Exploring Peptides as Template Scaffolds for Directed Olefin Metathesis Oligomerization Citation Poulin-Kerstien, Katherine G. (2008) Exploring Peptides as Template Scaffolds for Directed Olefin Metathesis Oligomerization. Master's thesis, California Institute of Technology. doi:10.7907/226t-1843. https://resolver.caltech.edu/CaltechTHESIS:10312023-230530609 Abstract [No Abstract] Item Type: Thesis (Master's thesis) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Tirrell, David A. Thesis Committee: Unknown, Unknown Defense Date: 2008 Record Number: CaltechTHESIS:10312023-230530609 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:10312023-230530609 DOI: 10.7907/226t-1843 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 16232 Collection: CaltechTHESIS Deposited By: Kathy Johnson Deposited On: 31 Oct 2023 23:59 Last Modified: 31 Oct 2023 23:59 Thesis Files PDF - Final Version See Usage Policy. 31MB Repository Staff Only: item control page

Exploring Peptides as Template Scaffolds for Directed Olefin Metathesis Oligomerization Thesis by Katherine G. Poulin-Kerstien In Partial Fulfillment of the Requirements for the Degree of Masters of Science California Institute of Technology Pasadena, California 2008 ©2008 Katherine G. Poulin-Kerstien All Rights Reserved 11 .. .for Adam ... 111 Acknowledgments First, I would like to thank my family for all their support and love - I couldn't have made it through this long journey without you. I would also like to thank all of the wonderful friends I've made during my time at Caltech, especially the "traveling scrapbook" ladies and my Snatch teammates, for making grad school a more enjoyable place to be all these years. To my labmates from both of the labs I spent time in, thank you for all your help, encouragement and commiseration along the way. Finally, I would like to thank Prof. David Tirrell for taking me into his lab and for always being so supportive of me and my goals. IV Table of Contents Acknowledgments ................................................................................... .iv Table of Contents ...................................................................................... v List of Figures and Tables .......................................................................... vii Chapter 1 Introduction: Taking a Cue from Nature ....................................... 1 Chapter 2 A Pseudorotaxane Recognition Motif.. ......................................... 8 Chapter 3 A Boronate Ester Recognition Motif.. ........................................ 51 List of Figures and Tables Chapter 1 Page Figure 1.1 PLGA product distribution 2 Figure 1.2 Ruthenium alkylidine metathesis catalysts 4 Figure 1.3 Types of metathesis chemistry 5 Figure 2.1 Structure of a [2]pseudorotaxane 9 Table 2.1 Influence of solvent on [2]pseudorotaxane formation 10 Table 2.2 Series of ammonium-bearing templates studied 11 Figure 2.2 Solid state structure of [3]pseudorotaxane 11 Figure 2.3 Helical peptides 12 Scheme 2.1 Synthesis of Pl 1 14 Figure 2.4 The BArF counterion 16 Figure 2.5 Series of alkane-linked templates, b-AmA 18 Scheme 2.2 Synthesis ofb-AmB 18 Scheme 2.3 Synthesis of M 19 Scheme 2.4 Synthesis of DB24C8-olefin derivatives 21 Table 2.3 R values for helical peptide templates 22 Figure 2.6 CD spectra of Pl 5 at different concentrations 22 Figure 2.7 CD spectra of Pl 1 with and without benzyl modification 22 Figure 2.8 CD spectra of Pll, P18 and P20 23 Figure 2.9 2D 1H NMR of b-Am F P11 24 Chapter 2 Vl Figure 2.10 2D NMR couplings 24 Table 2.4 Influence of solvent on M pseudorotaxane formation 26 Table 2.5 Influence of temperature on M pseudorotaxane formation 26 Table 2.6 Influence of equivalents ofDB24C8 on M pseudorotaxane formation Influence of Bn p-modification on M pseudorotaxane formation ESI-MS ofb-Amp 11 and DB24C8-CP pseudorotaxane 26 Figure 2.12 19 F NMR of M-p F and D B24C8-CP pseudorotaxane 30 Figure 2.13 19 F NMR of b-Amp 11 and DB24C8-CP pseudorotaxane 30 Figure 2.14 TOSCY spectra of b-Amp 11 without and with DB24C8-5en 32 Figure 2.15 Ruthenium alkylidene metathesis catalysts 33 Table 2.8 Influence of catalysts on template-directed metathesis 33 Table 2.9 Template-directed metathesis with DB24C8-e5en 34 Table 2.10 Influence of benzoquinone on template-directed metathesis 34 Figure 2.16 Binding modes for DB24C8 and benzyl-ammonium ion 35 Figure 2.17 36 Figure 2.18 Influence of different sizes of crown-ethers on templatedirected metathesis Influence of b-Amp 11 on DB24C8-CP ROMP Figure 2.19 Influence of b-Amp 11 on DB24C8-5en CM 38 Figure 2.20 39 Table 2.11 ESI-MS spectrum of CM reaction ofDB24C8-5en with bAmpn Comparison of peptide template-directed CM 40 Table 2.12 Comparison of template-directed AD MET 40 Table 2.13 Comparison ofb-AmA series of templates on ADMET 41 Table 2.14 Influence of template purity on AD MET 42 Table 2.15 Template-directed trimerization 43 Table 2.7 Figure 2.11 27 28 37 Vll Chapter 3 Table 3.1 Boronate ester equilibrium constants 53 Figure 3.1 Boronate ester formation 53 Table 3.2 Series ofboronic acid peptide template studied 54 Scheme 3.1 Synthesis of peptide templates 55 Scheme 3.2 Synthesis of diol-olefin monomers 57 Figure 3.2 59 Table 3.3 HPLC analysis of on-resin boronate ester fom1ation and hydrolysis Influence ofb-BAp 3 on CM Table 3.4 Influence ofb-BAp 3 on 1,3-diol-diene ADMET 60 Figure 3.3 HPLC analysis of MBHA-b-BAp 3 directed AD MET 61 Figure 3.4 HPLC analysis of on-resin tris-templates directed ADMET 62 59 vm Chapter 1 Introduction: Taking a Cue from Nature Background: Nature employs template-directed synthesis 1-4 by using noncovalent bonds and nn stacking interactions to obtain well-defined biopolymers - DNA, RNA, and proteins - in respect to their precise lengths, specific sequences, and stereochemical purity. This is in contrast to conventional polymerization techniques which give rise to heterogeneous populations of products. 5 For example, the traditional synthesis of poly(a ,L-glutamic acid) (PLGA) via ring-opening polymerization ofN-carboxy-a-glutamic acid anhydrides yields products with a broad distribution of molecular weights (Fig. 1.la). 6 On the other hand, harnessing biological machinery to generate PLGA via biosynthetic techniques yields a monodisperse product, with a molecular weight determined at the genetic level (Fig. 1.lb ). a) b) IJl&I. - ~ -.. J.__ - Do & ·L 2000 4000 6000 6000 10000 14000 20000 Figure 1.1: Comparison of the distribution of PLGA product produced by a) traditional polymer synthesis, and b) by biosynthesis. The versatility of genetically engineered polymer synthesis is limited by Nature's building blocks. DNA and RNA are limited to the five natural nucleotides and proteins are limited to the twenty natural amino acids. Extensive work has been done to expand the available pool of building blocks, including work with modified nucleic acids 7 and 2 non-canonical amino acids. 8' 9 While, peptides and proteins with novel characteristics have been generated by biosynthesis with non-canonical amino acids, 10, 11 polymers with drastically different properties might be accessible by means of to a different class of monomers. Examples of Template-Directed Polymerization: In an example of highly controlled polymerization, Liu and co-workers successfully used a DNA template to perform synthesis of peptide nucleic acids (PNAs ). 12 Employing a four base pair recognition codon, up to five PNA tetramer units were oligomerized in a sequence-dependent fashion via reductive amination chemistry. This experiment demonstrated tight template-controlled polymerization and high product yields. However, the resulting DNA-templated products are themselves DNA-based polymers, retaining characteristics of the template beyond sequence information. In addition to reductive amination, a range of other reactions have been directed by DNA templates in a sequence-specific fashion, including amine acylation, oxazolidine formation, Wittig olefination and the Huisgen cycloaddition. 13 None of these other reactions, however, have been employed for template-directed polymerization. In recent work, South and Weck demonstrated the template-directed synthesis of an artificial polymer by ruthenium alkylidene mediated olefin metathesis. 14 Ring opening metathesis polymerization (ROMP) of a norbomene-based thymine monomer was directed by a norbomene-polymer template functionalized with diaminopyridine recognition units, yielding a product with a PDI of 1.19. Without the template, ROMP of the thymine monomer yielded a polymer with a PDI of 1.73. While this work 3 demonstrated a template effect, the template itself was a norbomene polymer generated via ROMP, and so neither it nor the polymer product was monodisperse. Olefin metathesis, 15 ' 16 using ruthenium alkylidene catalysts, has been utilized in other templated-directed (non-polymer) syntheses, such as those of mechanically interlocked catenane- and rotaxane-based structures 17- 19 and for the dimerization of a hydrogen-bonded oligoamide duplex. 20 Scope of This Work: The goal of this research project was to establish a highly controllable method to synthesize non-biologically based heteropolymers. To accomplish this, proteins and peptides were chosen as template scaffolds. By using peptide biopolymers as templates, the cellular machinery of living cells can be utilized to generate the templates. In addition, the well-established techniques of directed evolution21 can be used to evolve the structure and function of the artificial, temp lated polymers via manipulation of the templates at the DNA level. There has also been extensive work done on the designing and controlling the conformational structure of peptides and proteins based upon their primary sequence, making them attractive for templates designed rationally. 22 ' 23 An additional benefit inherent to a peptide scaffold lies in the versatility of the side chains which have variable functionality and are separate from the linear polymer. The side chains provide a useful means of presenting multiple binding sites without manipulation of the amide backbone. To synthesize completely artificial heteropolymers with sequences as welldefined as those found in nature, an alternative set of catalysts and molecular recognition motifs had to be identified and developed. Given the previous successes in templated 4 syntheses and its convenience for ---- 0 polymerizations, olefin metathesis was chosen RCM ------- as the polymerization chemistry for this project. Both ROMP and acyclic diene methathesis (AD MET) - two methods of metathesis polymerization (Fig. 1.3) - are sensitive to the n ~ AD MET ~ ROMP n Figure 1.3: Types of metathesis chemistry. concentration of olefin in solution, and will not RCM = Ring ~losing Metathesis ROMP= Ring Qpening Metathesis folymerization ADMET = A.cyclic ]2.iene Metathesis proceed below a critical concentration that is specific for each olefin. For example, the critical concentration at room temperature for cyclopentene is 0.9M, while for cycloheptene it is 0.75M. Thus, if a templated reaction is run below the critical concentration, only monomer coordinated to the template will achieve a local concentration above the critical value. In this way, selective templatedirected polymerization will be achieved. To transfer information from the template to the artificial daughter polymer, selective interactions between the template and its binding partners (i.e., the metathesis monomers) must be designed. This recognition may be based on interactions such as hydrogen bonding, aromatic stacking, electrostatics, or metal-ligand interactions. Utilization of sets of recognition motifs that are orthogonal interactions to one another will enable the generation of ordered heteropolymers, with properties not available via simple homopolymers or random heteropolymers. Particularly relevant to establishing a template-directed polymerization system is the ability to use thermodynamic control to direct the formation of a well-defined molecular compound during a reaction which, if it were performed under kinetic control 5 would afford a myriad of products. Them1odynamic control is critical to supramolecular24 and dynamic covalent chemistry (DCC), 25 -27 both of which rely upon reversible non-covalent and covalent bond making and breaking processes as part of key proof-reading and error checking mechanisms. Thus, the success of a thermodynamically controlled protocol depends at the outset on the formation of complexes that are extremely stable prior to polymerization. This criterion must be taken into account when designing a recognition motif. In this research program, two recognition motifs are investigated: the [2]pseudorotaxane formed between secondary dialkylammonium ions and crown-ethers (Chapter 2), and the boronate ester formed between diols and boronic acids (Chapter 3). 