University of Groningen
Novel peptide replicators from dynamic combinatorial libraries
Altay, Yigit
DOI:
10.33612/diss.90041906
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2019
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Altay, Y. (2019). Novel peptide replicators from dynamic combinatorial libraries. University of Groningen.
https://doi.org/10.33612/diss.90041906
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
24 1. Introduction [69] Belousov, B. P. Collection of
ab-stracts on radiation medicine 1959, 147, 145.
[70] Zhabotinsky, A. M. Biophysics 1963, 9, 306-311.
[71] Debnath, S.; Roy, S.; Ulijn, R. V. J. Am. Chem. Soc. 2013, 135, 16789-16792.
[72] Boekhoven, J.; Brizard, A.;
Kowlgi, K.; Koper, G.;
Eelkema, Rienk andvan Esch, J. Angew. Chem. Int. Ed. 49, 4825-4828.
[73] Ilday, S.; Makey, G.; Akguc, G. B.; Yavuz, ¨O.; Tokel, O.; Pavlov, I.; G¨ulseren, O.; Ilday, F. ¨O. Nature Comm. 2017, 8, 14942.
[74] Bull, J. J.; Meyers, L. A.; Lach-mann, M. PLOS Comp. Biol. 2005, 1, 450-460.
[75] Nowak, M. A. Trends Ecol. Evol. 1992, 7, 118-121.
[76] Sadownik, J. W.; Mattia, E.; Nowak, P.; Otto, S. Nat. Chem. 2016, 8, 264 - 269.
Manuscript in preparation.
Chapter 2
A Structural Survey in the Search for
Novel Self-Replicating Peptides
“The hardest thing to explain is the glaringly evident which everybody has decided not to see.”
Ayn Rand
Abstract
One of the most puzzling and intriguing questions regarding the origin of life is not actually about solely finding the source of life but creating life de novo using the principles that we learnt on the way. Peptides that can self-replicate provide us with an important aspect needed to solve this puzzle by dealing with one of the three fundamental elements of life: replication. As we are unraveling the mechanisms of replication at the molecular level, we realized the importance of structural variation along with environmental effects. Here we present a struc-tural survey of self-replicating peptides in order to extend our portfolio with novel replicating peptide sequences and additionally decipher the factors that are playing a role in the replication behavior. For that purpose we designed and synthesized 16 different peptide sequences with different hydrophilicities. We also investigated the effect of the carboxy terminus on the self-assembly and the effect of ionic strength on the emergence of replicators. We have found that up to a certain hydrophilicity, the overall peptide hydrophilicity is the dominant factor on the selection of replicating macrocycles, and beyond that, several other factors start to take over. Additionally, we showed the importance of the car-boxy terminus on the self-assembly of peptides using experimental data along with molecular dynamic simulations. Finally, we discovered that dynamic com-binatorial libraries set up in high ionic strength buffers form relatively short fibers which affects the overall rate of the self-replication.
26 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
2.1
Introduction
Q
uestions regarding the molecular origin(s) of life have the potential of provid-ing answers to what life is and how it evolved. The RNA world hypothesis1,2 could provide an explanation for the evolution of biochemical networks but it does not explain the connection between the primeval molecules and the emergence of life. Molecules with high complexity are unlikely to emerge spontaneously. In contrast to RNA and other nucleic acids, amino acids and short peptides can be easily produced under primordial conditions.3Although natively folded proteins tend to have limited stability, the amyloid fold has excellent properties in terms of resistance to degra-dation and self-propagation via a nucleation-growth mechanism.4–10 Formation of tubular,11–14 vesicular15–18and fibrillar19–21structures via self-assembly of peptides as simple as dipeptides has been demonstrated extensively in the literature. In addi-tion to that, short peptides can be used as templates22and can act as catalysts23,24 for the synthesis of other peptides. In contrast to the RNA world hypothesis which features complex RNA molecules and nucleotides, the factors above favor the use of functional short peptides in the search for satisfactory answers that can help to understand life and possibly create life de novo.Peptide replicators23,25–34that have been developed using systems chemistry35–37 in combination with dynamic combinatorial chemistry38–43 constitute an important advance in connecting chemistry and biology on this important topic. Short peptide amphiphiles that have a chain in which hydrophobic and hydrophilic amino acids alternate, have a high propensity to assemble into β-sheets in their amyloid-like fibril structures.44–49 The peptide sequence50–54 and the size of non-polar residues55,56 have a significant influence on the self-assembly.
In our lab, we have studied such systems with different building blocks produc-ing replicators havproduc-ing different sizes. In a previous study,57 Malakoutikhah et al. revealed the selection criteria for the self-assembling macrocycles - i.e. as the hy-drophobicity of the building blocks decrease, the equilibrium shifts toward larger macrocycles, since more peptide strands are needed to drive self-assembly. It has been shown that minor structural changes in the replicator design can heavily in-fluence the library distribution and the selection of the most stable product in a dynamic combinatorial library (DCL). The most hydrophilic building blocks used in the previous study contained alanine (Ala) and serine (Ser) residues. The effect of further increase of the hydrophilicity of the peptide is unknown and is investigated in this chapter.
2.1. Introduction 27 Table 2.1: Chemical structures of the peptides used in this study. Each peptide is num-bered depending on the amino acid at position R3. Peptides having an amide group at R1
are represented with a suffix “a”.
R1= -COO- R2= R3=
Lys p-Ser p-Tyr Asn Ser Thr
1 2 3 4 5
Gly Ala Tyr Phe
6 7 8 9
R1= -CONH2 R2= R3=
Lys p-Ser Asn Ser Thr
1a 3a 4a 5a Gly Ala 6a 7a R1= -COO- R2= R3= Asp Ala 7b
In this study, we have investigated the effects of structural variation of the peptide sequence on the library composition in more detail. For that purpose, we have designed, synthesized and purified peptides listed in Table 2.1 and studied the DCLs
28 2. A Structural Survey in the Search for Novel Self-Replicating Peptides formed from them in detail. We have found that even minute changes in the peptide sequence can have dramatic and unanticipated effects on the product distribution of DCLs as well as on the the self-assembly of each peptide. The structural variations that are presented in this chapter include the variation of amino acids at specific positions, amide functionalisation of the C-termini and modification of the charged amino acids to change the overall charge state of the peptide. Additionally, the effect of the ionic strength is studied to unravel the factors that might play an important role in the emergence of replicators from the dynamic combinatorial libraries. These findings provide new insight into how the self-assembly process of a peptide replicator is affected by minute structural and environmental changes.
2.2
Results and Discussion
The chain of alternating hydrophilic and hydrophobic amino acids in the 5-amino-acid-long peptide sequence makes the building blocks prone to assemble into β-sheets.59–63 Our peptide based building blocks that are used to make dynamic com-binatorial libraries bear an additional dithiol unit, derived from 3,5-dithio-benzoic acid (DTB), at the N-termini to enable the reversible disulfide exchange reactions.28
The oxygen present in air oxidizes the thiols to disulfides (Figure 2.1). In slightly basic (pH 8.2) aqueous solutions, a mixture of differently sized macrocycles is formed that continuously exchanges building blocks. The composition of the DCL can be perturbed by a macrocycle that is able to stabilize itself through self-assembly. Such assembly tends to occur through a nucleation-growth mechanism.28 In such a case,
the composition shifts towards this specific macrocycle and in the presence of me-chanical agitation (shaking or stirring) exponential growth can be achieved, since the number of fiber ends, from which the fiber grows, increases as the growing fiber fragments by the mechanical energy provided.
A challenge in these systems is to obtain building blocks at high purity (>99%). Especially as the hydrophilicity of the peptides increase, synthesis of the building block with high purity and yield becomes more difficult, particularly if there is a phosphorylated amino acid present in the sequence. Yet, high purity is required as the libraries tend to be effected by small perturbations.
2.2.1
Sequence Variation
We have selected a certain position (R3in Table 2.1), that corresponds to the second
amino acid from the C-termini, to survey the effect of hydrophilicity of the overall peptide on the possible formation of replicators. We used phospho-serine (p-Ser) as the most and tyrosine (Tyr) as the least hydrophilic residues in this series.
2.2. Results and Discussion 29
Figure 2.1: Schematic representation of a) oxidation of the peptide building block, and the exponential replication of hexamer macrocycles obtained from a dynamic combinatorial library, b) formation of the disulfide bonds and chemical structures of oligomers in the mixture, c) representative TEM image (adapted from a previously published study)58of the
bundles of hexamer replicators formed.
