Final Report Summary - DYNAREP (Self-replication in dynamic combinatorial libraries)
It was recently reported by our group that oxidising a stirred solution of peptide XGLKLK (1, X: 3,5-dimercaptobenzoic acid) produced heptamer. Furthermore, we found that peptides XGLKFK (2) and XGLKAK (3), when stirred, formed hexamer and octamer respectively. The three peptides differ only in their hydrophobic C-terminal amino acids leucine (Leu in 1), phenylalanine (Phe in 2) and alanine (Ala in 3). These findings suggest that the more hydrophobic this amino acid is the smaller the replicating macrocycle becomes. To discover further self-replicating peptides and explore the effect of the strength of peptide-peptide interaction on the self-replication process a library of four peptides was designed XGLK(1-Nal)K (4), XGLK(Cha)K (5), XGLK(p-Cl-Phe)K (6) and XGLK(Ser)K (7). The new peptide sequences were designed to be either less or more hydrophobic than peptide 1, by substituting the C-terminal leucine (Leu) of peptide 1 with more hydrophobic amino acids such as 1-naphthylalanine (1-Nal in 4), cyclohexylalanine (Cha in 5) and para-chlorophenylalanine (p-Cl-Phe in 6) or less hydrophobic amino acid like serine (Ser in 7). Our expectation was that a more hydrophobic sequence would give rise to more stable sheets, so that fibre formation (and concomitant self-replication) becomes feasible for a reduced macrocycle size. Conversely, a less hydrophobic peptide should give fibres assembled from larger macrocycles. Behaviour of the new peptides was investigated using a wide range of analytical techniques. We found that the more hydrophobic peptides (4, 5, 6) gave hexamer or smaller macrocycles, while the less hydrophobic peptide (7) produced octamer. All four peptides formed long fibres through beta-sheet formation as observed by transmission electron microscopy (TEM), thioflavin T assay and circular dichroism (CD). These findings show that the size of replicating macrocycle can be tuned by changing the peptide hydrophobicity. We previously proposed that fibres ends catalyse the formation of the very macrocycle from which the fibres are constituted. To further show the relationship between the number of fibre ends and fibre growth rate a set of experiments was performed for peptide 7. Six solutions of peptide 7 were prepared and each solution underwent a different rate of stirring ranging from 200 to 1 500 rpm. The composition of each solution was monitored by high-performance liquid chromatography (HPLC). Results show that the higher stirring rate is, the faster octamer grows. This could be attributed to more efficiency of higher rate to fragment fibres and consequently produce more fibre ends. To confirm this explanation the samples were analysed by TEM. Quantitative analysis of fibres by measuring fibre length revealed that an increase in stirring rate resulted in shorter fibres. These findings support the proposed mechanism for fibre formation in which the number of fibre ends play a crucial role.
To have self-replication inside the bilayer vesicles, first we attempted to trap pre-formed fibres from peptide 7 inside the liposomes. Analysing liposomes with cryo-TEM revealed that fibres were not trapped inside the liposomes; it may be due to the fact that fibres are longer than liposomes diameter. In another experiment we attempted to trap fibres inside synthetic vesicles provided by Dr Reink Eelkema a collaborator at the Delft University of Technology. Using cryo-TEM we observed that fibres were not inside the vesicles. We also attempted to make fibres shorter by sonication process before trapping them inside synthetic vesicles; however, we did not observe fibres inside the vesicles. These results suggest that achieving replication inside vesicles is more challenging than expected. Since para-chlorophenylalanine-containing peptide (p-Cl-Phe in 6) produced hexamer and pentamer as the main species under agitated and non-agitated conditions respectively, therefore, we investigated the effect of other halogen atoms on the size of macrocycles. We synthesised a library of three peptides para-fluorophenylalanine (p-F-Phe in 8), para-bromophenylalanine- (p-Br-Phe in 9) and para-iodophenylalanine (p-I-Phe in 10). Peptide penta-fluorophenylalanine (F5-Phe in 11) was also synthesised to study the influence of more halogen atom on behavior of peptide. We found that peptide 8 gave hexamer under agitated and non-agitated conditions. Peptides 9, 10 and 11 formed trimer as the main species. They formed pentamer and hexamer but these macrocycles did not grow to become the dominant product. All four peptides formed long fibres through beta-sheet formation as observed by TEM, thioflavin T assay and circular CD. As expected, fibres in the agitated samples were shorter than those from non-agitated samples. These results suggest that the size of halogen atom may be additional factor to control the size of macrocycles. Peptides 6 and 9 have similar hydrophobicity, but only peptide 6 having a smaller halogen atom formed hexamer as the dominant product. In order to establish that the larger macrocycles that emerge in the libraries are indeed self-replicating and, hence, capable of enhancing the rate of their own formation, we have performed a series of seeding experiment for peptides 2, 3, 5, 6 and 7. Addition of a small amount (5 mol %) of a sample that is rich in the suspected replicator (hexamer for 2 and 5 and octamer for 3 and 7) should induce a substantial increase in the rate of replicator formation in these samples. The data shows that the rate of formation of the larger macrocycles increases dramatically upon seeding, confirming that 26, 56, 38 and 78 are indeed self-replicators. For every replicator we performed two seeding experiments: one where the sample was stirred after being seeded and one where it was left non-agitated. Self-replication is somewhat more efficient in the stirred samples, although the difference is surprisingly small. The growth of hexamer and octamer in the non-agitated samples indicates the ability of these macrocycles to grow in the absence of agitation, once they have reached a certain concentration, even though under these conditions these compounds do not reach significant concentrations in the absence of seed on the timescale on which we monitored our experiments. This result points towards the importance of mechnically-induced fibre breakage in the early stages of the emergence of the replicators. Seeding experiments for peptide 6 were performed in a different way. A mother solution of peptide 6 chemically oxidised by sodium perborate containing mostly 3mer and 4mer was split in 4 parts: one was seeded by 5mol % 6mer and stirred, the second was not seeded and stirred, the third sample was seeded by 5mol % 5mer and left without agitation and the last solution was not seeded and let without agitation. We observed faster formation of 6mer in the seeded sample compared to the non-seeded sample under stirring condition showing self-replication properties of this species. However, 5mer grew in similar rate for both seeded and non-seeded samples while left without agitation. By repeating the experiments using 15 mol % 5mer similar results were obtained. These results show that the 5mer is not a self-replicating species.