6 References 1. Cantrill, S. J.; Pease, A. R.; Stoddart, J. F., J. Chem. Soc., Dalton Trans., 2000, 37153734. 2. Hoss, R.; Vogtle, F., Angew. Chem. Int. Ed., 1994, 33, 375-384. 3. Anderson, S.; Anderson, H. L.; Sanders, J. K. M., Acc. Chem. Res., 1993, 26, 469-475. 4. Busch, D. H.; Stephenson, N. A., Coord. Chem. Rev., 1990, 100, 119-154. 5. Krejchi, M. T.; Atkins, E. D. T.; Waddon, A. J.; Fournier, M. J.; Mason, T. L.; Tirrell, D. A., Science, 1994, 265, 1427-1432. 6. Zhang, G. H.; Fournier, M. J.; Mason, T. L.; Tirrell, D. A., Macromolecules, 1992, 25, 3601-3603. 7. Cobb, A. J. A., Org. Biomol. Chem., 2007, 5, 3260-3275. 8. Dougherty, M. J.; Kothakota, S.; Mason, T. L.; Tirrell, D. A.; Fournier, M. J., Macromolecules, 1993, 26, 1779-1781. 9. Connor, R. E.; Tirrell, D. A., Polymer Rev., 2007, 47, 9-28. 10. Yoo, T. H.; Link, A. J.; Tirrell, D. A., Proc. Nat. Acad. Sci US.A., 2007, 104, 1388713890. 11. Montclare, J. K.; Tirrell, D. A., Angew. Chem. Int. Ed., 2006, 45, 4518-4521. 12. Rosenbaum, D. M.; Liu, D.R., J Am. Chem. Soc., 2003, 125, 13924-13925. 13. Calderone, C. T.; Liu, D.R., Curr. Opin. Chem. Biol., 2004, 8, 645-653. 14. South, C.R.; Weck, M., Macromolecules, 2007, 40, 1386-1394. 15. Grubbs, R.H., Ed., Handbook of Metathesis. Wiley-VCR: Weinheim, 2003. 16. Trnka, T. M.; Grubbs, R. H., Acc. Chem. Re., 2001, 34, 18-29. 17. Badjic, J. D.; Nelson, A.; Cantrill, S. J.; Turnbull, W. B.; Stoddart, J. F., Acc. Chem. Res., 2005, 38, 723-732. 18. Guidry, E. N.; Cantrill, S. J.; Stoddart, J. F.; Grubbs, R.H., Org. Lett., 2005, 7, 21292132. 19. Kilbinger, A. F. M.; Cantrill, S. J.; Waltman, A. W.; Day, M. W.; Grubbs, R.H., Angew. Chem. Int. Ed., 2003, 42, 3281-3285. 20. Yang, X. W.; Gong, B., Angew. Chem. Int. Ed., 2005, 44, 1352-1356. 21. Xu, H.F.; Zhang, X. E.; Zhang, Y. M.; Cass, A. E.G., Prog. Biochem. Biophys., 2002, 29, 518-522. 22. Toniolo, C.; Crisma, M.; Formaggio, F.; Peggion, C.; Broxtern1an, Q.; Kaptein, B., J Inclusion Phenom. Macrocyc. Chem., 2005, 51, 121-136. 23. Bolin, K. A.; Millhauser, G. L., Acc. Chem. Res., 1999, 32, 1027-1033. 24. Lehn, J.M., Supramo. Chem., Wiley-VCR: Weinheim, 1996. 25. Fuchs, B.; Nelson, A.; Star, A.; Stoddart, J. F.; Vidal, S. B., Angew. Chem. Int. Ed., 2003,42,4220-4224. 26. Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K. M.; Stoddart, J. F., Angew. Chem. Int. Ed., 2002, 41, 898-952. 27. Lehn, J.M., Chem. Eur. J, 1999, 5, 2455-2463. 7 Chapter 2 A Pseudorotaxane Recognition Motif 8 I. Introduction The Pseudorotaxane Recognition Motif: A pseudorotaxane was investigated as a recognition motif in the development of template-directed metathesis polymerization system. A pseudorotaxane is composed of a secondary ammonium ion threaded through an appropriately sized crown-ether molecule stabilized by [N+-H···O] hydrogen bonds and [CH···O] interactions (Fig. 2.1). 1 This motif is useful in a templating system because it is reversible allowing easy removal of the daughter template and its strength is tunable by changes in solvent, Figure 2.1: The structure of a r2lpseudorotaxane. temperature or pH. In addition, the secondary ammonium center can be directly generated on a peptide template by the alkylation and protonation of a natural lysine residue. A rotaxane - derived from the Latin rota meaning wheel and axis meaning axle is composed of a macrocylic compound encircling a rod with two bulky stopper groups at each end. A compound with that composition minus the stoppers is termed a pseudorotaxane and exists in equilibrium between threaded and separated components. The identity of the alkyl groups attached to the ammonium center influences the strength of the pseudorotaxane association. 2 The association constant in deuterated acetonitrile of a [2]pseudorotaxane composed of dibenzo[24]-crown-8 (DB24C8) with dibenzyl ammonium (Ka of 420 L mor 1) is about five times higher than that of a [2]pseudorotaxane containing dibutyl ammonium (Ka of 70 L mor 1). This observed difference has been attributed to the increased acidity of the benzylic methylene protons, 9 which enhances the pseudorotaxane hydrogen-bonding interactions, and to the decreased flexibility of the dibenzyl ammonium moiety, which gives it better preorganization and therefore enhanced binding. The association constant for pseudorotaxanes is also influenced by solvent. The Gutmann donor number2 - a semiquantitative measure of the ability of a solvent to participate in non-covalent bonding - varies inversely with the strength of the association in each solvent (Table 2.1). 3 For example, the association constant for the [2]pseudorotaxane formed between a dibenzyl ammonium ion and DB24C8 is highest in the solvent with the lowest Gutmann donor number, chloroform, and is zero in the solvent with the highest donor number, DMSO. Notably, in chloroform, the most effective solvent for promoting hydrogen-bonding complex formation, the association constants for the [2]pseudorotaxanes containing dibutyl and dibenzyl ammonium ions are both in the 104 range. Solvent Gutmann donor no. Ka (L mor 1) ~-DMSO 29.8 0 ~-acetone 17.0 360 CD3CN 14.1 420 CDCb 4.0 27000 Table 2.1: The effect of solvent on the association constant for a complex ofDB24C8 and dibenzyl ammonium ion. (Adapted from Ashton, P.R. , et. al. Chem, Eur. J., 1996, 2(6), 709-728) The Templates: In the investigation of the pseudorotaxane as a recognition motif for the templatedirected metathesis system, a number of different templates were evaluated (Table 2.2). Within this report, templates with two binding sites, also known as bis-ammonium templates, are denoted as b-Am, while the tris-ammonium templates, with three binding 10 sites, are denoted as t-Am. The subscript refers to the style of template. Table 2.2: The series of templates studied bearing ammonium ion recognition sites for DB24C8 binding. Type of Template Chemical Formula ID Peptide, 11-mer Ac-VALK\Bn)VALK+(Bn)VAL-NH 2 b-Amp11 Peptide, 15-mer Ac-VALK\Bn)VALK+(Bn)VALK\Bn)VAL-NH 2 t-Amp1 5 Peptide, 20-mer Ac-VALVALK+(Bn)VAL V ALK+(Bn)VAL VAL-NH 2 b-AmP2o Peptide, 18-mer Ac-LALLALLK+(Bn)LALAK+(Bn)LALLAL-NH 2 b-Amp 18 Peptide, 3-mer Ac-K+(Bn)AK+(Bn)-NH2 b-Amp3 Peptide, 5-mer Ac-K+(Bn)AK+(Bn)AK+(Bn )-NH 2 t-Amp5 Alkane, butane-linked Bn-WH2)-(CH2)4-WH2)-Bn b-AmA4 Alkane, hexane-linked Bn-WH2)-(CH2)6-(N~2)-Bn b-AmA6 Alkane, octane-linked Bn-WH2)-(CH2)s-(N+H2)-Bn b-AmAs Alkane, decane-linked Bn-WH2)-(CH2)1 o-WH2)-Bn b-AmAIO Alkane, dodecane-linked Bn-WH2)-(CH2)1rWH2)-Bn b-AmA12 Linear, benzyl-linked Bn-WH2)-CHr Ph-CHrWH2)-Bn b-AmB Mono-ammonium, model TBDPSO-(CH2)4-N+HrBn M Four different iterations of peptide templates were tested, the designs of which are described in more detail below. In addition, two different types of linear templates were used as standards against which the peptides were compared. Template b-Am8 was known from previous macromolecular studies to bind two DB24C8 molecules (Fig. 2.2). 1 The series of alkanelinked linear bis-templates, b-AmA, were designed to study the influences of template flexibility and binding site spacing on template-directed metathesis. Figure 2.2: The solid state structure of a [3 ]pseudorotaxane comprised of temp late b-Am 8 and two DB24C8. (Adapted from Reference 1) 11 Design of 1st Generation Peptide Templates: A helical structure was selected for the first peptide templates because it can be formed by short peptides and is highly ordered, allowing for controlled presentation of the crown-ether binding sites. Also, helices are favored by aprotic, non-polar solvents such as those suitable for metathesis and pseudorotaxane formation. B: Boc-Val-Ser-Leu-Aib-Val-Ser-Leu-OMe A: Boc~Val-Ala-LeujAib-Val-Ala-Leu-OMe Figure 2.3: Karie's helical heptapeptide A and Blackwell's helical peptide st B, plus 1 generation peptide template b-Amr 11 . The design of the 1st generation templates, b-Amp 11 and t-Amp 15 , was based on the Val-Ala-Leu repeat of the helical, hydrophobic heptapeptide (A in Fig. 2.3) of Karle 4 et. al. . Peptide A had also inspired the heptapeptide (B), developed by Blackwell et. al.,5 which was shown to adopt an a-helical confom1ation in the solid state and a 310 -helical conformation in chloroform. In addition, peptide B was used in ring closing metathesis reactions, after modification of the serine residues with allyl ethers. For the 1st generation peptides, the valine-alanine-leucine repeat was utilized to promote helix formation. The a-aminoisobutyric acid residue (Aib) is known to stabilize helical conformations. However, Aib residues were not incorporated into the peptide templates because their incorporation by standard biosynthetic expression techniques is not possible. Instead, to adapt the previous sequences to the requirements of a template for pseudorotaxane complex formation, ammonium binding sites were installed via lysine amino acids at 12 residues i and i+4 within the peptide sequence. This spacing positions two residues on the same side of the helix, approximately one tum away from each other. Design of 2nd Generation Peptide Templates: Two longer peptide templates were designed to provide space between the binding sites and greater helical stability compared to the 1st generation peptides. The template b-Amp2o used the same valine-alaine-leucine repeat as b-Amp 11 and t-Amp 15 . However, this peptide is extended by an additional three V-A-L repeats, one on each end and one in between the lysine residues. By increasing the peptide length by nine amino acids, the helical conformation of this template was hypothesized to have increased stability compared to the b-Amp 11 . In addition, the spacing between the ammonium centers is i and i+ 7, which places the sites on the same side of an a-helix two helical turns apart. This makes the binding sites approximately twice the distance of those with i and i+4 spacing in b-Amp 11 (approximately 11.4 angstroms vs. 5.7 angstroms). The other 2nd generation peptide was designed with the sequence AcLALLALKLALALKLALLAL-NH2 to incorporate the more helix-inducing leucine amino acid in the place ofvaline residues from the V-A-L sequence. The spacing between the lysine residue binding sites was i and i+6, which positions the sites on the same side of a 3 10-helix (approximately 12.6 angstroms apart). Unfortunately, a synthesis error yielded a peptide missing one leucine residue between the lysine binding sites and gave the template b-Amp 18 instead. This template has an i and i+5 spacing between the lysine residues, which in both a- and 3 10 -helices places the binding sites on opposite sides of the helix. 13 Design of 3 rd Generation Templates: The last set of peptide templates, b-Amp3 and t-Amps, was designed to simplify the template motif. These short peptides, consisting of alternating lysine residues - the crown-ether binding sites - and alanine-residue spacers, were not intended to adopt an ordered structure. Rather, they were designed to provide a simple scaffold to hold bound crown-ether molecules in close proximity. II. Results and Discussion Synthesis of b-Amp 11 and t-Amp1s Templates: The peptide starting materials for the 1st generation peptide templates, b-Amp 11 and t-Amp 15 , were synthesized by the Biopolymer Synthesis and Analysis Facility at the California Institute of Technology and were used without further purification. To provide binding sites for the DB24C8 monomers, the free lysine residues were modified with benzyl groups and protonated. Scheme 2.1: Synthesis of Pll. a) benzaldehyde, 0.IM NaOH (aq), TFE; b) NaBH4, 93% overall. 