Peptide 1 that contains the most hydrophilic p-Ser residue could potentially form a replicator macrocycle larger than an octamer considering the previously postu-lated selection rules, which state that replicator ring size increases with peptide hy-drophilicity. Kinetic data (Figure 2.3a) shows that 14 grows rapidly as the monomer
oxidizes. However, any other macrocycles larger than tetramer were not detected under the experimental conditions (50 mM borate buffer, pH 8.2, agitated library). Even though the library members appear to self-assemble into fibers, self-replication behavior could not be assessed with the standard seeding experiments as 14emerges
almost instantaneously after the library is set up. The exponential growth of the 14
28 2. A Structural Survey in the Search for Novel Self-Replicating Peptides formed from them in detail. We have found that even minute changes in the peptide sequence can have dramatic and unanticipated effects on the product distribution of DCLs as well as on the the self-assembly of each peptide. The structural variations that are presented in this chapter include the variation of amino acids at specific positions, amide functionalisation of the C-termini and modification of the charged amino acids to change the overall charge state of the peptide. Additionally, the effect of the ionic strength is studied to unravel the factors that might play an important role in the emergence of replicators from the dynamic combinatorial libraries. These findings provide new insight into how the self-assembly process of a peptide replicator is affected by minute structural and environmental changes.
2.2
Results and Discussion
The chain of alternating hydrophilic and hydrophobic amino acids in the 5-amino-acid-long peptide sequence makes the building blocks prone to assemble into β-sheets.59–63Our peptide based building blocks that are used to make dynamic com-binatorial libraries bear an additional dithiol unit, derived from 3,5-dithio-benzoic acid (DTB), at the N-termini to enable the reversible disulfide exchange reactions.28
The oxygen present in air oxidizes the thiols to disulfides (Figure 2.1). In slightly basic (pH 8.2) aqueous solutions, a mixture of differently sized macrocycles is formed that continuously exchanges building blocks. The composition of the DCL can be perturbed by a macrocycle that is able to stabilize itself through self-assembly. Such assembly tends to occur through a nucleation-growth mechanism.28 In such a case,
the composition shifts towards this specific macrocycle and in the presence of me-chanical agitation (shaking or stirring) exponential growth can be achieved, since the number of fiber ends, from which the fiber grows, increases as the growing fiber fragments by the mechanical energy provided.
A challenge in these systems is to obtain building blocks at high purity (>99%). Especially as the hydrophilicity of the peptides increase, synthesis of the building block with high purity and yield becomes more difficult, particularly if there is a phosphorylated amino acid present in the sequence. Yet, high purity is required as the libraries tend to be effected by small perturbations.
2.2.1
Sequence Variation
We have selected a certain position (R3in Table 2.1), that corresponds to the second
amino acid from the C-termini, to survey the effect of hydrophilicity of the overall peptide on the possible formation of replicators. We used phospho-serine (p-Ser) as the most and tyrosine (Tyr) as the least hydrophilic residues in this series.
2.2. Results and Discussion 29
Figure 2.1: Schematic representation of a) oxidation of the peptide building block, and the exponential replication of hexamer macrocycles obtained from a dynamic combinatorial library, b) formation of the disulfide bonds and chemical structures of oligomers in the mixture, c) representative TEM image (adapted from a previously published study)58of the
bundles of hexamer replicators formed.
Peptide 1 that contains the most hydrophilic p-Ser residue could potentially form a replicator macrocycle larger than an octamer considering the previously postu-lated selection rules, which state that replicator ring size increases with peptide hy-drophilicity. Kinetic data (Figure 2.3a) shows that 14grows rapidly as the monomer
oxidizes. However, any other macrocycles larger than tetramer were not detected under the experimental conditions (50 mM borate buffer, pH 8.2, agitated library). Even though the library members appear to self-assemble into fibers, self-replication behavior could not be assessed with the standard seeding experiments as 14emerges
almost instantaneously after the library is set up. The exponential growth of the 14
30 2. A Structural Survey in the Search for Novel Self-Replicating Peptides TEM images (Figure 2.119) showed highly laterally associated long fibers which can support our hypothesis.
Peptide 2 forms mostly trimers with tetramers in small amounts in the DCLs prepared following standard protocols (Figure 2.3c). TEM analysis revealed fibers having a width of 7.8 `´ 1.1 nm for 23 (see Figure 2.120). Kinetic data together with
the TEM analysis suggests that 23is able to self-replicate. However, verifying this by
a seeding experiment is complicated by the fact that the trimer emerges immediately upon oxidation of the building block (i.e. there is no lag phase). We used two different batches of this peptide in this study: one was synthesized by conventional methods and has a purity of 95% and the other was purchased and had a purity of 99%. When the library is fully oxidized by perborate just after it is set up formation of oligomers up to pentadecamer is observed only when the high purity building block is used (Figure 2.2). Purity of the building block (as long as it is above 90%) does not affect the outcome of the DCLs that gave rise to the most extensively studied replicators formed from peptides 4, 7 and 9. TEM analysis did not show any type of fiber formation or aggregation. Seeding experiments also failed to produce large oligomers in high percentages (Figure 2.66).
Figure 2.2: a) Kinetic profile of a library made from peptide 2 oxidized 100% (with respect to monomer) with 80 mM perborate solution. The library was prepared to have a 3.8 mM concentration in building block in 50 mM borate buffer, pH 8.2 and was stirred at 1200 rpm.
2.2. Results and Discussion 31
Figure 2.3: Kinetic profiles of libraries made from peptide a) 1, b) 1a, c) 2, d) 3, e) 3a, f) 4, g) 4a, h) 5, i) 5a, j) 6, k) 6a, l) 7, m) 7a, n) 7b, o) 8. All libraries were prepared to have a 3.8 mM concentration in building block in 50 mM borate buffer, pH 8.2 and were stirred at 1200 rpm. x1 denotes an unidentified peak.
30 2. A Structural Survey in the Search for Novel Self-Replicating Peptides TEM images (Figure 2.119) showed highly laterally associated long fibers which can support our hypothesis.
Peptide 2 forms mostly trimers with tetramers in small amounts in the DCLs prepared following standard protocols (Figure 2.3c). TEM analysis revealed fibers having a width of 7.8 `´ 1.1 nm for 23 (see Figure 2.120). Kinetic data together with
the TEM analysis suggests that 23is able to self-replicate. However, verifying this by
a seeding experiment is complicated by the fact that the trimer emerges immediately upon oxidation of the building block (i.e. there is no lag phase). We used two different batches of this peptide in this study: one was synthesized by conventional methods and has a purity of 95% and the other was purchased and had a purity of 99%. When the library is fully oxidized by perborate just after it is set up formation of oligomers up to pentadecamer is observed only when the high purity building block is used (Figure 2.2). Purity of the building block (as long as it is above 90%) does not affect the outcome of the DCLs that gave rise to the most extensively studied replicators formed from peptides 4, 7 and 9. TEM analysis did not show any type of fiber formation or aggregation. Seeding experiments also failed to produce large oligomers in high percentages (Figure 2.66).
Figure 2.2: a) Kinetic profile of a library made from peptide 2 oxidized 100% (with respect to monomer) with 80 mM perborate solution. The library was prepared to have a 3.8 mM concentration in building block in 50 mM borate buffer, pH 8.2 and was stirred at 1200 rpm.
2.2. Results and Discussion 31
Figure 2.3: Kinetic profiles of libraries made from peptide a) 1, b) 1a, c) 2, d) 3, e) 3a, f) 4, g) 4a, h) 5, i) 5a, j) 6, k) 6a, l) 7, m) 7a, n) 7b, o) 8. All libraries were prepared to have a 3.8 mM concentration in building block in 50 mM borate buffer, pH 8.2 and were stirred at 1200 rpm. x1 denotes an unidentified peak.
