«Kinetic Investigations of Thiolate Protected Gold Nanoparticles: Protein Interactions, Electron Transfer, and Precursor Formation By Brian N. Turner ...»
In the regions represented by peptides 1-6, the following escape mutations have been reported: N262S/Y/K, I266M, andN268I (previously mentioned in Table 1). In the area represented by peptides 7-15, K272N/Q/M/R/E/N/T/D, N276F, and N277Y (previously mentioned in Table 1) have been reported. Peptide A, a representation of the full region, binds PZ stronger than any of the peptide fragments. Joining the regions together gives an estimated, albeit semi-quantitative, increase in affinity of nearly 50% (peptide A versus peptide 6, based on a single concentration of PZ to an identical mass of peptide).
The same sequence presented in the reverse direction, peptide B, binds as effectively as sequences from peptides 7-15. Comparing segments from peptides 1-6 and 7-15 to the reversed sequence peptides 1b-7b, and the full length peptide A to B, it is clear that the orientation in which the peptides are displayed to PZ has a dramatic effect on binding affinity.
In conclusion, the most important observations to be made are that the full length epitope with the linker at the C terminus (B) shows little binding activity in ELISA (A = 0.195 ±0.019), while the full length epitope with the linker at the N terminus (A) shows significant binding (0.618 ± 0.086). The difference in binding was attributed to a directionally dependent misalignment in the case of peptide B, as demonstrated in Figure 14.
Figure 14: A conceptual interpretation of PZ approaching the synthetic peptide epitope in forward (left) and reverse (right) presentations. The illustrated concept might explain why PZ has such a strong preference for peptide A over peptide B.
Fragments 2,3,5, and 6 are more active than any of the peptides in fragments 7-15, suggesting that the sequence NSELLSLINDMPITN (amino acids 254-268) represents a significant binding motif for mAb interaction with HRSV F, particularly with PZ. The approach of PZ to the protein (or peptide fragment thereof) during recognition is, therefore, hypothesized to occur in a manner where the Fab region of PZ is pointing against the sequence NSELLSLINDMPITN, which constitutes an alpha helical and then kinked part of the native protein. While it was tempting to speculate that the hypothesized approach was due to projection of binding residues further from the surface in fragments 2, 3, 5, and 6, this was not consistent with identical experiments that used the same primary epitope sequences presented in the reverse direction. As the segments representing amino acids 269-278, which bend opposite from the direction of the earlier part of the sequence in the native protein structure, it was reasonable that presenting them in the same direction eliminated binding.
Given these conclusions, it would be interesting to study peptides 8-15 in reverse.
Alternative presentation might face these sequences in the same direction as they are presented to PZ in the native protein. Studying the two alpha helical portions with the following sequences would be informative: CSGSG-NSELLSLINDM and NDQKKLMSNN-GSGSC. These peptides would allow for interrogation of either alpha helix on its own without the non-interacting loop region (PIT). Chapter III discusses the possibility of placing these sequences together on a single nanoparticle of a specific size to mimic the distance between the two helices in the native protein.
If both halves appear to be distinct “hot spots” that bind the strongest when oriented in opposite direction of each other, this would confirm the importance of the loop conformation of this region in the native protein. In the full length linear sequence, an amorphous loop conformation might be represented by the natural kink introduced by the proline residue at position 265. If this theory is true, the data stands in sharp contrast to the work of Gerdon, et. al.110 In their study of a loop peptide epitope of anthrax protective antigen, they had to use a bidentate attachment strategy in order for their antibody to bind to its nanoparticle immobilized synthetic peptide epitope, purportedly in order to force the peptide into the loop conformation of the native structure. Considering that the peptide used in their study also featured a proline residue in the middle of the sequence does not necessarily mean that it is kinked into a favorable conformation in the free peptide or in the monodentate immobilized peptide, therefore still requiring the bidentate attachment for antigenic binding. The PZ epitope peptide epitope has two alpha helices, which might help control the conformation without bidentate attachment. The conformation of peptide A is therefore thought to be more identical to the PZ epitope in the native protein than synthetic anthrax PA epitope is to the native epitope. The conformational differences could be confirmed by observing the secondary structure of both peptides in solution, bound to maleimide coated plastic with a sing terminal cysteine, and also to a gold surface and analyzing with circular dichroism or solid state NMR.