14 Complete mono-benzyl protection of both lysine residues of the peptide starting material for Pll was achieved under reductive amination conditions (Scheme 2.1). First, a Schiff base was formed between the free amines of the lysine residues and benzaldehyde in basic solution. Then the intermediate Schiff base was reduced with sodium borohydride to give the desired benzyl-amine product. For Pll, it was necessary to add benzaldehyde batchwise in three rounds and to employ a time delay before addition of the reducing agent to achieve complete benzyl modification. Attempts to purify Pl 1 by normal phase column chromatography and prep TLC caused degradation of the peptide, so the crude product was used without further purification. The secondary nitrogen centers of the benzyl-lysine residues of Pll had to be converted to cationic ammonium centers to enable pseudorotaxane formation. The crown-ether ammonium interaction is maximized by using ammonium salts for which there is weak ion pairing between the ammonium ion and the counterion. 2 The standard counterion used to meet this criterion is hexafluorophosphate (PF 6} As an additional benefit, the PF 6 anion often enhances the organic solubility of ammonium salts. To protonate the peptide and generate the PF 6-salt, Pll was mixed with HPF 6 . ESI-MS confirmed that the reaction had quantitatively generated the doubly charged peptide Pl12+. However, the PF 6-salt of Pl12+ was not soluble in chloroform or dichloromethane without the addition of TFE. TFE reduces the pseudorotaxane interaction, making it unsuitable for this templating system. The large, hydrophobic counterion BArF (Fig. 2.4) was tested for its ability to improve the solubility of the protonated peptide. To generate the BArF-salt of Pll a HBArF-complex was generated in situ. 6 This complex, [H(Et20ht[BArFL was 15 successfully made by mixing an anhydrous solution of HCl in ether with NaBArF at reduced temperature. The NaCl by-product was removed by filtration and the ether was removed to yield the complex as a white Figure 2.4: The BArF counterion. solid. At reduced temperature, the [H(Et20)2t[BArFr solid was redissolved in anhydrous ether and was added to a solution of Pll. The reaction yielded a product with an accurate 1:2 peptide to BArF ion ratio. Generation of Pll2+ was confirmed by ESI-MS, which showed both the doubly charged peptide and the BArF counterion. The product, b-Amp 11 , was found to be readily soluble in chloroform and dichloromethane, and was used for complexation and metathesis studies. • The tris-ammonium template, t-Amp 15 , was synthesized using the same procedures described for b-Amp 11 . Synthesis of b-AmPls and b-Amp20 Templates: The peptide starting materials for two longer templates were ordered from the GenScript Cooperation in crude form. The peptide for P18 was synthesized by mistake instead of the desired sequence of P19, but given the extended time of synthesizing the peptides, complexation and metathesis studies were done with P18 before it was decided whether or not to order the correct peptide for P19. Using the same reductive amination procedure as with Pll, the lysine residues of P2O and P18 were benzyl protected. Because of the difficulties experienced while trying to purify Pll, purification ofbenzylated products P2O and P18 was not attempted. Protonation and coordination with the BArF anions was accomplished using the HBArFcomplex. 16 Synthesis of b-Amp3 and t-Amp 5 Template: The peptide for template b-Amp 3 was synthesized and purified to 97% purity by GenScript. After the initial experiments, further batches of b-Amp 3 and t-Amp 5 were synthesized by standard Fmoc solid phase peptide synthesis methods. For these templates, a slightly modified procedure was used for the benzyl protected of the lysine residues via reductive amination. As before, a Schiff base was formed between the amine residues of P3 or PS and benzaldehyde. However, in this procedure the imine material was isolated and dried to remove excess benzaldehyde. The reaction material was redissolved and the imine was reduced. The crude reaction was purified by reverse phase HPLC and pure p32+ and p52+ were isolated as TF A-salts. To enhance the solubility of p32+ and p52+ in organic solvents, the TF A counterions were changed to BArF ions. To accomplish this, pure peptide was dissolved in a minimal amount of water, and dichloromethane was added to make a biphasic solution. NaBArF was added (2 equiv. for P32+, and 3 equiv. for and PS2+), and after vigorous stirring, the templates were isolated from the organic layer. 1H NMR verified the correct 1:2 or 1:3 peptide to ion ratio. Synthesis of b-AmA Series of Templates: To study both the influence on template-promoted metathesis of template rigidity and the influence of spacing between the crown-ether binding sites, a series of alkanelinked bisammonium templates of varying lengths was synthesized (Fig. 2.5). Briefly, each commercially available alkane diamine was benzyl protected under standard reductive amination conditions and purified by column chromatography. The purified dibenzyl-amine products were protonated with HCl, and the resulting HCl-salt 17 precipitates were isolated. Then, the chloride counterions were replaced by BArF counterions in a biphasic solution of dichloromethane and water. H2+ N b-AmA4 2+ I ~ N H2+ :.v Q ~ H2+ N b-AmA6 2+ I ~N:.v H2+ ~ H~ B N~N:.vb-AmAs H2+ ~ I 2+ H2+ N~N:.vb-AmAJO H2+ ~ I H~ N B N:.vb-AmA12 H2+ ~ I Figure 2.5: Series of flexible, alkyl-linked bis-ammonium templates, b-Am,H through b-AmA 12 . All were synthesized as BArF-salts. Synthesis of b-AmB: 7 A rigid, linear bis-ammonium template inspired by a previously studied [3]pseudorotaxane system2 was synthesized as shown in Scheme 2.2. Under Dean Stark conditions, p-xylylenediamine and two equivalents of benzylaldehyde were reacted to give the diimine intermediate. Reduction with sodium borohydride, followed by protonation of the amines with hydrochloric acid and counterion exchange with NaBArF yielded the organic soluble bis-ammonium template. Scheme 2.2: Synthesis of template b-Am 8 . a) toluene, -H 2O (Dean-Stark), 84%; b) NaBH4 , MeOH, 0 °C ➔ rt, 89%; c) HCl, MeOH, then NaBArF, DCM, 68%. 18 Model Synthesis: A model compound (M) was designed to simplify the evaluation of the pseudorotaxane interaction within the peptide template system. To suit this purpose M was designed to mimic a benzyl-modified lysine residue. Mis composed of two parts, a benzyl-protected butylamine moiety that mimics the benzyl-lysine binding site and a bulky t-butyl diphenyl silyl (TBDPS)-protected alcohol that stands in for the peptide chain. To synthesize M (Scheme 2.3), 4-aminobutanol was TBDPS-protected using TBDPS-Cl and imidazole in DMF. The TBDPS group is not only bulky enough to block DB24C8 dissociation,8 but is also stable to both the basic conditions of the benzylation reaction (pH< 10) and the acidic conditions of the protonation step (pH> 2). Following silyl protection of the alcohol, the amine functionality was mono-benzyl protected using reductive amination conditions. The purified material was protonated in a solution of methanol acidified to pH 2.5 with an aqueous solution of 6M hydrochloric acid. Finally, NaBArF was used to replace the chloride counterion in a biphasic solution of water and chloroform. Protonated M was isolated from the organic layer. ~R NH 2 HO~ a ~ TBDPSO~ NH 2 b ~~ ~ TBDPSO~ R H2 ~ _,__lJ~N 8 BArF M,R=H M-pF, R=F M-pOMe, R=OMe M-pCN, R=CN Scheme 2.3: Synthesis ofM. a) TBDPSCI, imidazole, D:MF; b) benzaldehyde, 0.IM NaOH (aq), MeOH, then NaBH4 , 50% over three steps; c) 6M HCI (aq), MeOH, then NaBArF, H2O, DCM, 75%. 19 Modification of the benzyl group at the para-position with an electronwithdrawing functionality was hypothesized to promote ammonium interaction with the crown-ether. To test this, M compounds with para-modified benzyl functionalities were synthesized using the same procedure as for M but with modified benzyaldehyde reagents in the reductive amination step. The functionalities of the three para-modified M compounds synthesized - fluorine, methoxy and cyano - were chosen because of their range of electron-withdrawing character and because the necessary benzaldehyde reagents were commercially available. DB24C8-Monomer Syntheses: An alcohol modified crown-ether derivative (DB24C8-0H) was synthesized and further functionalized with various olefinic moieties (Scheme 2.4). To synthesize DB24C8-0H, one equivalent of dihydroxybenzaldehyde was treated with two equivalents of 2-[2-(2-chloroethoxy)ethoxy]ethanol. The hydroxyl groups of the diol were tosyl-protected. Macrocyclization with catechol, followed by reduction of the aldehyde with lithium aluminum hydride yielded DB24C8-0H. Subsequent EDC coupling to various olefinic acids yielded the ester-functionalized derivatives, DB24C8- CP, -Sen and -diene. Alternatively, coupling ofDB24C8-0H to 5-bromopentene yielded the ether-functionalized derivative DB24C8-e5en. The diene functionality ofDB24C8-diene was designed to minimize intramolecular metathesis reactions between the two olefins. 9 The acid-containing compound used to couple to DB24C8-0H was synthesized by alkylation of methyl 3,5dihydroxybenzoate with 1-bromobutene and potassium carbonate. Saponification using potassium hydroxide in water and ethanol yielded the acid compound. 20 a ______.. o nnl\ 0 0 OTs 0 0 OTs H l_J\__j\_./ oX) I t\o"o~ o HO f DB24C8-0H .,q 0 l_Jo\__Jo\_./ !' nt111 ("'~B24C8-e5en ~ 0 l_Jo\__jo\_./ R= X) DB24C8-CP DB24C8-diene ~ ~~~ DB24C8-Sen ~ ~ ~' Scheme 2.4: Synthesis ofDB24C8-olefin derivatives. a) K 2CO 3 , LiBr, DMF; b) pTsCl, TEA, DCM; c) catechol, Cs 2CO 3, DMF; d) LiAlH4 , THF; e) EDC, DMAP, DCM;j) 5-bromopentene, NaH, THF, 63%. Peptide Conformations: CD spectra of Pl 1, P15, P18 and P20 were taken in trifluoroethanol. TFE is known to induce peptide conformations similar to those induced by chloroform and dichloromethane, 5 which are the ideal solvents for metathesis chemistry and pseudorotaxane association, but are not suitable for CD spectroscopy. Wavelength scans from 260 nm to 190 nm were taken at room temperature for each peptide. The CD spectra indicate that all the peptides adopt right-handed 310-helical conformations in TFE. The 310-helix is more tightly wound than an a-helix and has 1~4 main chain hydrogen bonding, as compared to the 1~5 pattern found in an a-helix. 10 CD spectra of helical peptides are distinguished by two minima: one at 206 nm, which 21 Peptide R arises from a rr-rr* transition, and another at 222 Pll 0.55 nm, which arises from a n-rr* transition. A 3 10 - P18 0.64 P20 0.66 Table 2.3: R values for peptide templates. helix gives rise to a weaker n-rr* than rr-rr* transition. This is reflected in a ratio R - defined as [0]222/[0ho6 - of approximately 0.4. For an a-helix, the two transitions are approximately equal, giving an R close to 1.0. The R values for all the peptides are between 0.55 and 0.66, indicating that the structures of the peptides are mainly 3 10 -helix, with varying amounts of contaminating a-helix conformation (Table 2.3). The helical structure of Pll has the most 3 10 -helical character, while P20 has the least. The CD spectra of the bis-ammonium peptides were all found to be independent of concentration, which demonstrates that their structures are monomeric. However, the CD spectra of the tris-anunonium Pl 5 showed concentration dependence, indicating the formation of higher order aggregates in those samples (Fig. 2.6). 11 250 260 R =0 .59 -50 ------------ ~-------------~ Wavelength (nm) wavelength (nm) Figure 2.6: Concentration dependent CD spectra of P15; • 10µM, o 30µM, A 100µM. Figure 2.7: Influence of En-modification on P11 helicity; • peptide starting material for Pl 1 with free lysines, o P11. 22 As judged by CD analysis in TFE, upon benzyl modification of P 11 the strength of the helix formed by the peptide decreases. The starting material peptide for Pl 1 has more than twice the helical character of the benzyl-modified template Pll (Fig. 2.7). It is unclear why benzyl-protection of the lysine residues decreases the peptide helicity. The CD spectra of Pll, P18 45 ~ - - - - - - - - - - - - - - - ~ and P20 all have similar minima at 206 nm and 220 nm (Fig. 2.8). Thus, in opposition to the design principle of the P18 and P20 250 26 templates, the longer sequences did not endow them with increased helical character compared to Pl 1. -25 - ' - - - - - - - - - - - - - - - - - - - - - - _ _ J Wavelength (nm) Figure 2.8: CD analysis of 1st and 2nd generation peptides; □ Pl 1, • P18, • P20. A melting curve of P18 from 25 °C to 72 °C showed a gradual decrease in helicity, but no distinct phase transition. A similar result was seen with Pl 1. This type of melting behavior is indicative of a peptide existing in multiple conformations. NMR Analysis of the Conformation of b-Ampu: A series of 2D NMR spectra - tntocsy, and tnnoesy - were taken to explore the conformation of one of the peptide templates, fluorinated b-Amp 11 ( dubbed b-AmFPu), in deuterated dichloromethane. Tntocsy and tnnoesy data were acquired for three different samples of b-AmFp 11 . All eleven spin systems were identifiable in all three tntocsy spectra, although each sample gave different results (Fig. 2.9). Extensive heating and sonication of one of the samples changed its spectrum slightly, but not enough so that it matched either of the other two samples. This disparity indicates that each batch of b- 23 AmFPu folded differently, even though the preparation procedures were identical. A, V \/-Ac 9. 