32 2. A Structural Survey in the Search for Novel Self-Replicating Peptides Peptide 8, which corresponds to the dephosphorylation product of 2, forms 83 dominated libraries in DCLs prepared in borate buffer (Figure 2.3o and 2.4). Similar to the trimers of peptide 2, trimers of peptide 8 also showed self-assembled structures in TEM (8.7 `´ 0.9 nm width fibers in Figure 2.125 and once bundled fibers with 18.4 `´ 2.5 nm width in Figure 2.126). However, again, seeding experiments were not conducted as the trimers emerge immediately upon oxidation of the building block. Unexpectedly, we observed the emergence of 88when the library is rapidly oxidized to 90% with perborate followed by slow oxidation by air (Figure 2.113). Both of these peptides, 2 and 8, bear a residue having an aromatic ring, p-Tyr or Tyr. They are capable of forming smaller macrocycle stacks (trimers), presumably due to the hydrophobic interactions between these aromatic moieties in their side chain. We studied the behavior of DCLs made from peptide 8 and their propensity to give rise to replicators further by cross-catalysis as discussed in detail in Chapter 4.
Figure 2.4: Kinetic profiles of libraries made from peptide 8 a) oxidized by air (taken from Figure 2.3) and b) oxidized 80% with perborate solution (80 mM) followed by slow air oxidation. c) Comparison of the growth of 88 in seeded and non-seeded libraries. The seed
was prepared by fast oxidation using perborate. Seeding was done (using 10% seed with respect to monomer concentration) just after the library was set up. Libraries were prepared to have 3.8 mM concentration in building block in 50 mM borate buffer, pH 8.2 and were stirred at 1200 rpm.
We have introduced an asparagine (Asn) residue in peptide 3. Kinetic data in Figure 2.5a shows the emergence of 38 shortly after the library was prepared. The sigmoidal shape of the growth curve for 38readily is indicative of replication behavior. Using the 38formed in this library as a seed, we set up another identical library and added 10 mol % (with respect to monomers) of this seed just after the experiment started. In the kinetic plot shown in Figure 2.5b, the loss of the lag phase in the growth curve when the library is seeded with octamer confirms that peptide 3 is a self-replicator. Additionally, TEM micrographs show highly laterally associated
2.2. Results and Discussion 33 fibers of 38 (Figure 2.122). We presume that the Asn moiety, which can act as an H-donor and -acceptor, is playing a key role in the inter-fibrillar interactions that are pronounced in these assemblies.
Figure 2.5: Kinetic profiles of libraries made from peptide a) 3 (taken from Figure 2.3d) and b) comparison of the growth of 38in seeded and non-seeded libraries. Libraries were 3.8
mM in building block in 50 mM borate buffer, pH 8.2 and were stirred at 1200 rpm. Seeding was done on day 2.
Peptide 4 and peptide 7 were extensively studied by Malakoutikhah et al.57 and they were reported to produce octamer replicators in agitated DCLs set up in borate buffer. Peptide 5 and peptide 6 were designed to have hydrophilicities in between those of peptides 4 and 7. However, both of these peptides failed to form octamers or any macrocycle larger than tetramers. These results were counterintuitive, con-sidering the proposed rules for macrocycle size selection; and suggest that peptide hydrophilicity is not the only factor affecting the size of the replicating macrocycle. While we were searching for other factors that might dictate replicator ring size, we found a cross-catalytic pathway that induced formation of a hexamer replicator of peptide 5 which is discussed in detail in Chapter 3.
Emergence of the octamer replicators in the DCLs made from peptide 3 and 8 is in accordance with the results published before.57There is a progression from trimer replicators to octamer replicators by increasing the hydrophilicity of the second amino acid from the C-termini. However, we did not observe a further increase in ring size as we increased the hydrophilicity further. Dr. Frederix ran molecular dynamic (MD) simulations to understand whether the failure in forming rings larger than octamers is due to a conformational/steric problem or due to a possible inability to shield hydrophobic surface area from contact with water.
For the MD simulations we decided to use a peptide that is known to form self-replicating assemblies and that is widely studied by our group. We chose peptide 9 that is forming macrocycles having a size up to hexamers (with hexamers being the replicator). MD simulations were performed on macrocycles made from peptide 9
34 2. A Structural Survey in the Search for Novel Self-Replicating Peptides having 3, 4, 6, 8, 10 and 14 building blocks. We included macrocycles larger than hexamer in order to see the limits of our system. We have found that the core parts of the macrocycles, which consist of disulfide bridges, have different conformations and different preferences for interaction which likely affect the capability of forming self-assembling structures (Table 2.5). Firstly, trimers cannot form intramolecular stacks of 3,5-dithio-benzoic acid (DTB) group. In most of the simulations, all peptides are on the same side of the macrocycle. The phenylalanine (Phe) side chains fold back on to the core to π-stack. For tetramers, the DTB 1-3 interaction is preferred (see Figure 2.6 for the schematic representation of these interactions). In this conformation the formation of intramolecular H-bonds is hampered. Pentamers appear more dynamic, reversibly forming different 1-3 interactions of DTB. This is the first macrocycle in the series where intramolecular H-bonds and salt-bridges are observed. For hexamers; 1-3 and 1-3-5 stacking of DTB is commonly observed. Beyond this ring size the core becomes more elliptical than circular. For heptamers and octamers, stacking is ill-defined and the occurance of intramolecular H-bonds is infrequent. For the decamers, 1-3; 1-4; 1-5 DTB stacking is observed. Although we do not see many intramolecular H-bonds, there are a considerable number of salt-bridges present. Additionally DTBs are no longer significantly exposed to solvent. For tetradecamers, 1-3; 1-4; 1-5 and 1-3-5 stacking of DTBs is observed. In contrast to decamers, intramolecular H-bonds are present as well as salt-bridges. Similar to decamers, DTBs are not significantly exposed to solvent.
Figure 2.6: A schematic representation of interactions observed in the DTB core of the hexamer of peptide 9. a) 1-3, b) 1-4, c)1-5 and d) 1-3-5 interaction.
In order to compare the difference in shielding of hydrophobic surface area for the different ring sizes, we calculated solvent accessible surface area (SASA) for the core DTB, the core DTB with phenylalanine attached to it (DTB+PHE), and the core
2.2. Results and Discussion 35 DTB with phenylalanine and leucine attached to it (DTB+PHE+LEU). Additionally we calculated the hydrophilic and hydrophobic components of the total SASA. Figure 2.7a shows that the total SASA per monomer is decreasing with macrocycle size, as expected. Approximating the macrocycle as a sphere (which is reasonable for large ring sizes), we found that the volume is linear with the number of monomers (Figure 2.7d), while the surface of the sphere is proportional to (#monomers)2/3. Also, there does not seem to be a large change of ratio of hydrophobic-to-hydrophilic SASA, which could have been an indicator of differences in prosperities to assemble for different macrocycles. With small macrocycles, the DTB core is relatively exposed to the solvent. With hexamer and octamers, specific intramolecular stacking of DTB occurs, somewhat limiting the SASA. With larger macrocycles, all the surface area is covered by the peptide chains around it. Given that the DTB groups of different macrocycles have to find each other for them to stack, this shielding of the DTB groups by the peptide chains may be an important factor hampering stacking by larger macrocycles.
Furthermore, there appears to be a limit of space around the DTB core in a fiber scenario where the core is flat. In figure 2.10b, one can imagine fitting two more β-sheets around the slightly larger core, but the lateral sheet-sheet interaction sterically prevents fibers from larger macrocycles with a flat DTB core. In other words, the sheets take up more radial space than the DTB allows in the case of large cycles.
There is more than one variable that determines the ring size of the replicator macrocycle. Hydrophilicity of the building block is an important parameter, but also other variables play a role. These variables include the purity of the building block, the environment and the oxidation rate, as shown above. As we expected that also the charge of the peptide could play a role we decided to explore the effect of this parameter.
The overall charge of the peptides described thus far in this chapter is around +1 at pH 8.2 which is used in the entire thesis. We decided to introduce a modification changes the charge of the peptide: in peptide 7b, two lysine (Lys) residues present in the original design are replaced by aspartic acids (Asp) to change the overall charge of the peptide from +1 to -3 (assuming that the charged residues do not affect each others pKAs. We were expecting an increase in the macrocycle size due to the fact that charge repulsion would weaken the peptide-peptide interactions which would provide a bias towards the formation of larger macrocycles. However, peptide 7b almost exclusively formed trimers and tetramers in the DCLs prepared in borate buffer (Figure 2.3n). Although we did not detect larger macrocycles, we observed self-assembled fibers (with a width of 6.3 `´ 1.1 nm) by TEM (Figure 2.124). As seeding experiments failed to enhance the rate of formation of trimers and tetramers, there is no evidence for self-replication.
