into two steps (Fig. 4E). The first step is the immobilizationof PrP patterns on the reconstructed Au(111) surface with the helpof Au–S bonds and Au geometry patterns, and the second step isthe deposition of the second layer with various conformations ofdimer–trimer aggregates. The PrP monomers (from the dimer ortrimer unit) between the first and second layers are predicted tointeract with each other via the residues in their a-helices (Fig. 4A toD and Tables S10 to S14, ESI†).

For the non-covalent interactions involved in the aggrega-tion, hydrogen bonds are the strongest type, and hydrophobic

interactions also play an important role. At each interface, theratio of hydrogen bonds to total binding residues can be usedas the criteria to compare the stability of that interface (Table S1,ESI†). The aggregations between the first and second layers canbe analysed in the same way (Table S10, ESI†). The resultsshowed that the interfaces formed by basic units (dimers ortrimers) between the two layers can be more stable than the onesformed in the same layer (Tables S1 and S10, ESI†). Therefore,the PrP dimer and trimer in solution were supposed to bestabilized when they attached to the ones in the first layer andstarted the process of growing the second layer.

The aggregation process of human recombinant PrP (23–231)on a reconstructed Au(111) surface was monitored by AFMcontinuous imaging in real-time for more than 3 hours underpH 4 and 7. Two aggregation layers were formed by PrP dimersand timers as basic units. The patterns on the first layer followedthe herringbone structures of the Au(111) surface, while thesecond layer formed less regular patterns under the influences ofthe PrP in the first layer and the interactions among each basicunit in the second layer. Molecular docking and moleculardynamics simulations provide helpful structural information toelucidate the two-step aggregation mechanism.

We thank the US National Science Foundation (ECCS 1231967and CBET 1139057) for partial financial support of this research.

Notes and references1 E. Biasini, J. A. Turnbaugh, U. Unterberger and D. A. Harris, Trends

Neurosci., 2011, 35, 92–103.2 J. Hardy and D. J. Selkoe, Science, 2002, 297, 353–356.3 S. B. Prusiner, Science, 1982, 216, 136–144.4 H. J. Dyson and P. E. Wright, Nat. Rev. Mol. Cell Biol., 2005, 6,

197.5 A. Vallee, V. Humblot and C.-M. Pradier, Acc. Chem. Res., 2010, 43,

1297–1306.6 S. Z. Nergiz, J. M. Slocik, R. R. Naik and S. Singamaneni, Phys. Chem.

Chem. Phys., 2013, 15, 11629.7 A. Lakshmanan, S. Zhang and C. A. E. Hauser, Trends Biotechnol.,

2012, 30, 155–165.8 R. G. Nuzzo, B. R. Zegarski and L. H. Dubois, J. Am. Chem. Soc., 1987,

109, 733–740.9 M. I. Apostol, K. Perry and W. K. Surewicz, J. Am. Chem. Soc., 2013,

135, 10202–10205.10 R. Diaz-Espinoza and C. Soto, Nat. Struct. Mol. Biol., 2012, 19,

370–377.11 T. Kubota, Y. Hamazoe, S. Hashiguchi, D. Ishibashi, K. Akasaka,

N. Nishida, S. Katamine, S. Sakaguchi, R. Kuroki, T. Nakashima andK. Sugimura, J. Biol. Chem., 2012, 287, 14023–14039.

12 I. G. Biljan, G. Gabriele, G. Ilc, I. Zhukov, J. Plavec and L. Giuseppe,Biochem. J., 2012, 446, 243–251.

13 M. W. van der Kamp and V. Daggett, Top. Curr. Chem., 2011, 305,169–198.

14 C. Govaerts, W. Holger, S. B. Prusiner and F. E. Cohen, Proc. Natl.Acad. Sci. U. S. A., 2004, 101, 8342–8347.

15 D. A. Case, T. A. Darden, T. E. Cheatham III, C. L. Simmerling,J. Wang, R. E. Duke, R. Luo, M. Crowley, R. C. Walker, W. Zhang,K. M. Merz, B. Wang, S. Hayik, A. Roitberg, G. Seabra, I. Kolossvary,K. F. Wong, F. Paesani, J. Vanicek, J. Liu, X. Wu, S. R. Brozell,T. Steinbrecher, H. Gohlke, Q. Cai, X. Ye, J. Wang, M.-J. Hsieh,G. Cui, D. R. Roe, D. H. Mathews, M. G. Seetin, C. Sagui, V. Babin,T. Luchko, S. Gusarov, A. Kovalenko and P. A. Kollman, AMBER 11,University of California, San Francisco, 2010.

16 S. J. de Vries, A. D. J. van Dijk and A. M. J. J. Bonvin, Proteins: Struct.,Funct., Bioinf., 2006, 63, 479–489.

Fig. 3 Analysis of the second layer of PrP aggregations on the Au(111)surface. (A) Enlarged area inside the red frame in Fig. 1C, which shows boththe first and second layers. The cross-sectional profile shows the differentheight of the two layers. (B) Only the second layer by adjusting the off-setof the AFM image in the z direction. Image size of (A) and (B) is 400 nm by400 nm. The enlarged areas labelled 1 to 4 show representative patterns inthe second layers. Dimer and trimer units are marked as D and T,respectively. The three-unit aggregation in each area is compared withthe simulated structure. D and T stand for dimer and trimer, respectively.

Fig. 4 Overall schematics of PrP aggregation on an Au(111) surface. (A) to (D),the docking simulation of dimer–trimer interactions between the first layer (up)and second layer (down). (A) Dimer to dimer; (B) dimer to trimer, (C) trimer todimer, and (D) trimer to trimer. The predicted binding residues are shown instick representation for each structure. (E) The two steps of PrP aggregation onthe reconstructed Au(111) surface by schematics. The dimer–trimer hydro-phobic cores are shown in magenta, and the hydrophilic helices in blue.

ChemComm Communication

Publ

ishe

d on

08

Dec

embe

r 20

14. D

ownl

oade

d by

Uni

vers

ity o

f G

eorg

ia o

n 1/

17/2

019

6:50

:03

PM.

View Article Online