0 8,5 8,0 7.5 7 .o 6.5 6-0 9.0 8.5 8.0 1'2 7.5 7.0 6.5 6.0 (ppa) Figure 2.9: Differences in the tntocsy 1H NMR data for two samples ofb-AmFP 11 in CD 2 Cb, In tnnoesy NMR analysis, helical peptides are characterized by strong crosspeaks from sequential NN, aN(i, i+3) and ap(i, i+ 3) interactions (Fig. 2.10). 12 The tnnoesy data from one b-AmFPu sample showed strong crosspeaks for all the Figure 2.10: Crosspeaks seen in 2D NMR spectra of helical peptides. The dotted-line encloses an isolated spin system seen as through-bond coupling in TOCSY and COSY. The solid arrows denote through-space coupling seen in NOESY and ROESY analysis. residues except those at the C-terminus. This data suggested that b-AmFPu had helical character in dichloromethane, but that the C-terminus was less structured than the rest of the peptide. The tnnoesy spectra of the other two samples of b-AmFPu did not show any sequential NN, aN(i, i+ 3) and ap(i, i+ 3) crosspeaks. 24 Since the tnnoesy experiments failed to yield data, wroesy experiments were run on those two samples. Strong crosspeaks were seen in the wroesy spectra, analysis of which suggested that these samples also had helical character. As before, the C-terminus was less structured than the rest of the peptide in these samples of b-AmFp 11 . Tnnoesy experiments are useful for molecules that are either less than 1 kD or for larger molecules, while wroesy experiments are appropriate for molecules with a molecular weight around 2 kD. The molecular weight ofFPll2+ is 1.4 kD and that ofbAmFpn is 3.1 kD. The BArF counterions are thought to be loosely associated with the peptide template, such that it should be tumbling in solution with an effective molecular weight around 1.4 kD. Therefore, the fact that the first sample ofb-AmFpn showed good NOESY peaks suggests that it was forming some sort of self-aggregate which boosted its molecular weight into the NOESY range. In contrast, the second samples had ROESY, and not NOESY, peaks. This suggests that these samples were monomeric in solution. The conformation of b-AmFPn in dichloromethane is more complicated than the CD analysis in TFE would indicate. Pseudorotaxane Formation with M: The amount ofM complexed with DB24C8-CP was evaluated by 1H NMR in a variety of solvents, 13 at different temperatures, and with different ratios of M to crownether. Solvent had the largest influence on the amount of complex formed (Table 2.4), with the low dielectric solvents, chloroform and dichloromethane, promoting the highest amount of complexation - around 50% - and the high dielectric solvents, especially acetone and acetonitrile, allowing the least amount of complex to form. Surprisingly, 25 neither temperature (Table 2.5) nor the amount of DB24C8-CP added (Table 2.6) significantly influenced the amount of complex formed. It was hypothesized that para-benzyl modification with an electron-withdrawing of an ammonium entitity would increase its association in a pseudorotaxane complex. In agreement with this hypothesis, both M-pOMe and M-pF exhibited increased Solvent Dielectric Constant Ammonium(%) Complexed Ka (L mor 1) Chloroform 4.8 51 212 Dichloromethane 9.1 53 240 Nitromethane 39.4 39 105 Acetonitrile 37.5 25 44 Acetone 20.6 23 39 Table 2.4: Influence of solvent on complex ofM to DB24C8-CP at r.t. in a solution an equimolar solution. I 0mM solutions were analyzed by 600 MHz 1H NMR. Temperature(° C) Ammonium(%) Complexed Ka (L mor 1) -10.18 52 226 -0.17 51 212 20.15 51 212 30.13 51 212 40.06 51 212 Table 2.5: Influence of temperature on an equimolar complex ofM to DB24C8-CP ind-chloroform. Equivalents of Ammonium(%) DB24C8-CP Complexed 1.0 51 1.5 57 2.0 48 2.5 48 3.0 45 Table 2.6: Influence of increasing amounts of DB24C8-CP on complex with M (I 0mM) in d-chloroform at r.t. 26 complexation compared to the unmodified M. This trend correlated with the respective electron-withdrawing capabilities of the para-modifications (Table 2. 7). The fluorine modification confers strong inductive withdrawal, and accordingly M-pF exhibited the highest association constant. (As an aside, the fluorine modification also serves as a useful tag for NMR studies.) The methoxy modification also provides inductive withdrawal, though a slightly less well, but has an additional, weak electron-donating capabilities through hyperconjugation that reduces its electron-withdrawing property. Fittingly, M-pOMe shows less of an enhancement over M. The NMR peaks of free MpCN could not be resolved from those of the pseudorotaxane complex, making the data challenging to interpret. Regardless, no noticeable improvement in association was observed for M-pCN. Ammonium Compound (%) Complexed 1 Ka (L mor ) Krel M 51 218 1.0 M-pCN ~39 ~105 0.5 M-pOMe 63 460 2.1 M-pF 88 6.11 X 103 28.0 Table 2.7: Influence ofpara-modfication of the benzyl group on association constant. lOmM solutions in d-chloroform were analyzed by 600 MHz 1H NMR. Mass Spectrometry Analysis of Pseudorotaxane Complex: ESI-MS MS was used to analyze a solution containing the PF6 salt of P11 2+ and DB24C8-CP. The [2]pseudorotaxane, [Pll:::)DB24C8-CP], was detected, as was a small amount of the desired [3]pseudorotaxane, [Pll:::)2(DB24C8-CP)] (Fig. 2.11). The weak signal of pseudorotaxanes compared to that of free DB24C8-CP and Pl 1 is likely due to the lack of soft ionization settings on the ESI mass spectrometer. 14 The spray voltage setting had to be raised drastically from the ideal setting for any signal to be seen, 27 thereby causing fragmentation of the complex within the instrument. 100-::. 595.4 [DB24C8-CP•Na] • 60 ~ 30-=: ~[PlW 20 •• 673.5 o~ 684 .6 783.1 •I 600 959.5 800 1095.4 1000 1264.8 1345.7 1435.9 1200 mlz 1663.5 1733.4 1800.5 1939.5 1400 1600 1800 2000 960.2 100-:; [Pll::::>DB24C8-CPF- - - ) I 959.6 BO -: ! 960.7 ·'1 60-: I [PlbDB24C8-CP•NaF· l 971.1 I 971.7 20 -: I 91 10.:: 901 9 - j• 0 915 8 -~ 1 ~- 958 9 925 7 931 1 938 4 941 6 955 5 ' (~"{'I' \'\ J',~(1 ,J,i1,4f,.,,r,Jyt,\J>...,,..i,"''i, •I, ,11).~~ _.,.,;}~.. • ! .'HtJh;ll\~;A> 11'Mv- ·.i·1 N ~ 910 930 920 940 950 m lz 1245 6 • 100-=: i' I 967 .1 975.2 984.0 991.6 997.4 r1r\V")·'!-.\• l\.;J.,1.,., \ "'.'t\ 1.i1.,~'.,1.::JJ....v)//~.,~~'"-.-'~(1J' 960 970 980 990 1000 ~ [Plb2(DB24C8-CP)F· 1246.9 124 7.3 1 1269.1 1210 1220 1230 1240 1250 mlz 1260 1270 1280 1290 1300 Figure 2.11: ESI-MS of the PF6 salt ofb-Amr 11 and DB24C8-CP. 28 NMR Analysis of the Pseudorotaxane Complexes: The enhanced solubility of the BArF salt of Pll, as compared to the PF 6 salt, allowed for 1H NMR analysis of the pseudorotaxane formed between Pll and DB24C8- 5en. Formation of the complex decreases the inductive withdrawal of the crown-ether oxygen atoms, which is predicted to cause a small upfield shift in the DB24C8 methylene peaks.2 Formation of the complex also increases the deshielding effect of the oxygen atoms, which is predicted to cause a significant downfield shift in the Pl 1 peaks. Previous studies with similar pseudorotaxane systems have demonstrated approximately 0.05 ppm upfield shifts for crown-ether protons and 0.5 to 1.0 ppm downfield shifts for the ammonium and benzylic methylene protons. 3 The NMR spectra of free b-Amp 11 and DB24C8-5en were compared to the spectrum of a solution of the template with two equivalents of the crown-ether in d2 dichloromethane. Surprisingly, there were no obvious shifts of the crown-ether peaks in the spectrum of the complex compared to that of free DB24C8-5en. As expected, many of the template peaks were significantly shifted downfield. In particular, the ammonium protons - appearing as two peaks around 6.5 ppm in the spectrum ofb-Amp 11 -were shifted downfield in the dual component sample such that they overlapped with one of the DB24C8-5en aromatic peaks at 7.0 ppm. In addition, the benzylic methylene protons of Pl 1 were shifted downfield so that they partially overlapped with a DB24C8-5en at 5.02 ppm. Because the shifted peptide peaks both overlapped with unshifted crown-ether peaks, the increase in integration values of the DB24C8-5en peaks were used to locate the shifted peptide protons in the spectrum of the complex. The b-Amp11 peaks were shifted completely, indicating total generation of the [3]-pseudorotaxane. 29 No complex was observed between a) b-Amp 11 and DB24C8-5en in acetonitrile, BArFN ..... cri nitromethane or nitromethane/chloroform 0 ..... I because the peptide precipitated in each ...________,,..... .. ' ,~- ~ - 70 .l. ' ~~ - t.. O - 80 - fiO -10 0 -120 b) case. In the spectra collected in these co ~ 0 ..... ..... I solvents, only peaks arising from the crown molecule were observed. ·" 19 F NMR analysis ofb-AmFp11 with two equivalents ofDB24C8-5en also : ~· -~ - 60 ... -•o - 20 c) -8 0 -10a (j) I 0 ~ 0 ..... ..... ..... I I I l( ! \ . , , ---r-· ----r-.-----•- -2n 19 M-pF established that the fluorine LO cri CV') supported complete [3]-pseudorotaxane formation. Previous F NMR analysis of -120 -• a -8 0 - 1.i O ,.~ 19 Figure 2.12: F NMR spectra of a) M-pF, b) MpF with 1 equivalent of DB24C8-CP, and c) MpF with 0.5 equivalents of DB24C8-CP. The BArF counterion serves as an internal reference. resonance shifts upfield by approximately 2 ppm upon binding D B24C8-5en (Fig. 2.12). In the spectrum of [b- N I'- a) (j) co I'0 ..... LO CV') co 0 ..... ..... I ..... I I AmFp11::::,2(DB24C8-5en)] , the two fluorine resonances are shift upfield by 1. 7 and 2.1 - 104 ppm, respectively, as compared to those of h) -106 -108 -110 -116 -1~ co ..... ..... ~ N N b-AmFp 11 (Fig. 2.13). The NMR peaks of the [3 ]-pseudorotaxane are broader than those of the free b-AmFp 11 , a difference that was also seen for the protons of the lysine Figure 2.13: 19F spectra of a) b-AmF Pll and b) bAmF rit with 2 equivalents ofDB24C8-5en. Excess 4F-benzylalcohol side product served as an internal reference. residues in the 1H spectra. 30 NMR Analysis of Conformation of Pseudorotaxanes: One of the samples of b-AmFPn from the 2D NMR conformation studies was used to investigate the affect of adding crown ether on the conformation of the peptide. Tntocsy, tnnoesy and wroesy spectra were taken of a d2 -dichloromethane solution containing b-AmFp 11 and one equivalent ofDB24C8-5en. No crosspeaks were observed in the wroesy spectrum, while some crosspeaks were observed in the tnnoesy experiment. Taken together, these data suggest that a higher molecular weight complex of [bAmFp11::::)DB24C8-5en] has formed. The NOE crosspeaks were fewer and weaker than those seen with free peptide, suggesting that forming the [3 ]-pseudorotaxane reduces the helical character of b-AmFp 11 . In the tntocsy spectrum of the complex, the amide protons of the peptide converged. This is another indication of a less ordered structure. There were approximately two conformations identified in the tocsy spectrum, neither of which matched the spectrum of free b-AmFPll or b-AmFPll mixed with two equivalents of crown-ether (see below). This suggests the formation of two [2]-pseudorotaxanes, one with each ammonium center bound by one DB24C8-5en. The identities of the spin systems were determined, but because the tnnoesy data was weak, the connectivity was not. Another equivalent ofDB24C8-5en was added to the sample and more 2D NMR data was gathered. Again, the wroesy failed to produce data. The NOE crosspeaks seen in the tnnoesy spectrum were similar in terms of intensity and number to those seen in the spectrum with one equivalent ofDB24C8-5en. In the tntocsy spectrum, the amide protons converged even further, suggesting a further loss of structure (Fig. 2.14). Like residues - those of the same amino acid but at different positions in the sequence - had 31 amide and alpha protons with very similar chemical shifts. This is in contrast to those of free b-AmFpu, which showed shifts that varied depending on the position of the residue within the peptide. In the spectrum with two equivalents of crown-ether, only one set of protons was observed for each amino acid and two for each of the ammonium spin systems. This indicated the presence of only one species, presumably the [3]pseudorotaxane. Given the tnnoesy data and the shifts of the amide protons, the complex likely adopts a random coil conformation. b) VAAAN'NL L V 919'K 10I 10'K/L 8. 