36 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
Figure 2.7: Simulations of single macrocycles containing 3, 4, 5, 6, 7, 8, 10 and 14 building blocks in solution for 105 ns. Visual inspection and calculation of the sol-vent accessible surface area (SASA) were performed. The SASA was calculated in var-ious ways: a) the total SASA and the SASA divided into hydrophobic and hydrophilic parts by default GROMACS tools; and b) the specific SASA of DTB, DTB+PHE side chain (sc) and DTB+PHEsc+LEUsc; c) percentage of SASA of DTB, DTB+PHEsc and DTB+PHEsc+LEUsc in the total SASA; d) relationship between the macrocycle size and the volume through the (SASA)3/2. Numbers represent an average over the last 15 ns of the simulation, with a measurement error of `´ 0.1 nm2. All simulations were performed five times over single cycle simulations.
Circular dichroism (CD) is a widely used spectroscopic technique to identify the secondary structures of proteins and peptides.64,65Especially the spectroscopic data
at the far UV region (180-250 nm) is informative since the electronic transitions in the amide group falls into this region. When a peptide backbone adopts a β-sheet,
α-2.2. Results and Discussion 37
Figure 2.8: CD spectra of stirred DCLs made from peptide a) 1, b) 1a, c) 2, d) 3, e) 3a, f) 4, g) 4a, h) 5, i) 5a, j) 6, k) 6a, l) 7, m) 7a, n) 7b, o) 8 corresponding to Figure 2.3.
helix or random coil structure, each having specific Φ, Ψ angles and hydrogen bonding patterns, these give rise to characteristic CD spectra. So we used CD spectroscopy to analyze the mode of assembly of our macrocycles. CD spectra (Figure 2.8) of the DCLs made from peptide 2, 3, 4, 7, and 8 (corresponding to Figure 2.3), showed positive helicities below 200 nm and negative helicities between 205-220 nm which indicates the formation of β-sheets.66–70 Peptide 7b showed a mixed structure of
36 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
Figure 2.7: Simulations of single macrocycles containing 3, 4, 5, 6, 7, 8, 10 and 14 building blocks in solution for 105 ns. Visual inspection and calculation of the sol-vent accessible surface area (SASA) were performed. The SASA was calculated in var-ious ways: a) the total SASA and the SASA divided into hydrophobic and hydrophilic parts by default GROMACS tools; and b) the specific SASA of DTB, DTB+PHE side chain (sc) and DTB+PHEsc+LEUsc; c) percentage of SASA of DTB, DTB+PHEsc and DTB+PHEsc+LEUsc in the total SASA; d) relationship between the macrocycle size and the volume through the (SASA)3/2. Numbers represent an average over the last 15 ns of the simulation, with a measurement error of `´ 0.1 nm2. All simulations were performed five times over single cycle simulations.
Circular dichroism (CD) is a widely used spectroscopic technique to identify the secondary structures of proteins and peptides.64,65Especially the spectroscopic data
at the far UV region (180-250 nm) is informative since the electronic transitions in the amide group falls into this region. When a peptide backbone adopts a β-sheet,
α-2.2. Results and Discussion 37
Figure 2.8: CD spectra of stirred DCLs made from peptide a) 1, b) 1a, c) 2, d) 3, e) 3a, f) 4, g) 4a, h) 5, i) 5a, j) 6, k) 6a, l) 7, m) 7a, n) 7b, o) 8 corresponding to Figure 2.3.
helix or random coil structure, each having specific Φ, Ψ angles and hydrogen bonding patterns, these give rise to characteristic CD spectra. So we used CD spectroscopy to analyze the mode of assembly of our macrocycles. CD spectra (Figure 2.8) of the DCLs made from peptide 2, 3, 4, 7, and 8 (corresponding to Figure 2.3), showed positive helicities below 200 nm and negative helicities between 205-220 nm which indicates the formation of β-sheets.66–70 Peptide 7b showed a mixed structure of
38 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
Figure 2.9: Thioflavin T fluorescence emission of stirred DCLs made from peptide a) 1, b) 1a, c) 2, d) 3, e) 3a, f) 4, g) 4a, h) 5, i) 5a, j) 6, k) 6a, l) 7, m) 7a, n) 7b, o) 8 corresponding to Figure 2.3. Insets are magnifications of the parent plot.
Thioflavin T (ThT) fluorescence measurements71 are routinely used to confirm the amyloid type β-sheet conformation.72 ThT is an excellent chromophore with
a strong binding affinity towards β-sheets. It forms micellar structures in aqueous solutions in which its fluorescence is self-quenched. In the presence of amyloid type β-sheet structures, its fluorescence is recovered upon binding to the β-sheets. ThT fluorescence in the presence of DCLs made from peptide 2, 3, 4, 7, and 8 showed at least 15-fold increase at 490 nm (15-fold for 2 , 16-fold for 3, 32-fold for 4, 20-fold for 7, and 21-fold for 8) which confirms the formation of β-sheet amyloid-fibril-like
2.2. Results and Discussion 39
structures. Peptide 7b showed only a slight increase. ThT assay results for all other samples were negative.
2.2.2
C-termini Functionalization
MD simulations suggest that in our peptides, residues at or close to the C-termini are playing an important role in inter- and intra-molecular H-bonding and salt bridge formation and thus in the process of self-assembly.58 We decided to probe the effect
of introducing an amide functionality at the C-termini.
Without any exception, all the peptides with an amide group at the C-termini (1a-7a) yielded DCLs containing trimers and tetramers only. We synthesized 3a, 4a and 7a, the carboxy terminated analogues of which were found to make octamer replicators. The amide terminated peptide 3a, 4a, 7a (Figure 2.3e, 2.3g and 2.3m, respectively) formed mixtures of trimers and tetramers, we speculate that amidation stabilizes the smaller non-assembled macrocycles in the DCLs possibly through hy-drogen bonding, reducing the tendency to self-assemble. Only 3a formed tetramers almost exclusively. Morever, 3a is the only amidated peptide that self-assembled into fibers. In contrast to carboxyl analogue 3, which forms highly laterally associated short fibers of 38, TEM images of the DCL made from 3a shows very long and
flex-ible fibers (7.1 `´ 1.2 nm in width) of 3a4 (see Figure 2.123). Asn on the side chain
of 3a is a moiety that can act as a H-donor and H-acceptor simultaneously. Hence, this peptide is more prone to make intermolecular hydrogen bonding which might explain the ability of 3a4to self-assemble.
Amidation introduces extra amide H-bond possibilities, that can potentially in-crease β-sheet propensity. In short peptides that are amidated, the terminal amide is often more structured as evident from the IR spectrum: its peak is found at some-what higher frequency than peptide bonds, but still quite narrow.73 However, the
assemblies formed by 3a4 do not form a single stack but hierarchical aggregates,
complicating analysis by IR and preventing reliable analysis.
We decided to run MD simulations on the stacks of peptide 96with a carboxyl
C-terminus (Figure 2.10a) and with an amide C-C-terminus (Figure 2.10b). In the single macrocycle, the driving force for exposing all the Lys side chains to the solution is a bit stronger for the amidated analogue, as there is no salt-bridging possible. However, the core (DTB6) or internal backbone H-bonds did not show significant differences.
For the stack, ion retention at the surface of the fiber was somewhat increased upon amidation, and their occupancy roughly doubled as expected going from a 16x6 = +96 to a +192 electron system. Salt-bridges are released, increasing the space occupied by a single β-sheet, because the Lys residues need to be solvated now. For these amidated macrocycles, the structure appears to be more chiral (around 40%)
38 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
Figure 2.9: Thioflavin T fluorescence emission of stirred DCLs made from peptide a) 1, b) 1a, c) 2, d) 3, e) 3a, f) 4, g) 4a, h) 5, i) 5a, j) 6, k) 6a, l) 7, m) 7a, n) 7b, o) 8 corresponding to Figure 2.3. Insets are magnifications of the parent plot.
Thioflavin T (ThT) fluorescence measurements71 are routinely used to confirm the amyloid type β-sheet conformation.72 ThT is an excellent chromophore with
a strong binding affinity towards β-sheets. It forms micellar structures in aqueous solutions in which its fluorescence is self-quenched. In the presence of amyloid type β-sheet structures, its fluorescence is recovered upon binding to the β-sheets. ThT fluorescence in the presence of DCLs made from peptide 2, 3, 4, 7, and 8 showed at least 15-fold increase at 490 nm (15-fold for 2 , 16-fold for 3, 32-fold for 4, 20-fold for 7, and 21-fold for 8) which confirms the formation of β-sheet amyloid-fibril-like
2.2. Results and Discussion 39
structures. Peptide 7b showed only a slight increase. ThT assay results for all other samples were negative.