Alternatively, peptide A has a terminal asparagine residue which is capable of forming either an electrostatic or salt bridging interaction with cysteine carboxylate groups, available from the blocking agent used. Van der Waals bonding of this nature might also act to hold the peptide into a loop conformation more similar to the native structure.
Contrary to this theory is the observation that the C-terminal linked peptide B, which also presents a terminal asparagine residue, does not exhibit as much activity towards palivizumab. Notwithstanding, in this form the terminal asparagine may be held at an angle where the steric hindrance to moving the asparagine within close proximity of the blocking cysteine molecules is greater than the steric hindrance encountered in peptide A.
Another possibility is that other residues on that terminal would produce repulsive forces against the well-plate surface. Notably, the glutamic acid residue near the N terminal would electrostatically repel the like-charged carboxylate group. If this model holds true, peptide A might adopt a loop conformation on the well plate whereas peptide B does not.
Since these conclusions were drawn, crystal structures of HRSV F, post-fusion HRSV F, and the motavizumab peptide epitope (identical to the full peptide sequence used in this study) bound to HRSV F have been published. The authors determined that N262, N268, D269, K272, and S275 all have hydrogen bonding or salt bridge interactions with residues in CDRs of both the heavy and light chain of the motavizumab Fab.
Figure 15: Motavizumab peptide epitope of HRSV F (gray) interacting with the light chain (blue) and heavy chain (green). Labeled amino acids are those participating in hydrogen bonds or salt bridges. From McClellan, et al. 90 Copyright 2010, Nature Publishing Group, reprinted with permission.
Considering the importance of these residues, it is worth noting that peptide 6 represents the step where N268 is just included along with N262. Unexpectedly, binding is lost upon adding the next important binding residue, D269, in peptide 7, without losing any other important binding residues or escape mutants. At this point along the stepwise path it is possible that alpha helicity in the early part of the chain is lost with the disappearance of S259, forcing N262 out of a position suitable to bond formation.
The data collected in the peptide ELISA experiments was subsequently used to design antigenic mimics with small gold nanoparticle scaffolds. Chapter III will go on to describe the conjugation of peptides A and 6 to tiopronin protected gold nanoparticles and applying these conjugates in a QCM immunoassay. Furthermore, a loop configuration, with CSGSG linkers on both ends of the sequence, will be evaluated on two sizes of nanoparticle. Additionally, it would be useful to see if the full length peptide maintains the helix-loop-helix structure in solution with the linkers attached, as glycine residues can disrupt helicity. If it does, it would be interesting to see whether certain deletions off of either end disrupt this secondary structure.
I would like to begin by thanking Dr. David Wright, Dr. James Crowe, and Dr. David Cliffel for their advice on approaching the epitope mapping experiments on the discontinuous, winding path that it ended up taking. Similarly, I would like to thank Dr.
Sam Kuhn and Dr. Chris Keefer for their earlier assistance with this project. Most importantly, I would like to acknowledge my initial training/mentoring from Dr. Aren Gerdon, who carved out much of the early research that led to this study. I would like to acknowledge Malgorzata Broncel, Ryan Rutledge, Holly Carrell, and Joshua Swartz for their aid in peptide synthesis, characterization, and purification. I would also like to thank Catherine Prudom, Alex Rutledge, and Ryan Rutledge for their help with ELISA assay optimization. Finally, I would like to thank Jay Forsythe, Michal Kliman, Kellen Harkness, and Andrzej Balinski for their help with mass spectrometry.