0 7.5 7.0 rl (ppm), , 9.5 9.0 8.5 I 8.0 7.5 7.0 6.5 6.0 9.5 9.0 r2 (ppa) 8.5 1'2 6.5 6.0 (ppm) Figure 2.14: TOCSY spectra of a) b-Amp 11 and b) b-Amr 11 with 2 equivalents ofDB24C8-5en. Binding Constants of b-Amn: Isothermal titration calorimetry was used to find the association constants for [bAmn::>DB24C8-5en] and [b-Amn::>2{DB24C8-5en)] in dichloromethane. The association constant, Ka 1, for forming the one-to-one complex was 4.53 x 106 M- 1, and the association constant, Ka 2, for subsequently forming the one-to-two complex was 7 .17 x 10 7 M-1. The relationship between Kai and Ka2 (i.e., Ka2fKa1 > 0.25) implies that there is positive cooperativity between the two binding sites. 15 - 17 32 Catalyst 1 Catalyst 3 I\ Cy3P Cl I~ )~:('(~ c,~ju'~Ph PCy 3 Figure 2.14 Ruthenium alkylidene metathesis catalysts. ~~ Optimization of Metathesis Conditions Three different alkylidene ruthenium metathesis catalysts 18 (Fig. 2.15) were studied to find the best one for the pseudorotaxane templation system. Cross metathesis reaction ofDB24C8-5en, with or without the b-Amn template, were run with each of the three catalysts and the crude reaction mixtures were analyzed by HPLC. The results (Table 2.8) showed that catalyst 3 is most effective at achieving a balance between promoting dimerization and keeping the yield of the background reaction (i.e., without template) to a minimum. Interestingly, using 10 mol¾ 3 gave a 6.5-fold template enhancement over background, while using twice the amount of catalyst only gave a 2fold enhancement. + + + + Catalyst Yield(%) of Dimer 10 mol¾ 1 12 10 mol¾ 1 21 10 mol¾ 2 22 10 mol¾ 2 65 10 mol¾ 3 6 10 mol¾ 3 39 20 mol¾ 3 26 20 mol¾ 3 50 Table 2.8: A survey of metathesis catalysts. Reactions were run with 1mM DB24C8-5en at 40 °C. 33 A monomer with an olefin that is linked to the crown ether through an ether bond should be more electron rich, and therefore more reactive, than one attached via an ester bond. To test this theory, metathesis reactions were run with the ether-linked monomer, DB24C8-e5en. The results (Table 2.9) demonstrate that this monomer is more effective within the templation system than the ester-linked DB24C8-5en. Therefore, the crownether derivative for ADMET reactions (DB24C8-diene) was synthesized with etherlinked n-butene functionalities. + + Catalyst Yield (%) of Dimer 10 mol¾ 2 35 10 mol¾ 2 72 25 mol¾ 3 17 25 mol¾ 3 78 Table 2.9: Reactions with lmM DB24C8-e5en at 40 °C. A continual problem in the metathesis reactions with the crown-ether monomers was the production of side-products arising from olefin isomerization. Addition of benzoquinone has been shown to reduce the amount of isomerization in ruthenium catalyzed metathesis reactions. 19 However, for this system (Table 2.10), the addition of benzoquinone did not reduce the amount of side-products generated nor did it significantly improve the yield of dimer. Therefore, benzoquinone was not used further. Template Catalyst Benzoquinone 12 10 mol¾ 1 10 mol¾ 1 + 10 mol¾ 1 + 10 mol¾ 1 + 10 mol¾ 3 + 10 mol¾ 3 Yield(%) of Dimer + 6 21 + 26 38 + 27 Table 2.10: The influence ofbenzoquinone on metathesis dirnerization of lmM DB24C8-5en at 40 °C. 34 Verification of [3]-Pseudorotaxane Formation: There are two potential modes for a crown-ether to associate with an ammonium ton. In one mode, the crown-ether threads over the benzyl moiety (or other end-group) and encircles the ammonium center (Fig. 2.16a). The alternate "cupping" mechanism involves the crown-ether wrapping around the ammonium center without sliding over the end-group (Fig. 2.16b ). Previous work with a stoppered dibenzyl ammonium ion has shown that threading is not necessary for significant shifts to be observed in 1H NMR spectra (Smidt, S. unpublished). The association due to a cupping conformation is undesirable for two reasons: first, it is a more transient type of interaction, and second, it is not specific for the DB24C8-sized crown ether. a Figure 2.16: Two potential modes for a DB24C8 to bind a benzyl-ammonium ion, a) by threading over the benzyl group, orb) by wrapping around the ammonium center in a cup-like conformation. To test the mode of association, metathesis reactions were run with addition of one of two different sized crown-ethers molecules without olefin modification. It was hypothesized that addition of unmodified DB24C8, a crown-ether large enough to bind the template ammonium center by threading over the benzyl functionality, would block DB24C8-5en from threading onto the template and thereby reduce template-directed metathesis due to a threading interaction. On the other hand, addition of unmodified DB 18C6, a crown-ether too small to thread over the benzyl group, would only reduce 35 template-directed metathesis arising from a cup-like binding interaction. Analysis of the reactions by HPLC showed that addition of a large excess of DB24C8 had a minimal effect on the reaction without template, but drastically reduced the yield of the reaction containing template b-AmB (Fig. 2.17). On the other hand, addition of a large excess of DB 18C6 had no influence on either reaction. Taken together these results verified that the desired [3]-pseudorotaxane metathesis complex was forming. 90 - ~ - - - - - - - - - - - - - - - - - - - - ~ D No Additional Crown ■ 10x DB24C8 80 -----------------1 ■ 10x DB18C6 70 - l - - - - - - - = = = = = = = = - - - - - - - - - - ~ 60 + - - - - - - - - - - - - - - - - - lr""""""""""-:1 (I) E o 50 +------------------1 LL (I) E 40 + - - - - - - - - - - - - - - - - - - 1 c ("01 0: 0 ~ DB18C6 0 ~30 + - - - - - - - - - - - - - - - - - - 1 0~ 0 ~o~ 10 + - - ---1 No Template b-Am 8 Figure 2.17: Influence of a threading (DB24C8) versus a non-threading (DB18C6) crown-ether on metathesis efficiency. Reactions run with 1 mM DB24C8-5en, 25 mol % catalyst 2, and either no template or b-Am 8 at 40 °C. b-AmP1 1 Influence on ROMP of DB24C8-CP: 1 H NMR spectroscopy was used to follow a ROMP reaction ofDB24C8-CP in the presence and absence of the b-Amp 11 template. The reaction was run at a concentration of crown-ether monomer less than 10 mM to stay significantly below (90 times) the critical concentration needed for ROMP to proceed in solution. The reaction of [b-AmPI 1::,2(DB24C8-CP)] using catalyst 2 was established to be almost 45% complete within 16 hours, whereas the reaction ofDB24C8-CP in the absence of the template b-AmP1 1 resulted in less than 15% conversion under identical reaction conditions in the same amount of time (Fig. 2.18). The template also influenced the rate 36 of the reaction. In the presence of b-Amp 11 , ROMP of the monomer occurred faster than in the absence of template. 50 --- --·· -·-·-----·-- - ! 45 40 C: 0 30 g! 25 -~ C: 0 (J ~ 0 20 . .. • • • • • • - • • • - • • - 35 • • • =. • b-Amp11 I , I□ No Template ____, I - •• • 15 10 --- "· □ o □ DD D D D D D D D D D D D D D D D D D D D .r:£ [5)- I 5 I I 0 0 200 400 600 8 00 1000 1200 time (min) Figure 2.18: Influence ofb-Amr 11 on ROMP as detem1ined by reduction of olefin peak in 1H NMR. Reactions run with 8.25 mM DB24C8-CP, 10 mol¾ catalyst 2, and b-Amr 11 or no template at rt. Influence of b-Amp 11 on Cross Metathesis: Cross metathesis reactions ofDB24C8-5en (10 mM) using catalyst 2 with and without template b-Amp 11 were monitored by 1H NMR. As with the ROMP experiment, the template enhanced both the rate and yield of the dimerization (Fig. 2.19). The reaction of [b-Amp 11:::,2(DB24C8-5en)] was greater than 70% complete in 80 minutes, while the reaction ofDB24C8-5en without template was only 57% within the same amount of time. 37 80 70 • 60 • • C D 50 • • • • D D D D D D D D D D D D D D 1•:□ b-Amp11 I No Template: ; D > 40 I I D Q) . 1 I 0 ·;;; ... • • • • • • ----- ----• C 0 (.) ~ 0 • 30 I 20 I ! I 10 II 0 20 0 40 60 80 100 120 reaction time (min) Figure 2.19: Influence ofb-Amr 11 on CM as determined by 1H NMR. Reactions run with 10 mM DB24C8-5en, 10 mol % catalyst 2, and b-Amr 11 or no template at 40 °C. 5.0e7 · 41.887 -i 4.6e7 i 4.4c7 'il.287 ~.087 [DB24C8-5en•Na] ➔ t 583.0 3.8e7 • 3.607 3.4,r,7 3.287 S.De7 e- 2.~7 1!i, 2.&,1 4 I 24&7 .ij 2.2e7 2.0e7 1.8e7 [(DB24C8-5en) 2•Na] '---i, 11 16,2 1.6e7 1.<1e7 1.2e7 1.0c7 8.086 6.096 4 .0e6 Figure 2.20: ESI-MS analysis of CM reaction with 10 mM DB24C8-5en, catalyst 2 and b-Amr 11 . 38 Aliquots of both reactions were analyzed by ESI-MS. Peaks arising from the DB24C8-5en starting material and dimer product were visible in both spectra. In the spectrum (Fig. 2.20) of the reaction mixture with b-Amp 11 , there were also peaks from bAmpu, a doubly charged ions for two [2]-pseudorotaxanes: [b-Amp 11::::)DB24C8-5en] and [b-Amp 11 ::::)(DB24C8-5en)i]. Influence of b-Amn on Cross Metathesis: The same cross metathesis experiment was run using template b-Am8 . 7 The dimerization of [b-Am 8 ::::)2(DB24C8-5en)] was established to be 73% complete within 40 minutes, whereas the dimerization of DB24C8-5en in the absence of the template resulted in only 48% conversion in the same amount of time. The template also influenced the rate of the reaction. In the presence of b-Amn, cross metathesis of DB24C8-5en occurred faster than in the absence of template. Comparison of Templates for Cross Metathesis: The first three bis-ammonium peptide templates, b-Ampu, b-Amp 18 and b-Amp2o, were compared to the linear template b-Amn (Table 2.11), which had previously been shown to significantly influence cross metathesis with the templation system. In a 2 hour metathesis reaction ofDB24C8-5en, b-Amn showed almost a three-fold enhancement in dimerization compared to the background reaction. On the other hand, the three peptide template had almost no influence on the yield of dimer. From these results it can be concluded that b-Amn promotes dimer formation, while peptide templates b-Amp 11 , bAmp1s and b-Amp20 do not. 39 Template Yield(%) of Dimer Enhancement over background 23 b-AmB 65 2.8 b-Amp 11 31 1.3 b-Amp1s 25 I.I b-Amp20 27 1.2 Table 2.11: Comparison of bis-ammonium templates in cross metathesis reactions using 10 mol% 2 and lmM DB24C8-5en at 40 °C. Comparison of Templates for ADMET Dimerization: ADMET reactions were carried out using the bis-ammonium templates and catalysts 2 and 3 (Table 2.12). A very high background reaction yield was seen in the reactions with catalyst 2, but was reduced when the concentration ofDB24C8-diene was reduced by half. Using this catalyst, b-Amn showed a template-directed enhancement of dimer yield at both concentrations of monomer, although the effect was more pronounced at the lower concentration. On the other hand, there was only a minimal template effected observed for b-Amp 11 . Table 2.12: Comparison of different templates and different catalyst in ADMET reactions with lmM DB24C8-diene at 40 °C. Template Catalyst Yield (%) of Dimer 10 mol¾ 2 56 10 mol¾ 2 76 10 mol¾ 2 34* 10 mol¾ 2 72* 10 mol¾ 2 58 10 mol¾ 2 66 25 mol¾ 3 33 B 25 mol¾ 3 87 P3 25 mol¾ 3 85 B B P11 * [DB24C8-diene] = 0.5mM 40 Using catalyst 3 significantly reduced the background reaction seen with catalyst 2. The templates b-Amn and b-Amp3 were found to be equally effective with this catalyst, effecting close to a three-fold enhancement in dimer yield over background. Metathesis Dimerization with the b-AmA Templates: The series of alkane-linked bis-ammonium templates, b-AmA4 through b-AmA 12 , was used to asses the influence of template rigidity on the ability of a template to promote metathesis. Comparison of the metathesis reactions with the alkane-templates to those with the rigid template b-Amn showed that increasing the flexibility of the template did not compromise template efficacy (Table 2.13). All of the b-AmA templates gave significantly higher dimerization yields than the background reaction, and as high, or higher, than the rigid template b-Amn. In addition, within the range studied there was no correlation between linker length and template efficacy. However, it is likely distance dependence would be seen with increasingly long spacing between binding sites, an experiment that was beyond the scope of this study. Template Yield(%) of Dimer b-AmA4 63 b-AmAs 61 b-AmA6 68 b-AmAs 60 b-AmA10 63 b-AmA12 66 b-Amn 52 25 Table 2.13: Influence oflinear templates on ADMET dimerization. Reactions were run with lmM DB24C8-diene and 20 mol% 3 at 40 °C. 41 Template Purification: One distinct difference between the effective templates and the ineffective ones was their purity. The peptides Pl 1, P18 and P20 could not be purified after the reductive amination reaction, and they were the least effective templates. On the other hand, P3, B and the A-series were all purified prior to their use as templates, and they all significantly enhanced the yield of metathesis dimerization. To investigate this apparent correlation between efficacy and purity, b-AmA6 was synthesized without purification after the benzyl protection. As with the peptides, selective mono-benzyl protection of each amine functionality was confirmed by MS. It was found that template purification had a significant impact on the ability of bAmA6 to influence metathesis dimerization (Table 2.14). Comparison of the same template with and without purification shows that purification increased the dimer yield by 50%. Template Yield (%) of Dimer b-AmA6 78 Crude b-AmA6 39 21 Table 2.14: Influence of template purity on metathesis dimerization. Reactions were run with lmM DB24C8-diene and 20 mol¾ 3 at 40 °C. ADMET Trimerization: In all of the AD MET dimerization reactions no trimer or longer products were observed by HPLC or mass spec analysis. This suggests that a 1 mM concentration of monomer is not high enough to allow polymerization to occur. However, if the template motifs can be extended beyond dimerization, the addition of a tris-ammonium template should promote the formation of trimer products. 