2.2.2
C-termini Functionalization
MD simulations suggest that in our peptides, residues at or close to the C-termini are playing an important role in inter- and intra-molecular H-bonding and salt bridge formation and thus in the process of self-assembly.58 We decided to probe the effect
of introducing an amide functionality at the C-termini.
Without any exception, all the peptides with an amide group at the C-termini (1a-7a) yielded DCLs containing trimers and tetramers only. We synthesized 3a, 4a and 7a, the carboxy terminated analogues of which were found to make octamer replicators. The amide terminated peptide 3a, 4a, 7a (Figure 2.3e, 2.3g and 2.3m, respectively) formed mixtures of trimers and tetramers, we speculate that amidation stabilizes the smaller non-assembled macrocycles in the DCLs possibly through hy-drogen bonding, reducing the tendency to self-assemble. Only 3a formed tetramers almost exclusively. Morever, 3a is the only amidated peptide that self-assembled into fibers. In contrast to carboxyl analogue 3, which forms highly laterally associated short fibers of 38, TEM images of the DCL made from 3a shows very long and
flex-ible fibers (7.1 `´ 1.2 nm in width) of 3a4 (see Figure 2.123). Asn on the side chain
of 3a is a moiety that can act as a H-donor and H-acceptor simultaneously. Hence, this peptide is more prone to make intermolecular hydrogen bonding which might explain the ability of 3a4 to self-assemble.
Amidation introduces extra amide H-bond possibilities, that can potentially in-crease β-sheet propensity. In short peptides that are amidated, the terminal amide is often more structured as evident from the IR spectrum: its peak is found at some-what higher frequency than peptide bonds, but still quite narrow.73 However, the
assemblies formed by 3a4 do not form a single stack but hierarchical aggregates,
complicating analysis by IR and preventing reliable analysis.
We decided to run MD simulations on the stacks of peptide 96with a carboxyl
C-terminus (Figure 2.10a) and with an amide C-C-terminus (Figure 2.10b). In the single macrocycle, the driving force for exposing all the Lys side chains to the solution is a bit stronger for the amidated analogue, as there is no salt-bridging possible. However, the core (DTB6) or internal backbone H-bonds did not show significant differences.
For the stack, ion retention at the surface of the fiber was somewhat increased upon amidation, and their occupancy roughly doubled as expected going from a 16x6 = +96 to a +192 electron system. Salt-bridges are released, increasing the space occupied by a single β-sheet, because the Lys residues need to be solvated now. For these amidated macrocycles, the structure appears to be more chiral (around 40%)
40 2. A Structural Survey in the Search for Novel Self-Replicating Peptides when the distance restraints between the Cα atoms and the CONH2 group were switched off in the simulations. However, we see that the attraction between the sheets is lost, they take up more space around the z-axis and bind more Cl- ions. Also the macrocycles seem to be bound a bit more weakly to the stack. While this effect was not quantified, there is more “fraying” and more disorder at the C-termini because of the lost lateral inter-sheet interaction (although still not a lot). Overall, the conformations of the fiber do not appear very different, at least not within the short simulation time (50 ns). Note that, based on our results we cannot exclude that the amide group could potentially stabilize the smaller macrocycles (i.e. trimers and tetramers).
Figure 2.10: Side (left) and top (right) view snapshots from the simulation experiments on the peptide 96 fiber with a) the carboxylated C-termini and b) the amidated C-termini
2.2. Results and Discussion 41
2.2.3
Environmental Effects: Use of Phosphate-Buffered Saline
Leonetti et al.74 showed that when 30% trifluoroethanol (TFE) is used as a co-solvent in a DCL of peptide 4, the emergence of 46 replicator was observed instead of the emergence of the 48replicator. This result showed the effects of environmental factors on the emergence of replicators from DCLs and extended the possible means to control the ring size of self-replicating macrocycles. It is well established that buffer along with the buffer-salt concentration has significant effects on the conformational stability of proteins.75 Additionally, different buffers at the same pH can affect the disulfide-thiol exchange differently.76 These considerations prompted us to change the buffer and increase the ionic strength using phosphate-buffered saline (PBS) (50 mM in phosphate, 150 mM in NaCl, pH 8.2) for some of the peptide libraries.
Figure 2.11: TEM micrographs of DCLs consisting of a) 48 b) 78 and c) 96 prepared in
PBS (50 mM in phosphate, 150 mM in NaCl, pH 8.2), d) 48 e) 78 and f) 96 prepared in
borate buffer (50 mM, pH 8.2).
As discussed above, 2 forms trimers and tetramers in DCLs prepared using stan-dard protocols, while fast oxidation of high-purity building block leads to formation of large oligomers. Similar to fast oxidation, DCLs set up in PBS also lead to the formation of oligomers up to tridecamer (Figure 2.57). We observed large amorphous aggregates by TEM (Figure 2.121) whereas no ordered or disordered structure was
42 2. A Structural Survey in the Search for Novel Self-Replicating Peptides observed for the DCL of similar composition obtained via fast oxidation. Similarly, peptide 8 forms 88 which self-assembles into long fiber bundles of width 14.8 `´ 1.6 nm in DCLs in PBS (Figure 2.126), whereas in borate buffer it forms 83. We also observed morphological difference between the assemblies formed in DCLs set up in PBS and in borate buffer. Figure 2.11a-c shows short and laterally associated fibers when PBS is used for the preparation of DCLs of 4, 7 and 9. In contrast, Figure 2.11d-f shows long single fibers or long fiber bundles of two single fibers for the cor-responding peptides prepared in borate buffer. This enhanced tendency for lateral association of fibers in PBS may be a result of charge-reinforced hydrogen bonding interactions between the ammonium groups on the fibers and phosphate from the buffer, as shown in Figure 2.13.
Figure 2.12: Comparison of growth rates of a) 48 and b) 96in PBS (50 mM in phosphate,
150 mM in NaCl, pH 8.2), red lines; and in borate buffer (50 mM, pH 8.2), blue lines.
In addition to the morphological differences, we observed that the rate of conver-sion of the monomers into replicators is much faster in PBS as compared to borate buffer at the same pH (Figure 2.12). This enhanced rate in PBS may be related to the fact that in PBS fibers tend to be shorter; possibly because of the high lateral association provided by the high salt concentration. PBS can reduce the charge re-pulsions between the fibers due to high salt concentration and salt bridging between NH3+and the HPO42- (Figure 2.13). With a constant amount of available material, shorter fibers means a large number of fiber ends (from which the fibers grow), re-sulting an increase in the rate of fiber growth; i.e. replication. This beneficial effects of PBS on replication rates is why in Chapter 5, starting solutions of replicators are grown in the PBS at high concentration and then diluted with borate buffer. We can grow a family of replicators from two building block libraries in three weeks in borate buffer whereas in PBS same results can be obtained within a week. As a sidenote we have to mention that peptide batches purchased from a commercial supplier some-times have low purity (lower than 90%) or degrade over time to some extent. The
2.2. Results and Discussion 43 resulting impurities can have a significant effect on the overall library behavior and in many cases, emergence of the replicators is hindered. In some instances, where there were problematic batches of 4, 7 and 9 which failed to produce the corresponding replicators of 48, 78 and 96 in borate buffers, their emergence was observed using PBS.
Figure 2.13: Schematic representation of salt bridging between NH3+ and the HPO42-.
2.2.4
Conclusions
In summary, we have designed and synthesized 13 novel peptide building blocks and analyzed the DCLs of these along with the 3 peptide building blocks that are widely studied in our group. We varied the overall hydrophobicity of a building block by replacing a residue at a certain position, including the introduction of amide groups at the C-termini or the replacement of Lys with Asp to change the overall charge state of the building blocks. Furthermore, we studied the effects of the medium on the emergence of replicators in DCLs in phosphate buffered saline.
The results have improved our understanding of the emergence of replicators and the behavior of DCLs which are highly dependent on the peptide sequence. Minor structural changes can significantly influence the product distribution in the DCL. For most of the peptides, a negative charge at the C-terminus is essential for self-assembly and self-replication.