In order to study the importance of aromatic interactions on fibre formation we studied mixed libraries of peptide 4 and 5 containing aromatic and non-aromatic amino acids respectively. After 3 weeks we observed a mixture of heterogeneous hexamer distributed statically. These findings may indicate the lack of preference for aromatic interactions in fibres formed from macrocycles. Similar results were obtained when peptides 2 and 11 were mixed confirming the effect of the size of amino acid on the size of macrocycles. Peptide 11 produced trimer as the dominant species, however, when mixed with peptide 2 it was incorporated in heterogeneous hexamer.
To have self-replication inside the bilayer vesicles, first we attempted to trap pre-formed fibres from peptide 7 inside the liposomes. Analysing liposomes with cryo-TEM revealed that fibres were not trapped inside the liposomes; it may be due to the fact that fibres are longer than liposomes diameter. In another experiment we attempted to trap fibres inside synthetic vesicles provided by Dr Reink Eelkema a collaborator at the Delft University of Technology. Using cryo-TEM we observed that fibres were not inside the vesicles. We also attempted to make fibres shorter by sonication process before trapping them inside synthetic vesicles; however, we did not observe fibres inside the vesicles. These results suggest that achieving replication inside vesicles is more challenging than expected. Since para-chlorophenylalanine-containing peptide (p-Cl-Phe in 6) produced hexamer and pentamer as the main species under agitated and non-agitated conditions respectively, therefore, we investigated the effect of other halogen atoms on the size of macrocycles. We synthesised a library of three peptides para-fluorophenylalanine (p-F-Phe in 8), para-bromophenylalanine- (p-Br-Phe in 9) and para-iodophenylalanine (p-I-Phe in 10). Peptide penta-fluorophenylalanine (F5-Phe in 11) was also synthesised to study the influence of more halogen atom on behavior of peptide. We found that peptide 8 gave hexamer under agitated and non-agitated conditions. Peptides 9, 10 and 11 formed trimer as the main species. They formed pentamer and hexamer but these macrocycles did not grow to become the dominant product. All four peptides formed long fibres through beta-sheet formation as observed by TEM, thioflavin T assay and circular CD. As expected, fibres in the agitated samples were shorter than those from non-agitated samples. These results suggest that the size of halogen atom may be additional factor to control the size of macrocycles. Peptides 6 and 9 have similar hydrophobicity, but only peptide 6 having a smaller halogen atom formed hexamer as the dominant product. In order to establish that the larger macrocycles that emerge in the libraries are indeed self-replicating and, hence, capable of enhancing the rate of their own formation, we have performed a series of seeding experiment for peptides 2, 3, 5, 6 and 7. Addition of a small amount (5 mol %) of a sample that is rich in the suspected replicator (hexamer for 2 and 5 and octamer for 3 and 7) should induce a substantial increase in the rate of replicator formation in these samples. The data shows that the rate of formation of the larger macrocycles increases dramatically upon seeding, confirming that 26, 56, 38 and 78 are indeed self-replicators. For every replicator we performed two seeding experiments: one where the sample was stirred after being seeded and one where it was left non-agitated. Self-replication is somewhat more efficient in the stirred samples, although the difference is surprisingly small. The growth of hexamer and octamer in the non-agitated samples indicates the ability of these macrocycles to grow in the absence of agitation, once they have reached a certain concentration, even though under these conditions these compounds do not reach significant concentrations in the absence of seed on the timescale on which we monitored our experiments. This result points towards the importance of mechnically-induced fibre breakage in the early stages of the emergence of the replicators. Seeding experiments for peptide 6 were performed in a different way. A mother solution of peptide 6 chemically oxidised by sodium perborate containing mostly 3mer and 4mer was split in 4 parts: one was seeded by 5mol % 6mer and stirred, the second was not seeded and stirred, the third sample was seeded by 5mol % 5mer and left without agitation and the last solution was not seeded and let without agitation. We observed faster formation of 6mer in the seeded sample compared to the non-seeded sample under stirring condition showing self-replication properties of this species. However, 5mer grew in similar rate for both seeded and non-seeded samples while left without agitation. By repeating the experiments using 15 mol % 5mer similar results were obtained. These results show that the 5mer is not a self-replicating species.
In order to study the importance of aromatic interactions on fibre formation we studied mixed libraries of peptide 4 and 5 containing aromatic and non-aromatic amino acids respectively. After 3 weeks we observed a mixture of heterogeneous hexamer distributed statically. These findings may indicate the lack of preference for aromatic interactions in fibres formed from macrocycles. Similar results were obtained when peptides 2 and 11 were mixed confirming the effect of the size of amino acid on the size of macrocycles. Peptide 11 produced trimer as the dominant species, however, when mixed with peptide 2 it was incorporated in heterogeneous hexamer.