42 Yield (%) Dimer Yield(%) Trimer 28 0 t-Amp1s 31 0 t-Amps 58 19 Template Table 2.15: Template-directed metathesis trimerization. Reactions were run with lmM DB24C8-diene and 20 mol¾ 3 at 40 °C. Neither the control reaction nor the reaction with t-Amp 15 , a template that was not purified, yielded any trimer product. The t-Amp 15 template did enhance the metathesis yield compared to background. However, the yield of trimer was modest and most of the product formed was dimer. Analysis of the dimer and trimer products showed them to be cyclic - arising from a final intramolecular metathesis reaction to close the molecule on itself. III. Conclusions The series of experiments outlined above was designed to study the suitability of the pseudorotaxane binding motif for a template-directed metathesis polymerization system. A peptide template, bAmp3 , was identified that was as effective as the control linear templates at enhancing metathesis dimerization. It was also discovered that template rigidity and defined structure are not necessary to promote metathesis dimerization, although they may be important for longer polymerization reactions. The successful bis-ammonium peptide template design did not extend to the trisammonium template. While t-Amp 5 was able to influence the overall yield of the metathesis reaction, the majority of the product formed was the cyclic dimer. It was hypothesized that a stronger non-covalent or a covalent recognition motif might aid in the synthesis of oligomers larger than dimers. 43 IV. Experimental General: The peptide starting materials for template b-Ampu and t-Amp1s were synthesized by the Biopolymer Synthesis and Analysis Facility at the California Institute of Technology and was used without further purification. The peptide starting materials for templates bAmp 18 , b-Amp 20 , and b-Amp3 were synthesized by the GenScript Corporation. The peptide starting material for b-Amp3 was also purified to 97% by GenScript. All chemicals were obtained from commercial sources and used as received. Deuterated NMR solvents were from Cambridge Isotopes. 1 H-NMR spectra were recorded on Varian 300 or 500 MHz NMR. Electrospray ionization (ESI) and MALDI mass spectrometry were performed at the Mass Spectrometry Center in the Department of Chemistry at the California Institute of Technology. Circular dichromism (CD) spectra were recorded on an AVIV CD spectrometer, model 62DS. Synthesis of b-Amru, b-Ampn~, and b-Amp20 Templates: Benzylation: The peptide starting material was dissolved in trifluoroethanol (TFE) and IM aq. NaOH. Benzaldehyde (2 equivalents) was added and the solution was stirred for 20 min. Sodium borohydride (4 equivalents) was added and the solution was stirred for 30 min. Consecutive additions ofbenzaldehyde and sodium borohydride were repeated two more times, at which point MALDI-MS analysis confirmed the reaction was complete. A small amount of acetone was added to the solution to quench excess NaBH4 . The reaction was diluted with chloroform and washed with water (2x). The organic layer was dried over Na2SO 4 and concentrated to yield the benzyl-modified product. 20 Pll: 10.0 mg of peptide Ac-VALKVALKVAL-NH2 (0.00858 mmol) in 500 µL of TFE and 50 µL of IM NaOH (aq.) yielded 10.7 mg of product (93%): MALDI-MS (m/z) 1346 [M + HJ\ 1369 [M + Nat. PlSF: 13.3 mg of peptide Ac-VALKVALKVALKVAL-NH 2 (0.00843 mmol) in 700 µL of TFE and 10 µL of IM NaOH (aq.) yielded 14.4 mg of product (90%): MALDI-MS (m/z) 1902 [M + HJ\ 1924 [M + Nat. P18F: 10.0 mg of peptide Ac-LALLALKLALAKLALLAL-NH2 (5.34 µmol) in 500 µL of TFE and 10 µL of IM NaOH (aq.) yielded 5.6 mg of product (50%): MALDI-MS (m/z) 2091 [M + HJ\ 2113 [M + Nat. P20F: 10.0 mg of peptide Ac-VALVALKVALVALKVALVAL-NH2 (4.96 µmol) in 500 µL of TFE and 10 µL of IM NaOH (aq.) yielded 7.8 mg of product (70%): MALDI-MS (m/z) 2233 [M + HJ\ 2255 [M + Nat. Pl12+·(PF61 2 : Pll (23.2 mg, 0.0172 mmol) was dissolved in 1 mL of TFE, and a 60 wt % solution ofHPF6 in water (20.9 µL, 0.086 mmol) was added dropwise. The solution 44 was stirred for 2 hours and then concentrated to give a clear gel. Addition of water caused precipitation of a white solid, which was isolated by filtration. The solid was washed with water (2x), then ether (2x). It was collected and dried over night to give the desired product (16.1 mg, 57%): ESI-MS (m/z) 674 [M + 2H]2+. [H(Et20)2t[BArFr: A 0.045M solution ofNaBArF (49.7 mg, 0.0561 mmol) in anhydrous EtiO was dried over 4 angstrom molecular sieves for 12 hr. The solution was cooled on ice, and lM HCl in anhydrous Et2O was added dropwise. The solution was allowed to sit under argon on ice for 1.5 hr, during which time it turned cloudy as NaCl precipitated. The solution was filtered through a 0.45µm PFTE filter and concentrated to give a white solid. At all times, the reaction and product were kept on ice. b-Amp 11 : Pll (13.4 mg, 0.00996 mmol) was cooled on ice under argon. A 0.lM solution of [H(Et20)2t[BArFr in anhydrous EtiO (199.0 µL) was added. The material was dissolved/suspended in 800 µL anhydrous EtiO, and was stirred on ice for 1.5 hr. The reaction was concentrated, and the material was taken up in DCM. The solution was filtered to remove the insoluble particulate, and re-concentrated to yield the desired product (14.0 mg, 46%): ESI-MS (m/z) 674 [M + 2H]2+, 863 [BArFr. b-AmFp11 : The same procedure was used as for b-Amp 11 , but starting with Pll F (17.1 mg, 0.0124 mmol) and yielding 18.9 mg product (49%). t-AmFp 15 : The same procedure was used as for b-Amp 11 , but starting with P15F (5.0 mg, 2.63 µmol) and yielding 8.0 mg product (68%), the identity of which was confirmed by 1 HNMR. b-AmFPls: The same procedure was used as for b-Amp 11 , but starting with P18F (5.6 mg, 2.68 µmol) and yielding 7.4 mg product (72% ). b-AmFp20 : The same procedure was used as for b-Amp 11 , but starting with P20F (7.8 mg, 3.49 µmol) and yielding 8.3 mg product (60%). Synthesis of b-Amn and t-Am~: Solid-phase Synthesis of Peptide Starting Materials: The peptide starting materials Ac-KAK-NH2 and Ac-KAKAK-NH2 were synthesized using standard Fmoc chemistry on Rink Amide MBHA resin from Novabiochem. During the synthesis, Fmoc deprotection was achieved with 20% piperidine in DMF. Amino acid coupling was achieved using the coupling reagents HOBt and DIPCDI in DMF. After the final Fmoc deprotection, the N-terminus was acetylated using acetic anhydride and DIEA in DMF. Lysine deprotection and peptide cleavage were achieved using a solution of TFA and TIS in water. After cleavage, the peptides were precipitated with cold ethyl ether, isolated by filtration, dissolved in MeOH and concentrated. The identity of each peptide was confirmed by 1H NMR and ESI MS. 45 Benzylation: The peptide was dried under vacuum in a flame-dried flask overnight. An oven-dried stirbar and activated 4 A molecular sieves were added to the flask under argon. The peptide was dissolved in anhydrous methanol and allowed to stir under argon for approximately 20 minutes. Benzyaldehyde (6 equivalents) was added to the solution. The reaction was allowed to stir for 30 minutes, and then the solvent was removed under vacuum. The material was dried under vacuum for at least 2 hours to try to remove as much unreacted benzaldehyde as possible. The material was redissolved in anhydrous methanol and sodium borohydride (12 equivalents) was added. After stirring for 30 minutes, excess reducing agent was quenched with acetone and the reaction was concentrated. The reaction was dissolved in ethyl acetate and washed with water (3x), dried and concentrated. The crude material was dissolved in 9:1 0.1% TFA (aq):CH3CN, with the addition of extra CH3CN to allow complete dissolution. The material was then purified by HPLC on a C 18 reverse phase column, using a gradient from 10% to 100% CH3CN in 0.1 % TFA (aq). The run was monitored by UV at 250 nm. The desired product was collected and concentrated as the TFA-salt. 1H NMR and ESI-MS were used to verify the product identity. P3: 35.5 mg of peptide Ac-KAK-NH2 (0.0919 mmol) yield 14.7 mg of product as a TFAsalt (28%). P5: 45.0 mg of peptide Ac-KAKAK-NH2 (0.0768 mmol) yielded 7.0 mg of product as a TFA-salt (11 %). Generation of BArF-salt: The TF A-salt of benzylated peptide was dissolved in a minimal amount of water. Dichloromethane was added to make a biphasic solution. Two or three equivalents ofNaBarp were added to the solution. The solution was stirred vigorously for 12 hours. The organic layer was separated, washed with water (2x), dried and concentrated. 1H NMR was used to verify the correct ratio of peptide to ion. b-Amp 3 : 3.4 mg of the TFA-salt of P3 (4.28 µmol) yield 7.8 mg of product (79%). t-Amp 5 : 7.0 mg of the TFA-salt of P5 (5.84 µmol) yielded 16.0 mg of product (79%). Synthesis of b-Am_a Series: The series of alkyl linked bisammonium templates was synthesized from commercially available alkyl-diamine reagents using standard reductive amination conditions. Each alkyl-diamine was dissolved in anhydrous MeOH to a concentration of IM, and 2.2 equivalents of benzaldehyde was added. After stirring for 30 min., 4.4 equivalents of NaBH4 was added. The reaction was stirred for at least 1 hr, and excess reducing agent was quenched with acetone. The reaction was diluted with EtOAc, washed with water (2x) and then with sat. brine (2x), dried over Na2SO4 and concentrated. Purification by silica gel column chromatography yielded pure dibenzyl-protected product. The dibenzyl amines were protonated with HCl, and the resulting HCl-salt precipitate was isolated. Then, the chloride counterion was replaced by the BArF counterion in a biphasic solution 46 of dichloromethane and water using 2 equivalents ofNaBArF. 1H NMR was used to confirm the correct template to counterion ratio. Synthesis ofM, M-pF, M-pOMe, M-pCN: O-TBDPS-butanol-4-amine: Butanol-4-amine (200.0 mg, 2.244 mmol) was dissolved in 1.8 mL anhydrous DMF. Imidazole (3.366 mmol) was added, then TBDPSCl (2.468 mmol) was added dropwise. The reaction was complete in 4 hr, as judged by TLC. Removal of the solvent terminated the reaction, and the crude material was used without further purification. O-TBDPS-butanol-4-benzylamine: Crude O-TBDPS-butanol-4-amine (1.122 mmol) was dissolved in methanol, and the pH was taken to 8.5 with 0.5 mL of 1.0 M NaOH. Benzaldehyde (2.244 mmol) was added and the solution was stirred for thirty min at rt. NaBH4 (4.488 mmol) was added, causing the evolution of gas, and the reaction was stirred for thirty min. The consecutive additions of benzaldehyde and N aB~ were repeated one more time, with thirty min of stirring after each addition. At that point, TLC showed that 1 had been consumed. Excess NaBH 4 was quenched with acetone and the reaction was concentrated. The residue was taken up in EtOAc and washed with water (2x). The organic layer was dried over Na2 SO4 and concentrated. The crude material was purified by silica gel column chromatography in 1: 1 EtOAc:Hex to yield pure material (231.9 mg, 50%): ESI-MS (m/z) 418 [M + Ht. M: 231.9 mg of O-TBDPS-butanol-4-benzylamine (0.555 mmol) was dissolved in methanol. 