In our survey, we discovered new replicators, 38 and 88. The emergence of these replicators is in line with the previously postulated correlation between replicator ring size and peptide hydrophobicity. However, introducing polar groups (phosphorylated amino acids) in the sequence close to the C-termini breaks this rule probably due to competing inter- and intra-molecular interactions. Among these libraries we obtained 14, 23 and 83 as (possible) replicators. Attempts to access large sized replicators (larger than octamers) were not successful, but we cannot exclude that under certain conditions they can be accessible, although the MD results suggest that there is not sufficient space for the peptide side chain to align oligomers larger than octamers to stack into fibers.
44 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
2.3
Acknowledgements
M. Altay is gratefully acknowledged for the in-depth discussions on peptide design and synthesis, negative staining TEM analysis and the exhaustive discussions on the results. Dr. M. Malakoutikhah is acknowledged for the project proposal and the preliminary results. Dr. P. W. J. M. Frederix is acknowledged for the molecular dy-namic simulations and proof-reading this chapter and for giving precious feedback. G. Leonetti and J. Ottele are acknowledged for the cryo-TEM analysis and the negative staining TEM imaging, respectively.
2.4. Materials and Methods 45
2.4
Materials and Methods
2.4.1
Materials
Doubly distilled water was used in all experiments. Boric anhydride (Sigma-Aldrich) and sodium hydroxide (Merck Chemicals) were utilized for buffer preparation and pH adjustments. For UPLC measurements, UPLC grade acetonitrile, water and trifluoroacetic acid were purchased from Biosolve BV.
Solvents used in the peptide synthesis were reagent grade. Dichloromethane, dimethylformamide, oxyma and N,N-diisopropylcarbodiimide (DIC) were used with-out further purification. Reagents were purchased from Sigma-Aldrich, Merck, Fluka, or Acros and used as received. Preloaded resins and Fmoc-protected amino acids were purchased from Merck Chemicals and Sigma Aldrich.
2.4.2
Peptide Synthesis
Peptides 1, 2, 3, 4, 5, 6, 7, 7b and 8 were synthesized by conventional peptide synthesis using pre-loaded Wang resins and peptides 1a, 3a, 4a, 5a, 6a and 7a were synthesized by conventional peptide synthesis using rink amide resins by cou-pling 3,5-bis(tritylthio)-benzoic acid, which was synthesized via a previously reported procedure.77We used DIC and oxyma as coupling agent and additive, respectively,
throughout the synthesis. In the synthesis of peptide 1, 1a, 2, 2a; phosphorylated amino acids were coupled 3 times consecutively to avoid deletion products as much as possible. Crude peptides were purified by preparative HPLC using a Phenomenex, Jupiter (10µm, C5, 300 ˚A, 250ˆ21.2 mm) column. Peptides 1, 2, 3, 4, 5, 7, 8 and 9 were also synthesized by Cambridge Peptides Ltd. (Birmingham, U.K.) for purities higher than 98%.
All peptides had purity higher than 95%. Impurities were mainly resulting from the formation of disulfides (i.e. dimers, trimers etc.)
2.4.3
Library Preparation and Sampling
Building blocks were dissolved in borate buffer (50 mM, pH 8.2) to prepare a library of peptides at 3.8 mM concentration. The volume of each library was at least 500µL. All libraries were equilibrated in an HPLC vial (12ˆ32 mm) with a Teflon-coated screw cap. All HPLC vials were equipped with a cylindrical stirrer bar (2ˆ5 mm, Teflon coated, purchased from VWR) and were stirred at 1200 rpm using an IKA RCT basic hot plate stirrer. All experiments were performed at ambient conditions. For sampling, 10µL of each sample was transferred to another vial containing 10 µL DMF and diluted to 200µL with UPLC grade water prior to UPLC or LC-MS analysis.
44 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
2.3
Acknowledgements
M. Altay is gratefully acknowledged for the in-depth discussions on peptide design and synthesis, negative staining TEM analysis and the exhaustive discussions on the results. Dr. M. Malakoutikhah is acknowledged for the project proposal and the preliminary results. Dr. P. W. J. M. Frederix is acknowledged for the molecular dy-namic simulations and proof-reading this chapter and for giving precious feedback. G. Leonetti and J. Ottele are acknowledged for the cryo-TEM analysis and the negative staining TEM imaging, respectively.
2.4. Materials and Methods 45
2.4
Materials and Methods
2.4.1
Materials
Doubly distilled water was used in all experiments. Boric anhydride (Sigma-Aldrich) and sodium hydroxide (Merck Chemicals) were utilized for buffer preparation and pH adjustments. For UPLC measurements, UPLC grade acetonitrile, water and trifluoroacetic acid were purchased from Biosolve BV.
Solvents used in the peptide synthesis were reagent grade. Dichloromethane, dimethylformamide, oxyma and N,N-diisopropylcarbodiimide (DIC) were used with-out further purification. Reagents were purchased from Sigma-Aldrich, Merck, Fluka, or Acros and used as received. Preloaded resins and Fmoc-protected amino acids were purchased from Merck Chemicals and Sigma Aldrich.
2.4.2
Peptide Synthesis
Peptides 1, 2, 3, 4, 5, 6, 7, 7b and 8 were synthesized by conventional peptide synthesis using pre-loaded Wang resins and peptides 1a, 3a, 4a, 5a, 6a and 7a were synthesized by conventional peptide synthesis using rink amide resins by cou-pling 3,5-bis(tritylthio)-benzoic acid, which was synthesized via a previously reported procedure.77 We used DIC and oxyma as coupling agent and additive, respectively,
throughout the synthesis. In the synthesis of peptide 1, 1a, 2, 2a; phosphorylated amino acids were coupled 3 times consecutively to avoid deletion products as much as possible. Crude peptides were purified by preparative HPLC using a Phenomenex, Jupiter (10µm, C5, 300 ˚A, 250ˆ21.2 mm) column. Peptides 1, 2, 3, 4, 5, 7, 8 and 9 were also synthesized by Cambridge Peptides Ltd. (Birmingham, U.K.) for purities higher than 98%.
All peptides had purity higher than 95%. Impurities were mainly resulting from the formation of disulfides (i.e. dimers, trimers etc.)
2.4.3
Library Preparation and Sampling
Building blocks were dissolved in borate buffer (50 mM, pH 8.2) to prepare a library of peptides at 3.8 mM concentration. The volume of each library was at least 500µL. All libraries were equilibrated in an HPLC vial (12ˆ32 mm) with a Teflon-coated screw cap. All HPLC vials were equipped with a cylindrical stirrer bar (2ˆ5 mm, Teflon coated, purchased from VWR) and were stirred at 1200 rpm using an IKA RCT basic hot plate stirrer. All experiments were performed at ambient conditions. For sampling, 10µL of each sample was transferred to another vial containing 10 µL DMF and diluted to 200µL with UPLC grade water prior to UPLC or LC-MS analysis.
46 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
2.4.4
Seeding Experiments
A freshly prepared library of a peptide 3 (3.8 mM) was left stirring for 2 days such that it contained mostly monomer, trimer and tetramer. On day 2, a previously prepared solution containing predominantly octamer, 38 (10 mol %) was added. Analogues
seeding experiments were conducted on a freshly prepared library made from peptide 8 (3.8 mM) pre-oxidized 80 mol % using a perborate solution (80 mM) such that it contained mostly monomer, trimer and tetramer. After stirring for 30 min, a previously prepared solution containing predominantly octamer, 88 (10 mol %) was
added. The libraries were monitored with UPLC and UPLC-MS.
2.4.5
Circular Dichroism (CD)
Samples were prepared by diluting libraries to 0.24 mM (with respect to building block). All spectra were collected in a Jasco J-715 CD spectropolarimeter using HELMA quartz cuvettes with path length of 1.0 mm. Continuous spectra were obtained from 190 nm to 300 nm at room temperature. All reported spectra were an average of three scans recorded at a scan rate of 200 nm/min with 2-nm step interval and measured in millidegrees. Solvent spectra were subtracted from all spectra.