Having a better understanding of the nature of peptide epitopes based on HRSV F antigenic site A, the research went on to employ these materials in combination with tiopronin protected gold nanoparticles, discussed in Chapter I. The research aimed to determine whether or not different peptide epitopes were active (bound by an antibody) on gold nanoparticles, whether the recognition was specific, the effect of simultaneously changing the size of the nanoparticle and identity of an inert spacer between the nanoparticle and peptide, and quantifying the resulting binding interaction for a working “mimic” that was produced. The significance of these results lies in the fact that if an antibody can recognize a nanoparticle, then it is possible that an immune response to the same nanoparticle could result in serum specific to both the nanoparticle and the native protein that inspired it. A nanoparticle conjugate of this design may even be capable of evoking the production of memory cells in vivo, validating it as a novel subunit nanovaccine. Additionally, as the peptide and nanoparticle research progressed, a novel immunosensor platform for the diagnosis of HRSV was developed.
As was discussed in Chapter II, some of the early results in this chapter were not informed by information from literature or research presented in Chapter II. However, the results that are most emphasized and the future directions that are suggested, presented later in this chapter, follow the findings discussed in Chapter II.
Nanoparticles Presenting Peptides as Antigenic Mimics (Note: much of this background material appears in a previously written review article by this author, Brian Huffman, and David Cliffel, however rewritten and truncated for clarity and a narrowed focus).1 Given their similar size dimensions to biological macromolecules such as enzymes, receptors, immunoglobulins, antigens, DNA, and carbohydrates, small gold nanoparticles have gathered much interest as biomimetic platforms. These nanomaterials can be surface functionalized during particle assembly with a biomolecule in situ, by direct covalent attachment, or by place-exchange with the biomolecule after particle assembly.
In many cases, these conjugate materials have long-term stability. Additionally, nanoparticles can impart conformational structure to a biomolecule, contain a number of different biomolecules within a single supramolecular structure, and are often non-toxic.
Furthermore, the methods by which these materials are made provide simpler routes to biofunctionality than classical biological methods such as genetic engineering and directed evolution. Given these fundamental ideas and motivations, the research took from the work of interdisciplinary scientists and proceeded to adapt their concepts to the problem of HRSV, discussed in the previous chapter. It is thought that, ultimately, there are a number of situations in which nanotechnology will prove more effective than the currently used clinical technology.
A significant body of work exists describing the application of nanomaterials to biological problems. Here, the focus is on a subset of this work that concentrated on using peptides and small metal nanoparticles to mimic the antigenic behavior of whole proteins. An early and straightforward example of this is the synthesis of nanoparticles with a protecting peptide from the histidine-rich protein II (HRP-II) of Plasmodium falciparum.111 Using standard fmoc procedures, Wright and co-workers recreated this peptide and used it as a stabilizing ligand on different metal core particles: ZnS, Au 0, Ag0, TiO2, and AgS. These particles were recognized with good specificity by a monoclonal antibody specific for Plasmodium falciparum. Their group was able to detect the peptide-encapsulated particles as they would the whole protein with a colorimetric antigen capture assay, showing that their particle mimics the protein-protein interaction of the intact native structure.
Recently, our group has developed several gold nanoparticles that mimic antigens of clinical interest. The first was a glutathione (GSH) passivated gold cluster (GS-MPC) that was then detected with a polyclonal anti-GSH antibody.105 The antibody very specifically recognized the GSH-MPC compared to a tiopronin protected nanoparticle, even though both surface ligands only differ by, effectively, one amino acid, as seen in Figure 16. While glutathione is not a traditional antigen, it serves as a proof of concept that a very small ( 3 nm) MPC can be functionalized with a surface peptide, and then specifically recognized by a corresponding antibody.
Figure 16: Comparison of the glutathione (top) and tiopronin (bottom) ligands used to functionalize MPCs. Tiopronin is a truncated form of the tripeptide glutathione, with overlap shown in red.105 Copyright of the American Chemical Society.