6.05M HCl was added to give a pH 2.5 solution, which was stirred at rt for 3 hr. The reaction was concentrated and the residue was taken up in biphasic solution of water and CHCb, and NaBArF (<0.555 mmol) was added batch-wise with vigorous shaking after each addition until it no longer went into solution immediately. The organic layer was separated, dried over Na2 SO4 and concentrated to yield product with the correct 1:1 M+ to BArF-ratio (534.6 mg, 75%). O-TBDPS-butanol-4-pFluorobenzylamine: The same procedure was used as for OTBDPS-butanol-4-benzylamine, replacing benzaldehyde with 4-fluorobenzaldehyde. A reaction with 100 mg O-TBDPS-butanol-4-amine (0.305 mmol) yielded 33.4 mg pure product (25%): ESI-MS (m/z) 436 [M + Ht. O-TBDPS-butanol-4-pMethoxybenzylamine: The same procedure was used as for 0TBD PS-butanol-4-benzylamine, replacing benzaldehyde with 4-fluorobenzaldehyde. A reaction with 100 mg O-TBDPS-butanol-4-amine (0.305 mmol) yielded 70.0 mg pure product (51 %): ESI-MS (m/z) 448 [M + Ht. O-TBDPS-butanol-4-pCyanobenzylamine: The same procedure was used as for OTBDPS-butanol-4-benzylamine, replacing benzaldehyde with 4-fluorobenzaldehyde. A reaction with 100 mg O-TBDPS-butanol-4-amine (0.305 mmol) yielded 3.4 mg pure product (3%): ESI-MS (m/z) 443 [M + Ht. 47 M-pF: The same procedure was used as for M, but starting with O-TBDPS-butanol-4pFluorobenzylamine (55.0 mg, 0.126 mmol), to yield 110.0 mg product (67%). M-pOMe: The same procedure was used as for M, but starting with O-TBDPS-butanol4-pMethoxybenzylamine (70.0 mg, 0.126 mmol), to yield 98.0 mg product (55%). M-pCN: The same procedure was used as for Model, but starting with O-TBDPSbutanol-4-pCyanobenzylamine (3.4 mg, 0.00768 mmol), to yield 8.4 mg of product (84%). Synthesis of DB24C8-derivatives: AD MET diene: The hydroxyl functionalities of methyl 3,5-dihydroxybenzoate were alkylated using 1-bromobutene and K2CO 3 . The resulting compound was saponified using potassium hydroxide in a water and ethanol solution to give the desired diolefin acid compound. Representative Synthesis of Ester-linked DB24C8-Derivatives: EDC (1.36 mmol) was added to a stirring solution of the carboxylic acid derivative (1.36 mmol) and a few crystals ofDMAP in DCM (10 mL), followed by the addition ofDB24C8-0H (1.04 mmol). The reaction mixture was allowed to stir at ambient temperature for 16 hr, before removing the solvent in vacuo. Purification by silica gel column chromatography (2% MeOH/DCM) afforded each product as a white solid. DB24C8-5en: 1H NMR (500 MHz, CD 2Ch) 8 = 2.32-2.38 (m, 2H), 2.40-2.44 (m, 2 H), 3.75 (s, 8H), 3.83-3.86 (m, 8 H), 4.09-4.12 (m, 8H), 5.01 (s, 2H), 4.96-5.06 (m, 2H), 5.79-5.85 (m, lH), 6.82-6.86 (m, 7H); BC NMR (125 MHz, CDCb) d = 28.7, 33.4, 66.1, 69.3, 69.4, 69.7, 69.8, 71.19, 71.22, 113.4, 114.2, 115.4, 121.3, 121.6, 128.9, 136.5, 148.7, 148.8, 172.8; MALDI-ICR: [C30H40O10Nat = 583.2515, calcd = 583.2519. DB24C8-diene: 1H NMR (300 MHz, CDCb) 8 = 2.58 (dt, 4H), 3.85 (s, 8H), 3.91-3.98 (m, 8H), 4.06 (t, 4H), 4.12-4.20 (m, 8H), 5.08-5.30 (m, 6H), 5.82-5.91 (m, 2H), 6.62 (s, lH), 6.78-7.0 (m, 7H), 7.19 (s, 2H); HRFAB: [C40HsoO12Nat= 722.3338, calcd 722.3303. DB24C8-e5en: A solution ofDB24C8-0H (2.0 g, 4.2 mmol) in THF (40 mL) was slowly added dropwise to a suspension ofNaH (60% in oil, 252 mg, 10.5 mmol) in THF (40 mL). This mixture was stirred at ambient temperature under an N2 atmosphere until no further H2 evolution was observed. The mixture was cooled with an ice bath and a solution of 5-bromopentene (934 mg, 6.3 mmol) in THF (40 mL) was added dropwise. The reaction was stirred 2 days under ambient conditions and subsequently quenched with isopropanol, and then with water. All solvents were removed under reduced pressure and the residue puried by silica gel column chromatography with EtOAc to give the desired product (1.44 g, 63%) as a white solid: 1H NMR (300 MHz, CDCh) 8 = 1.611.80 (m, 2H), 2.14-2.20 (m, 2H), 3.43 (t, 2H), 3.82 (s, 8H), 3.83-3.89 (m, 8H), 4.07-4.09 48 (m, 8H), 4.41 (s, 2H), 4.93-5.03 (m, 2H), 5.71-5.88 (m, lH), 6.80-6.91 (m, 7H); 13 C NMR (125 MHz, CDCb) d = 28.9, 30.2, 69.0, 69.4, 69.7, 69.8, 71.6, 73.0, 113.9, 114.2, 115.4, 121.3, 121.6, 128.9, 136.5, 148.7, 148.8; HRFAB: [C30~2O9Nat = 546.2801, calcd = 546.2829. Analysis of Complexes between Ammonium Centers and Crown-Ether Monomers: Representative Analysis of Complex Formation by ESI-MS: A solution of b-Amr 11 with 2 equivalents ofDB24C8-CP in 1:20 CHCb:TFE was allowed to evaporated slowly under a stream of argon. The dry material was taken up in 3:1 MeOH:CH3CN. To make the injection sample, 1 µL of the solution was diluted with 200 µL of 4: 1 MeOH:CH3CN. The sample was injected via syringe pump at 5 µL/min. To facilitate complex formation, the following recommended settings for the spectrometer were used: capillary temperature (200° C) and tube lens offset (-39 V). 14 Because oflimitations of the instrument, the ideal spray voltage setting of 0.01 kV could not be used; instead the default setting of 4.5 kV was used. Typical Procedure for Metathesis Experiments: A solution containing the DB24C8derivative (1.0 mmol) and template (0.5 mmol for bis-templates, 0.33 mmol for tristemplates) in DCM was stirred at ambient temperature for 0.5 h. The catalyst, dissolved in DCM, was added to the reaction mixture such that the final concentration of the crown derivative was 1.0 mM (or 0.5 mM for DB24C8-diene ). The solution was heated to 40 °C for 2 h, before quenching the catalyst with ethyl vinyl ether (0.1 mL). The solvent was removed and the sample was subjected to HPLC analysis. 49 V. References: 1. Cantrill, S. J.; Pease, A. R.; Stoddart, J. F., J. Chem. Soc., Dalton Trans., 2000, 37153734. 2. Ashton, P. R.; Chrystal, E. J. T.; Glink, P. T.; Menzer, S.; Schiavo, C.; Spencer, N.; Stoddart, J. F.; Tasker, P.A.; White, A. J.P.; Williams, D. J., Chem. Eur. J., 1996, 2, 709-728. 3. Gutmann, V. W., E., Inorg. Nuc. Chem. Lett., 1966, 2, 257-260. 4. Karle, I. L.; Flippenanderson, J. L.; Uma, K.; Balaram, P., Biopolymers, 1993, 33, 827837. 5. Blackwell, H. E.; Sadowsky, J. D.; Howard, R. J.; Sampson, J. N.; Chao, J. A.; Steinmetz, W. E.; O'Leary, D. J.; Grubbs, R.H., J. Org. Chem., 2001, 66, 5291-5302. 6. Brookhart, M.; Grant, B.; Volpe, A. F., Organometallics, 1992, 11, 3920-3922. 7. Cantrill, S. J.; Grubbs, R.H.; Lanari, D.; Leung, K. C. F.; Nelson, A.; Poulin-Kerstien, K. G.; Smidt, S. P.; Stoddart, J. F.; Tirrell, D. A., Org. Lett., 2005, 7, 4213-4216. 8. Raymo, F. M.; Houk, K. N.; Stoddart, J. F., J. Am. Chem. Soc., 1998, 120, 9318-9322. 9. Beil, J.B.; Lemcoff, N. G.; Zimmerman, S. C., J. Am. Chem. Soc., 2004, 126, 1357613577. 10. Toniolo, C.; Benedetti, E., Trends Bioc. Sci., 1991, 16, 350-353. 11. Marqusee, S.; Robbins, V. H.; Baldwin, R. L., Proc. Nat. Acad. Sci. U.S.A., 1989, 86, 5286-5290. 12. Wagner, G., Prog. Nucl. Magn. Reson. Spectrosc., 1990, 22, 101-139. 13. Dalley, N. K.; Bradshaw, J. S.; Larson, S. B.; Simonsen, S. H., Acta Crystallogr., Sect. B: Struct. Sci., 1982, 38, 1859-1862. 14. Julian, R.R.; Beauchamp, J. L., Int. J. Mass Spectrom., 2001, 210, 613-623. 15. Badjic, J. D.; Nelson, A.; Cantrill, S. J.; Turnbull, W. B.; Stoddart, J. F., Acc. Chem. Res., 2005, 38, 723-732. 16. Mulder, A.; Huskens, J.; Reinhoudt, D. N., Org. Biomol. Chem., 2004, 2, 3409-3424. 17. Ercolani, G., J. Am. Chem. Soc., 2003, 125, 16097-16103. 18. Grubbs, R.H., Ed., Handbook of Metathesis. Wiley-VCR: Weinheim, 2003. 19. Hong, S. H., Sanders, D.P., Lee, C.W., Grubbs, R.H.,J. Am. Chem. Soc., 2005, 127, 17160 - 17161. 20. Saxena, I.; Borah, R.; Sarma, J. C.,JndianJ. Chem., Sect. B, 2002, 41, 1970-1971. 50 Chapter 3 A Boronate Ester Recognition Motif 51 I. Introduction Alternate Recognition Motifs: Previous investigation of template-directed metathesis demonstrated the pseudorotaxane was a useful recognition motif for metathesis dimerization, but not for trimerization or longer polymerizations. The tris-templated reactions afforded high yields of a cyclic dimer product, and only small amounts of trimer. It was hypothesized that after the initial dimerization reaction, one of the crown-ether moieties was dethreading from one of the ammonium binding sites which allowed an intramolecular metathesis cyclization reaction to occur yielding a cyclic dimer. To eliminate the formation of cyclic dimer and therefore allow a higher yield of trimer, a more stable recognition motif was sought. Experiments with palladium and platinum pincer binding motifs indicated that the pincer metal caused olefin isomerization, leading to the formation of myriad dimer products (Nelson, A., Hou, C., Grubbs, R.H, unpublished). For this reason, the pincer recognition motif was untenable for the metathesis template system. A cytosine-guanine recognition motif was also explored (Schmidt, S., Leung, K., Grubbs, R.H., Stoddart, J.F., unpublished). However, the limited solubility in aprotic solvents of the guanine half of the recognition unit made this motif unsuitable for metathesis polymerization. The Boronate Ester Recognition Motif: As an alternative, a boronate ester motif was examined. This interaction arises from a covalent reaction between a boronic acid and a diol and was anticipated to provide a stronger interaction than the pseudorotaxane, which is a non-covalent association. The pseudorotaxane needs low Gutmann donor number solvents to form a strong complex and 52 has an equilibrium that is shifted to favor the uncomplexed form in water. On the other hand, the boronate ester is covalently bound in anhydrous solvents and can also be formed in water (Table 3.1 ). The equilibrium in water is affected by the arrangement of the diol alcohols, with the catechol molecule forming the strongest boronate ester in water, followed by 1,2-diols and then 1,3-diols. 1 Diol Keq (M-1) Catechol 830 D-sorbitol 370 D-fructose 160 Table 3.1: Equilibrium constants with phenylboronic acid in aqueous buffer, pH 7.4. The boronate ester can be charged and in a tetrahedral geometry or can be neutral and in a trigonal planar geometry (Fig. 3.1 ). Within aprotic, organic solvents suitable for metathesis, the neutral boronate ester will have more favorable solubility properties. 1,2diols favor the formation of charged, tetrahedral boronate esters while 1,3-diols favor formation of neutral, trigonal planar boronate esters. 2 ' 3 Therefore, even though 1,2-diols form stronger boronate esters, the 1,3-diol functionality is more desirable for the template system. An additional benefit of the boronate ester motif is that it does not involve additional metals and so avoids the complications seen with the pincer motif. 53 Template Design: To adapt the boronate ester functionality into the two component template and binding monomer system, the two functionalities - boronic acid and the diol- must be incorporated into the two components - a peptide template and an olefin monomer. Because the boronate ester strength can be tuned by changing the diol reagent and it was more synthetically accessible to make functionality changes to the olefin-containing monomer rather than the peptide template, the system was established to involve diololefin monomers and boronic acid containing peptides. An additional benefit to this configuration was that the commercially available L-(4-Boronophenyl)alanine reagent could be used to incorporate the boronic acid functionality into the templates. The basic sequence of the series of peptide templates bearing boronic acid functionality (Table 3.2) was based upon the successful pseudorotaxane template, b- Type of Peptide Chemical Formula ID 3-mer Ac-F(pBA)-A-F(pBA)-NH2 b-BAp3 On-resin, 3-mer Ac-F(pBA)-A-F(pBA)-Rink Amide MBHA MBHA-b-BAp3 On-resin, 5-mer Ac-F(pBA)-A-F(pBA)-A-F(pBA)- Rink Amide MBHA MBHA-t-BAps On-resin, 5-mer Ac-F(pBA)-A-F(pBA)-A-F(pBA)-Tenta Tenta-t-BAp 5 Table 3.2: The series of templates studied bearing boronic acid functionalities for boronate ester formation. II. Results and Discussion Template synthesis: As a first step towards the synthesis of the peptides BAp3 and BAps, the free amine of L-(4-Boronophenyl)alanine was Fmoc-protected (Scheme 3.1 a). Using standard Fmoc solid phase synthesis on a Rink Amide MBHA resin, a tripeptide and 54 pentapeptide were synthesized (Scheme 3.1 b-i). After protecting the N-terminus with an acetyl group, acidic conditions were used to cleave the peptide from the resin with a C- E~ terminal amide. Reverse phase HPLC purification yielded the desired peptides. E B(OH)i ~ H2N HFmoc ~ B(OH)i ::,-_ H,N ~ OOH \ H ¥ HFmoc~C)-N0 C B(OH)i NHFmoc 8 OH I B'oH Scheme 3.1: Synthesis ofb-BAr 3 and t-BAr 5 : a) Fmoc-O-Su, 10% Na 2CO 3 , dioxane, 86%; b) piperidine, DMF; c) HOBt, DIPCDI, DMF; d) piperidine, DMF, then Fmoc-Ala-OH, HOBt, DIPCDI, DMF; e) piperidine, Fmoc-Phe(pBA)-OH, HOBt, DIPCDI, DMF;.f) piperidine, then AcP, NMP, pyridine, then TFA, TIS, H2O, 20%; g) piperidine, DMF, then Fmoc-Ala-OH, HOBt, DIPCDI, DMF; h) piperidine, Fmoc-Phe(pBA)-OH, HOBt, DIPCDI, DMF; i) piperidine, then Ac 2O, NMP, pyridine, then TFA, TIS, H2O, 14%. On-resin Template Synthesis: Using a resin-bound template is an effective method to drive up the yield of a reaction and to avoid unwanted side-product contaminants. 