2.4.6
Thioflavin T (ThT) Fluorescence Assay
A ThT stock solution (2.2 mM) was prepared in 10 mL phosphate buffer (50 mM phosphate, 150 mM NaCl, pH 8.2) and filtered through a 0.2µm syringe filter. On the day of analysis, 50µL of the stock solution was diluted into 5 mL phosphate buffer (50 mM phosphate, 150 mM NaCl, pH 8.2) to generate the working solution of 22µL. The fluorescence intensity of 450µL ThT solution was measured by excitation at 440 nm (slit width 5 nm) and emission between 480-700 nm (slit width 5 nm), averaging 3 accumulations. An aliquot of 80µL of peptide solution (100 µL in borate buffer) was added to a HELMA 10ˆ2 mm quartz cuvette, incubated for 2 min, and the intensity was measured over 3 accumulations. All fluorescence measurements were done on a JASCO FP6200 fluorimeter equipped with a 480 nm high pass cut-off filter on the emission channel to avoid high order diffractions coming from the excitation.
2.4.7
Negative Staining Transmission Electron Microscopy
A small drop (5µL) of sample was deposited on a 400 mesh copper grid covered with a thin carbon film (Agar Scientific). After 30 s, the droplet was blotted on filter paper. The sample was then stained twice (4µL each time) with a solution of 2% uranyl acetate deposited on the grid and blotted on the filter paper after 30 s each time.
2.4. Materials and Methods 47
The grids were observed in a Philips CM120 cryo-electron microscope operating at 120 kV. Images were recorded on a slow scan CCD camera.
2.4.8
Molecular Dynamic Simulations
All MD simulations were performed using the GROMOS 54a8 force field78,79 with
in-house developed parameters for the aromatic disulfide head group (see below). Simulations were started from a straight (achiral) stacked structure. SPC water80
was added in a layer of at least 1 nm thickness around the fiber and enough chloride ions were added to neutralize the system. All explicit hydrogen atoms were converted to virtual sites and all molecular bonds were constrained in production runs using the LINCS algorithm,81 except for water molecules which were constrained by the
efficient SETTLE algorithm.82Additionally, the center-of-mass motion of the solute
(both translation and rotation) were removed every 100 time steps. After a short minimization of 5000 steepest descent steps, all runs were performed with time steps of 2.0 fs in the NPT ensemble using the velocity-rescaling thermostat83to separately
couple the peptide macrocycles and the rest of the system (water and ions) to an external bath at 298 K with a coupling constant of 1.0 ps. The pressure was kept at 1.0 bar using the Berendsen barostat84 with a coupling constant of 1.5 ps. A
Barker-Watts reaction field with ϵRF= 62 was used to treat long-range electrostatic
interactions with Coulombic and van der Waals forces cut off at 1.4 nm. All MD simulations were run using GROMACS 4.6.7.85Trajectories of 100 ns were generated
in two steps: during the first 50 ns distance restraints with a force constant of 1000 kJ/mol were applied to keep Cα atoms of nearest neighbors in the other peptide
strands at a constant distance of 0.48 nm. These restraints were then removed for the second period of 50 ns.
2.4.9
Parametrization of aromatic disulfide head group
For peptide 9, the parameters available in the GROMOS54a8 force field were used. However, not all parameters for the aromatic disulfide head group of the peptide were available.
Bonded parameters: As flexibility around the disulfide bond and the linker to the peptide strand was considered important, the dihedrals related to these bonds were parameterized by comparing the MD single point energy to QM single point energies at the BP86/DZVP level of theory using GAMESS UK on the system of the truncated dimer depicted in Figure 2.14. Ryckaert-Bellemans (RB) or standard dihedral potentials were fitted to the difference in energy and added to the force field.
46 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
2.4.4
Seeding Experiments
A freshly prepared library of a peptide 3 (3.8 mM) was left stirring for 2 days such that it contained mostly monomer, trimer and tetramer. On day 2, a previously prepared solution containing predominantly octamer, 38 (10 mol %) was added. Analogues
seeding experiments were conducted on a freshly prepared library made from peptide 8 (3.8 mM) pre-oxidized 80 mol % using a perborate solution (80 mM) such that it contained mostly monomer, trimer and tetramer. After stirring for 30 min, a previously prepared solution containing predominantly octamer, 88 (10 mol %) was
added. The libraries were monitored with UPLC and UPLC-MS.
2.4.5
Circular Dichroism (CD)
Samples were prepared by diluting libraries to 0.24 mM (with respect to building block). All spectra were collected in a Jasco J-715 CD spectropolarimeter using HELMA quartz cuvettes with path length of 1.0 mm. Continuous spectra were obtained from 190 nm to 300 nm at room temperature. All reported spectra were an average of three scans recorded at a scan rate of 200 nm/min with 2-nm step interval and measured in millidegrees. Solvent spectra were subtracted from all spectra.
2.4.6
Thioflavin T (ThT) Fluorescence Assay
A ThT stock solution (2.2 mM) was prepared in 10 mL phosphate buffer (50 mM phosphate, 150 mM NaCl, pH 8.2) and filtered through a 0.2µm syringe filter. On the day of analysis, 50µL of the stock solution was diluted into 5 mL phosphate buffer (50 mM phosphate, 150 mM NaCl, pH 8.2) to generate the working solution of 22µL. The fluorescence intensity of 450µL ThT solution was measured by excitation at 440 nm (slit width 5 nm) and emission between 480-700 nm (slit width 5 nm), averaging 3 accumulations. An aliquot of 80µL of peptide solution (100 µL in borate buffer) was added to a HELMA 10ˆ2 mm quartz cuvette, incubated for 2 min, and the intensity was measured over 3 accumulations. All fluorescence measurements were done on a JASCO FP6200 fluorimeter equipped with a 480 nm high pass cut-off filter on the emission channel to avoid high order diffractions coming from the excitation.
2.4.7
Negative Staining Transmission Electron Microscopy
A small drop (5µL) of sample was deposited on a 400 mesh copper grid covered with a thin carbon film (Agar Scientific). After 30 s, the droplet was blotted on filter paper. The sample was then stained twice (4µL each time) with a solution of 2% uranyl acetate deposited on the grid and blotted on the filter paper after 30 s each time.
2.4. Materials and Methods 47
The grids were observed in a Philips CM120 cryo-electron microscope operating at 120 kV. Images were recorded on a slow scan CCD camera.
2.4.8
Molecular Dynamic Simulations
All MD simulations were performed using the GROMOS 54a8 force field78,79 with
in-house developed parameters for the aromatic disulfide head group (see below). Simulations were started from a straight (achiral) stacked structure. SPC water80
was added in a layer of at least 1 nm thickness around the fiber and enough chloride ions were added to neutralize the system. All explicit hydrogen atoms were converted to virtual sites and all molecular bonds were constrained in production runs using the LINCS algorithm,81 except for water molecules which were constrained by the
efficient SETTLE algorithm.82Additionally, the center-of-mass motion of the solute
(both translation and rotation) were removed every 100 time steps. After a short minimization of 5000 steepest descent steps, all runs were performed with time steps of 2.0 fs in the NPT ensemble using the velocity-rescaling thermostat83to separately
couple the peptide macrocycles and the rest of the system (water and ions) to an external bath at 298 K with a coupling constant of 1.0 ps. The pressure was kept at 1.0 bar using the Berendsen barostat84 with a coupling constant of 1.5 ps. A
Barker-Watts reaction field with ϵRF = 62 was used to treat long-range electrostatic
interactions with Coulombic and van der Waals forces cut off at 1.4 nm. All MD simulations were run using GROMACS 4.6.7.85Trajectories of 100 ns were generated
in two steps: during the first 50 ns distance restraints with a force constant of 1000 kJ/mol were applied to keep Cα atoms of nearest neighbors in the other peptide
strands at a constant distance of 0.48 nm. These restraints were then removed for the second period of 50 ns.
2.4.9
Parametrization of aromatic disulfide head group
For peptide 9, the parameters available in the GROMOS54a8 force field were used. However, not all parameters for the aromatic disulfide head group of the peptide were available.
Bonded parameters: As flexibility around the disulfide bond and the linker to the peptide strand was considered important, the dihedrals related to these bonds were parameterized by comparing the MD single point energy to QM single point energies at the BP86/DZVP level of theory using GAMESS UK on the system of the truncated dimer depicted in Figure 2.14. Ryckaert-Bellemans (RB) or standard dihedral potentials were fitted to the difference in energy and added to the force field.
