Final Report Summary - FOLDAPOP (An integrated peptide and foldamer chemistry approach towards pro-apoptotic trail mimetics)
Several death receptors (DR), including the receptors of the protein tumour necrosis factor related apoptosis inducing ligand (TRAIL), namely DR4 and DR5, have come into focus as targets for the selective activation of the cell death (apoptosis) pathway in tumours in vivo without causing toxicity to healthy cells. Pharmacological strategies presently pursued to exploit the unique cancer selectivity of the TRAIL pathway include the development of recombinant TRAIL and agonistic anti-DR5 antibodies. In this context, the identification of small molecules (e.g. peptidomimetics) that can both bind to TRAIL receptors and activate the TRAIL signalling pathway is of prime interest towards the development of new treatments complementary to conventional cancer therapy. Expected advantages of peptide-based drugs over therapeutic proteins include reduced immunogenicity, flexible storage and more accessible chemical production. The aim of the project FOLDAPOP was to develop peptidomimetic foldamers (unnatural oligomers with predictable conformations) as proapoptotic DR ligands by capitalising on recent achievements of the host group in the fields of peptide and foldamer chemistries. The research plan included the following specific objectives:
(1) exploration of structural requirements for peptide-binding to DR5;
(2) evaluation of foldameric systems for polyvalent display;
(3) structure-guided design of DR5-binding helical foldamers; and
(4) construction and screening of focused foldamer libraries for binding to DRs.
To explore the structural requirements of peptide-binding to DR5, we have designed and synthesised a series of analogues of TRAILmim (M1c), a 16-mer DR5-binding peptide with a disulfide bridge between Cys3 and Cys13. A new macrocyclisation procedure with iodine as an oxidising agent was developed to obtain the cyclic peptides in shorter time (30 minutes instead of 72 hours) with the same or better purity. Two analogues of M1c with single point mutations in position 1 were synthesised by solid phase peptide synthesis (SPPS) methodology with fluorenylmethyloxycarbonyl chloride (Fmoc) strategy and diisopropylcarbodiimide / hydroxybenzotriazole (DIC / HOBt) as coupling reagents. Tryptophan (Trp) at P1 was substituted by more constrained cyclic derivatives with tetrahydro-beta-carboline and azepinone type rings. Unfortunately both Trp-mutants did not interact with DR5, suggesting the importance and the sensitivity of the P1 residue. Disulfide bridges such as in M1c are susceptible to reduction and scrambling in vivo by glutathione and redox enzymes potentially resulting in the loss of the desired biological activity of the peptide. To increase the redox stability of M1c, another analogue in which the disulfide bridge (Cys3 - Cys13) is substituted for a lanthionine thioether linkage was prepared. Its synthesis involved the desulfurisation of the M1c peptide under basic conditions. It is known that M1c as a monomer is not activating the TRAIL-DR5 apoptosis pathway, but following oligomerisation (dimerisation or trimerisation) it is converted to a full agonist. Therefore, the desulfurised monomer was also dimerised with an appropriate linker and the resulting monomers and dimers were tested for molecular interaction with recombinant hDR5 using surface plasmon resonance (SPR) method and showed very promising DR5 binding properties.
As mentioned above, multimerisation of TRAIL-mimetic results in higher avidity for the receptor and is a prerequisite for receptor activation. More elaborated multivalent ligands with controlled distance and topology can be designed to investigate multivalency effects. The programmability of helical oligourea foldamers developed by the Guichard group is an attractive solution towards this goal. A unique solid-phase synthesis (SPS) methodology has been developed during this project to chemoselectively attach DR5-binding peptides including M1c onto short-chain multivalent helical foldameric templates (< 10 residues). The methodology works very well, giving satisfactory purity and yield. In the meantime in the host laboratory another post-doc has been working on the solution phase synthesis of urea-based foldamers as a platform for biologically active peptides. The solution methodology is highly time consuming (difficulties during the purification) so the development of the new SPS-protocol may significantly accelerate the obtaining of new compounds.
Another specific objective of the project FOLDAPOP was the construction and the screening of focused on-bead foldamer libraries for binding to DRs. Helix forming urea/gamma-amide oligomer hybrids have been chosen for library construction because of the predictability of their folding pattern and the possibility to use a vast array of monomers with high side chain diversity. We introduced a new type of activated monomer, suitable for the rapid preparation of high quality libraries in which an azide is used as a masked amine and acid sensitive groups (Boc / tBu) protect functionalised side chains. Sixteen new monomers were synthesised in 6 steps from alpha-amino acids in yields ranging from 10 - 49 %.