48 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
Figure 2.14: Aromatic head group dimer used for parameterizing dihedral angles in the GROMOS54a8 force field. The colors of the arrows correspond to the colors in Table 2.2
Table 2.2: Dihedral angle parameters for the three dihedral angles indicated in Figure 2.14
Dihedral (RB) C0 C1 C2 C3 C4 C5
1 (CSSC) RB -2.922 -5.231 13.782 8.33 19.196 -14.391
2(CSSC) RB -0.232 2.767 4.904 0.814 -9.281 -2.235
Dihedral Phase Force
Constant Multiplicity
3(CCCO) 180 11.0 2
Validation of non-bonded parameters: CH3Cl : H2O partitioning was mea-sured on a model system of the aromatic dithiol head group attached to a triethylene glycol tail (Figure 2.15) by UPLC. The equivalent partitioning was calculated us-ing thermodynamic integration usus-ing the default GROMOS54a8 atom types. Good agreement was found between experimental and calculated results.
∆∆GCH3ClÑ water, experimental = 3.3 `´ 1.3 kJ/mol
∆∆GCH3ClÑ water, calculated = 3.6 `´ 0.8 kJ/mol
Figure 2.15: Model system used for parameterization of non-bonded parameters of the dithiol aromatic head group. Synthesis of MePEG3 is described elsewhere.86
2.5. UPLC, HPLC and LC-MS analyses 49
2.5
UPLC, HPLC and LC-MS analyses
UPLC analyses were performed on a Waters Acquity UPLC I-class and H-class system equipped with a PDA detector. A reversed-phase UPLC column (Aeris 1.7µm. XB-C18 150ˆ 2.10 mm, purchased from Phenomenex) was used in the analyses of all samples, while UV absorbance was monitored at 254 nm. The column temperature was equilibrated at 35˝C prior to injections.
UPLC-MS analyses were performed using a Waters Acquity UPLC H-class system coupled to a Waters Xevo-G2 TOF. The mass spectrometer was operated under the following conditions: electrospray ionization mode: positive; capillary voltage: 2.5 kV; sampling cone voltage: 30 V; extraction cone voltage: 4.0 V. Samples were prepared by dissolving 10µL of a 3.8 mM library in 10 µL DMF and then further diluting them with 180µL UPLC grade water. The operating parameters were the following: eluent flow rate 0.3 mL/min; eluent A: UPLC grade water with 0.1 v% trifluoroacetic acid; eluent B: UPLC grade acetonitrile with 0.1 v% trifluoroacetic acid.
2.5.1
UPLC Methods
Libraries were analysed using the following method (linear gradient) and Phenomenex Aeris Peptide column.
Solvent A: ULC/MS grade water purchased from Biosolve (0.1 v% trifluoroacetic acid added)
Solvent B: ULC/MS grade acetonitrile purchased from Biosolve (0.1 v% triflu-oroacetic acid added)
Time, min. A% B% 0.00 90 10 1.00 90 10 1.30 75 25 3.00 72 28 11.00 60 40 11.50 5 95 12.00 5 95 12.50 90 10 17.00 90 10
50 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
2.5.2
HPLC Methods
Solutions containing peptides 4 and 7 and their oxidation products were analyzed using the following methods (all gradients are linear) using phenyl hexyl column with the following eluents:
Solvent A: double distilled water (0.1 v% trifluoroacetic acid added)
Solvent B: ULC/MS grade acetonitrile purchased from Biosolve (0.1 v% triflu-oroacetic acid added)
Time, min. A% B%
0.00 81 19
40.00 74 26
2.5. UPLC, HPLC and LC-MS analyses 51
2.5.3
UPLC and LC-MS analysis of Peptide 1
Figure 2.16: UPLC trace (monitored at 254 nm) of monomer of the peptide 1 (3.8 mM in 50 mM borate buffer, pH 8.2) after purification.
Figure 2.17: Mass spectrum of monomer 11 from the LC-MS analysis of peptide 1 after
purification (corresponding to Figure 2.16). Observed isotopic profile for [M+H]+(species,
abundance): 780.24 (M, 100%), 781.25 (M+1, 40.74%), 782.26 (M+2, 18.45%); m/z cal-culated: 780.28 [M+H]1+, 391.14 [M+2H]2+; m/z observed: 780.24 [M+H]1+, 390.62
52 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
2.5.4
UPLC and LC-MS analysis of Peptide 1a
Figure 2.18: UPLC trace (monitored at 254 nm) of monomer of the peptide 1a (3.8 mM in 50 mM borate buffer, pH 8.2) after purification.
Figure 2.19: Mass spectrum of monomer 1a1 from the LC-MS analysis of peptide 1a
after purification (corresponding to Figure 2.18). Observed isotopic profile for [M+H]+
(species, abundance): 779.26 (M, 100%), 780.27 (M+1, 45.34%), 781.26 (M+2, 30.26%); m/z calculated: 779.30 [M+H]1+, 390.65 [M+2H]2+; m/z observed: 779.26 [M+H]1+,
2.5. UPLC, HPLC and LC-MS analyses 53
2.5.5
UPLC and LC-MS analysis of Peptide 2
Figure 2.20: UPLC trace (monitored at 254 nm) of monomer of the peptide 2 (3.8 mM in 50 mM borate buffer, pH 8.2) after purification.
Figure 2.21: Mass spectrum of monomer 21 from the LC-MS analysis of peptide 2
af-ter purification (corresponding to Figure 2.20). Observed isotopic profile for [M+2H]2+
(species, abundance): 428.76 (M, 100%), 429.26 (M+1, 58.79%), 429.76 (M+2, 30.23%); m/z calculated: 856.31 [M+H]1+, 429.16 [M+2H]2+; m/z observed: 856.13 [M+H]1+,
54 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
2.5.6
UPLC and LC-MS analysis of Peptide 3
Figure 2.22: UPLC trace (monitored at 254 nm) of monomer of the peptide 3 (3.8 mM in 50 mM borate buffer, pH 8.2) after purification.
Figure 2.23: Mass spectrum of monomer 31 from the LC-MS analysis of peptide 3 after
purification (corresponding to Figure 2.22). Observed isotopic profile for [M+H]+(species,
abundance): 727.23 (M, 100%), 728.23 (M+1, 38.08%), 729.23 (M+2, 14.89%); m/z cal-culated: 727.33 [M+H]1+, 364.67 [M+2H]2+; m/z observed: 727.23 [M+H]1+, 364.29
2.5. UPLC, HPLC and LC-MS analyses 55
2.5.7
UPLC and LC-MS analysis of Peptide 3a
Figure 2.24: UPLC trace (monitored at 254 nm) of monomer of the peptide 3a (3.8 mM in 50 mM borate buffer, pH 8.2) after purification.
Figure 2.25: Mass spectrum of monomer 3a1 from the LC-MS analysis of peptide 3a
after purification (corresponding to Figure 2.24). Observed isotopic profile for [M+H]+
(species, abundance): 726.24 (M, 100%), 727.25 (M+1, 30.18%), 728.24 (M+2, 11.69%); m/z calculated: 726.34 [M+H]1+, 364.17 [M+2H]2+; m/z observed: 726.24 [M+H]1+,
56 2. A Structural Survey in the Search for Novel Self-Replicating Peptides
2.5.8
UPLC and LC-MS analysis of Peptide 4
Figure 2.26: UPLC trace (monitored at 254 nm) of monomer of the peptide 4 (3.8 mM in 50 mM borate buffer, pH 8.2) after purification.
Figure 2.27: Mass spectrum of monomer 41 from the LC-MS analysis of peptide 4 after
purification (corresponding to Figure 2.26). Observed isotopic profile for [M+H]+(species,
abundance): 700.24 (M, 100%), 701.25 (M+1, 43.28%), 702.24 (M+2, 15.48%); m/z cal-culated: 700.32 [M+H]1+, 351.16 [M+2H]2+; m/z observed: 700.24 [M+H]1+, 350.80
2.5. UPLC, HPLC and LC-MS analyses 57
2.5.9
UPLC and LC-MS analysis of Peptide 4a
Figure 2.28: UPLC trace (monitored at 254 nm) of monomer of the peptide 4a (3.8 mM in 50 mM borate buffer, pH 8.2) after purification.
Figure 2.29: Mass spectrum of monomer 4a1 from the LC-MS analysis of peptide 4a
after purification (corresponding to Figure 2.28). Observed isotopic profile for [M+H]+
(species, abundance): 699.26 (M, 100%), 700.26 (M+1, 42.40%), 701.26 (M+2, 14.87%); m/z calculated: 699.33 [M+H]1+, 350.67 [M+2H]2+; m/z observed: 699.26 [M+H]1+,