Several model hybrids were then synthesised successfully, thus validating the methodology with azido succinimidyl carbamates and Fmoc-gamma-AA. Microwave irradiation was used to accelerate the synthesis. To convert the azido function into amino group, Staudinger reaction was selected with PMe3 as a reducing agent. All Fmoc-gamma-AA were coupled in the presence of HBTU/HOBt as a coupling reagent. The purity of all model compounds was always > 70 %. The combinatorial libraries were synthesised by standard split-and-mix procedure (after each coupling step, the beads were pooled (randomised) and then splitted into new groups for coupling with the next set of selected residues). 7 sublibraries of 8-mers with five variable (X) positions and three fixed were designed and prepared. Of several types of solid supports available, the TentaGel (polyoxyethylene-grafted polystyrene) was selected for OBOC library construction.
The X positions were varied by six of thirteen residues using a method of unique pairs, which reduces the number of redundant compounds, so that one sublibrary contained 31104 distinct foldamers. The difference between each sublibrary was the position of Trp-type residue. All 7 sublibraries were screened for binding to fluorescently labeled DR5 receptor. Screening and sorting of positive beads was performed as secondments in Barcelona using Complex Object Parametric Analyser and Sorter (COPAS) instrument located in prof. Albericio group at the Institute for Research in Biomedicine Barcelona (IRBB). A total of 10 hit (positive) beads from each library were manually selected. The deconvolution of libraries was performed using mass spectrometry in collaboration and with the help of Prof Schmitter in Bordeaux. It is worth mentioning that tandem mass spectrometry matrix-assisted laser desorption / ionisation (MS / MS MALDI) analyses gave fully exploitable results even though 1 bead contains only 40 pmol of the compound. 15 full sequences were identified of which 9 were resynthesised as dimers (because the activity of TRAIL mimics is higher upon multimerisation) and tested for apoptotic activity using colon cancer cells either expressing or KO for DR5 (collaboration with the Micheau group in Dijon). Preliminary data show that some of the compounds display apoptotic activity, but the mechanism of action seems unlikely to be mediated by DR5. SPR measurements are undergoing to confirm these results and further investigate the binding of dimeric foldamers to DR5.
Overall this work shows that it is actually feasible to generate and screen on-bead libraries of hybrid foldamers containing urea linkages and to sequence single beads using MS / MS. Though the approach still remains to be fully exploited to identify high affinity hits for DR5, it paves the way towards new applications of foldamers in chemical biology.
(1) exploration of structural requirements for peptide-binding to DR5;
(2) evaluation of foldameric systems for polyvalent display;
(3) structure-guided design of DR5-binding helical foldamers; and
(4) construction and screening of focused foldamer libraries for binding to DRs.
To explore the structural requirements of peptide-binding to DR5, we have designed and synthesised a series of analogues of TRAILmim (M1c), a 16-mer DR5-binding peptide with a disulfide bridge between Cys3 and Cys13. A new macrocyclisation procedure with iodine as an oxidising agent was developed to obtain the cyclic peptides in shorter time (30 minutes instead of 72 hours) with the same or better purity. Two analogues of M1c with single point mutations in position 1 were synthesised by solid phase peptide synthesis (SPPS) methodology with fluorenylmethyloxycarbonyl chloride (Fmoc) strategy and diisopropylcarbodiimide / hydroxybenzotriazole (DIC / HOBt) as coupling reagents. Tryptophan (Trp) at P1 was substituted by more constrained cyclic derivatives with tetrahydro-beta-carboline and azepinone type rings. Unfortunately both Trp-mutants did not interact with DR5, suggesting the importance and the sensitivity of the P1 residue. Disulfide bridges such as in M1c are susceptible to reduction and scrambling in vivo by glutathione and redox enzymes potentially resulting in the loss of the desired biological activity of the peptide. To increase the redox stability of M1c, another analogue in which the disulfide bridge (Cys3 - Cys13) is substituted for a lanthionine thioether linkage was prepared. Its synthesis involved the desulfurisation of the M1c peptide under basic conditions. It is known that M1c as a monomer is not activating the TRAIL-DR5 apoptosis pathway, but following oligomerisation (dimerisation or trimerisation) it is converted to a full agonist. Therefore, the desulfurised monomer was also dimerised with an appropriate linker and the resulting monomers and dimers were tested for molecular interaction with recombinant hDR5 using surface plasmon resonance (SPR) method and showed very promising DR5 binding properties.
As mentioned above, multimerisation of TRAIL-mimetic results in higher avidity for the receptor and is a prerequisite for receptor activation. More elaborated multivalent ligands with controlled distance and topology can be designed to investigate multivalency effects. The programmability of helical oligourea foldamers developed by the Guichard group is an attractive solution towards this goal. A unique solid-phase synthesis (SPS) methodology has been developed during this project to chemoselectively attach DR5-binding peptides including M1c onto short-chain multivalent helical foldameric templates (< 10 residues). The methodology works very well, giving satisfactory purity and yield. In the meantime in the host laboratory another post-doc has been working on the solution phase synthesis of urea-based foldamers as a platform for biologically active peptides. The solution methodology is highly time consuming (difficulties during the purification) so the development of the new SPS-protocol may significantly accelerate the obtaining of new compounds.
Another specific objective of the project FOLDAPOP was the construction and the screening of focused on-bead foldamer libraries for binding to DRs. Helix forming urea/gamma-amide oligomer hybrids have been chosen for library construction because of the predictability of their folding pattern and the possibility to use a vast array of monomers with high side chain diversity. We introduced a new type of activated monomer, suitable for the rapid preparation of high quality libraries in which an azide is used as a masked amine and acid sensitive groups (Boc / tBu) protect functionalised side chains. Sixteen new monomers were synthesised in 6 steps from alpha-amino acids in yields ranging from 10 - 49 %.
Several model hybrids were then synthesised successfully, thus validating the methodology with azido succinimidyl carbamates and Fmoc-gamma-AA. Microwave irradiation was used to accelerate the synthesis. To convert the azido function into amino group, Staudinger reaction was selected with PMe3 as a reducing agent. All Fmoc-gamma-AA were coupled in the presence of HBTU/HOBt as a coupling reagent. The purity of all model compounds was always > 70 %. The combinatorial libraries were synthesised by standard split-and-mix procedure (after each coupling step, the beads were pooled (randomised) and then splitted into new groups for coupling with the next set of selected residues). 7 sublibraries of 8-mers with five variable (X) positions and three fixed were designed and prepared. Of several types of solid supports available, the TentaGel (polyoxyethylene-grafted polystyrene) was selected for OBOC library construction.
The X positions were varied by six of thirteen residues using a method of unique pairs, which reduces the number of redundant compounds, so that one sublibrary contained 31104 distinct foldamers. The difference between each sublibrary was the position of Trp-type residue. All 7 sublibraries were screened for binding to fluorescently labeled DR5 receptor. Screening and sorting of positive beads was performed as secondments in Barcelona using Complex Object Parametric Analyser and Sorter (COPAS) instrument located in prof. Albericio group at the Institute for Research in Biomedicine Barcelona (IRBB). A total of 10 hit (positive) beads from each library were manually selected. The deconvolution of libraries was performed using mass spectrometry in collaboration and with the help of Prof Schmitter in Bordeaux. It is worth mentioning that tandem mass spectrometry matrix-assisted laser desorption / ionisation (MS / MS MALDI) analyses gave fully exploitable results even though 1 bead contains only 40 pmol of the compound. 15 full sequences were identified of which 9 were resynthesised as dimers (because the activity of TRAIL mimics is higher upon multimerisation) and tested for apoptotic activity using colon cancer cells either expressing or KO for DR5 (collaboration with the Micheau group in Dijon). Preliminary data show that some of the compounds display apoptotic activity, but the mechanism of action seems unlikely to be mediated by DR5. SPR measurements are undergoing to confirm these results and further investigate the binding of dimeric foldamers to DR5.
Overall this work shows that it is actually feasible to generate and screen on-bead libraries of hybrid foldamers containing urea linkages and to sequence single beads using MS / MS. Though the approach still remains to be fully exploited to identify high affinity hits for DR5, it paves the way towards new applications of foldamers in chemical biology.