4-6 In this approach, any nonassociated monomers and their template-independent products may be washed away at the end of the reaction so that they do not contaminate the temp lated-dependent product profile. Controlled removal of the bound daughter polymer yields a completely templatedirected product. 55 To generate a useful on-resin peptide template, the same procedure as before was followed for the amino acid couplings to synthesize the b-BAp3 and t-BAp 5 peptides on Rink Amide MBHA resin. However, the original peptide synthesis used pyridine in the final N-terminal acetyl protection step. It was found that in order to perform metathesis on-resin, pyridine had to be excluded from the peptide synthesis because having the amine base trapped within the resin poisoned the catalyst. 7 To avoid the use of pyridine, the synthesis was revised to use a pentafluorophenyl acetate reagent8' 9 to acetylate the Nterminus instead of acetic anhydride in pyridine. A second type of resin, TentaGel S RAM, was also used as a templatepresentation platfom1. The TentaGel S resin is composed of a lightly crosslinked polystyrene matrix with alkyl-linked PEG grafts. The resin contains ~50-70% PEG and its properties are highly influenced by the PEG moeities. These properties manifest themselves in swelling behavior that is almost independent of the solvent used to swell the matrix and in the solution-like reactivity of the reactive sites. The b-BAp3 peptide was synthesized on the TentaGel S resin using the same procedure established on the Rink Amide BMHA resin. Monomer syntheses: An ADMET monomer containing a diol functionality (1,3-diol-diene) was synthesized (Scheme 3.2a) Starting from the 1-bromobutene alkylated 3,4dihydroxybenzoic acid compound used in the synthesis of DB24C8-diene, a symmetrical ADMET-diol monomer was synthesized in one step via esterification with 2hydroxymethyl-1,3-propanediol. Having a symmetrical diol-monomer simplified the 56 analysis of the boronate ester and the subsequent metathesis products; use of an unsymmetrical diol would yield diastereomeric boronate esters and metathesis products. A second monomer (catechol-diene) was also synthesized, using a catechol in place of the 1,3-diol (Scheme 3.2b) It was anticipated that using this diol functionality would give more stable boronate esters. To synthesize the compound, the alcohol groups of 3,4-dihydroxybenzaldehyde were silyl protected and the aldehyde was reduced to an alcohol. The resulting compound was coupled to the ADMET-acid intermediate and the silyl groups were removed to yield the desired product. Catechol-diene turned out to be unstable and difficult to store, and so was not used in the metathesis studies. (a) a • (b) HO:dO HO 1"' ...-::: b .. TBSO:dj TBSO 1"' ~ TBSO)G(~~ TBSO ...-::: c., I "' ...-::: TBSO)G(H 1"' TBSO e ► d • + ...-::: 1"' ~ HO)G(~~ HO ...-::: Scheme 3.2: Synthesis of (a) 1,3-diol-diene: a) EDC, DMAP, DCM, DMF, 35%; and (b) catecholdiene: a) TBDMSCI, imidazole, DMF, 77%, b) mCPBA, DCM, then K 2 CO 3 , MeOH, 55%, c) EDC, DMF, DCM, 57%, d) TBAF, THF, 13%. Boronate Ester Formation in Solution: Commercially available 7-octene-1,2-diol was used to establish the optimal conditions to form the boronate ester in solution. 1H NMR confirmed complete boronate 57 ester formation after the b-BAp3 peptide and 1,3-diol-diene were heated at 40 °C for two hours with molecular sieves in a solution of dichloromethane and methanol. Boronate Ester Formation On-Resin: Optimal on-resin boronate ester formation and hydrolysis conditions were studied using the MBHA-b-BAp3 template via HPLC analysis. In a series of steps, a standard metathesis reaction was replicated: 1,3-diol-diene was complexed to MBHA-b-BAp 3 , the monomer solution was collected and the resin was washed, the on-resin boronate ester complex was subjected to mock metathesis conditions (40 °C in DCM for 2hr, no catalyst), and the boronate ester was hydrolyzed to elute all diol-containing molecules. Using this method the ideal conditions for forming the boronate ester on-resin were found to be mixing for 4 hours at 40 °C in dichloromethane with 4 A molecular sieves and the ideal reagent to hydrolyze the boronate ester complex was found to be a 60% acetonitrile solution in aqueous 1% TF A. The HPLC trace of the elution solution showed a strong peak for 1,3-diol-diene, integration of which indicated that 95% of the monomer loaded onto the resin was recovered (Fig. 3.2). In addition, only a small amount of 1,3-diol-diene was seen in the mock metathesis solution, indicating that the boronate ester is stable under those conditions. 58 1.00 - Monomer 0 90 Mock Metathesis Solution Resin Elution Solution 0.80 0.70 060 :) 0 50 < 040 0.30 0.20 010 o.oo..i------~--_,_, ~=-----------..L.>O,,.....c:;_____,.'----~==-_,_j 35 00 36 .00 37 00 38 00 39 00 40 .00 41 00 42 00 43 00 44 00 45 00 46 00 47 00 Minutes Figure 3.2: HPLC analysis of on-resin mock metathesis reaction and diol elution of 1,3-diol-diene and l\IBHA-b-BP3. In an experiment with unmodified MBHA Rink Amide resin, neither the mock metathesis solution nor the elution solution showed any 1,3-diol-diene monomer by HPLC analysis. This demonstrated that the 1,3-diol-diene monomer bound selectively to the peptide attached to the resin, and was not interacting non-selectively with the resin. Template-Directed Cross Metathesis: To test the ability of the boronate ester system to directed metathesis dimerization, the boronate ester complex formed from 7-octene-1,2-diol and b-BAp3 was subjected to cross metathesis conditions with 25 mol % catalyst 3. The reaction was run 1 with an olefin concentration of 1 mM, and the reaction was analyzed by H NMR. Because the reaction was dilute, it had to be quenched and concentrated before analysis. b-BAp3 Yield(%) Dimer 25 + 65 Table 3.3: Influence ofboronate ester template on cross metathesis. Reactions were run with lmM 7octene-1 ,2-diol and 25 mol % 3 at 40 °C. 59 This experiment demonstrated the applicability of the boronate ester motif to the template-directing system. Complexation of 7-octene-1,2-diol to b-BAp3 prior to the metathesis reaction increased the yield of dimer by more than 2.5 times (Table 3.3). Template-Directed ADMET: ADMET dimerization reactions were run with b-BAp3 and 1,3-diol-diene and analyzed by HPLC. The results from the templated reactions were not as selective as those seen in the cross metathesis reactions. Specifically, the dimerization yields were consistently as high without template as with. This was unexpected given the high association constant ofboronate esters. However, the template did influence the distribution of products. The reaction without template yielded exclusively a cyclic dimer, while the reaction with b-BAp3 yielded a mixture of linear and cyclic dimers (Table 3.4). Attempts to increase the yield in the reaction with template - longer reaction time and using different metathesis catalysts - were not successful. b-BAp3 + Yield(%) Linear Dimer Yield(%) Cyclic Dimer Overall Yield (%) 0 56 56 15 22 37 Table 3.4: Influence of peptide template of AD MET dimerization of 1mM 1,3-diol-diene. The possibility was investigated that the high background reaction was due to an association between the diol functionality of the free monomer and the catalyst. In a reaction without template, p-phenylboronic acid (PBA) was complexed to the monomer prior to the metathesis reaction to block the interaction. However, addition of PBA had no effect on the background reaction. 60 On-Resin Template-Directed ADMET: Using the conditions previously established, a boronate ester complex was formed between MBHA-b-BAp 3 and 1,3-diol-diene. Analysis of the elution solution from the on-resin ADMET dimerization with this complex showed an 83% yield of cyclic dimer (Fig. 3.3). In addition, the solution from the metathesis step of the reaction (i.e., before boronate ester hydrolysis) showed a large catalyst peak, but no peaks from diolcontaining compounds. Interestingly, unlike the template-directed reactions in solution, the reaction on solid support yielded only cyclic dimer. The yield was significantly higher than that seen for the reactions in solution. The reaction on-resin also appeared to be much cleaner - with fewer side products - than those in solution. 1.20 Cyclic Dimer - Metathesis Solution Resin Elution Solution Catalyst 1 00 0 80 040 Linear Dimer Monomer 0.20 000 ? l 35.00 36.00 37 .00 38.00 39 .00 40 00 41 00 42.00 43 .00 44 .00 45 00 46.00 47 00 48. 00 49 .00 50.00 Minutes Figure 3.3: HPLC analysis ofMBHA-b-BAr 3 template-directed AD MET dimerization of 1,3-dioldiene. The effective template-directed metathesis dimerization using on-resin bistemplates did not transfer to the tris-templates, MBHA-t-BAp 5 and Tenta-t-BAp5 . In general, the recovery of the diol-containing compounds was impaired compared to that with the bis-templates was lower and trimer products could be identified. 61 Across trials, the reactions with MBHA-t-BAps were reproducible, giving a similar HPLC profiles each time. In those reactions there was always a significant amount of 1,3-diol-diene monomer and a new large peak at eluted after 30 minutes (Fig. 3.4a). While this seemed like an appropriate retention time for a cyclic trimer product, the identity of this peak could not be confirmed by MS. The reactions with Tenta-t-BAps were more variable. In some reactions a significant amount of 1,3-diol-diene monomer remained and in others a high yield of cyclic dimer was seen (Fig. 3.4b ). However, no new peak of significant size was seen in any of the reactions with Tenta-t-BAp5 . b) a)_ ···i tclic Trimer? :Cyclic Dimer f1Monomer Cyclic Dimer i' f .'· I -l, i 0:-1 Linear Dimer I. -l, Monomer r, LJvi\. __~ :<-r l, ; •' -- ~ Figure 3.4: Examples of HPLC analyses of template-directed ADMET trimerization of 1,3-diol-diene with a) MBHA-t-BArs and b) Tenta-t-BArs- III. Conclusions As with the pseudorotaxane recognition motif, the boronate ester motif showed a significant template effect for metathesis dimerization, but not for trimerization. Efforts to reduce the level of background and side reactions by using a resin-attached template improved the dimerization reactions with a bis-template but did not yield trimers when used with tris-templates. 62 IV. Experimental Fmoc-L-(4-Boronophenyl)alanine: A solution ofFmoc N-hydroxysuccinimide ester (334.1mg, 0.99 mmol) in 3.0 mL dioxane was added to a solution ofL-(4Boronophenyl)alanine (207.0 mg, 0.99 mmol) in 1.9 mL 10% Na2CO 3 (aq) solution. The reaction was stirred overnight. It was then diluted with water and washed with diethyl ether (3x). The pH of the aqueous layer was adjusted to 3 with IM HCI, which caused a white precipitate to form. The precipitate was removed by filtration, dissolved in EtOAc, dried over Na2SO4 and concentrated to a translucent film. The material was dissolved in MeOH/EtOAc, and the insoluble white crystals were removed by filtration. Concentration of the solution yielded 368.3 mg of product as a white solid (86%): 1H NMR (CD3OD) 8 7.77 (4H, d), 7.57 (2H, d), 7.23-7.43 (6H, m), 4.45 (IH, dd, J= 4.8, 9.9 Hz), 4.28 (IH, dd), 4.10-4.4.23 (3H, m), 3.23 (IH, dd, J= 4.8, 14.1 Hz), 2.94 (IH, dd, J = 9.9, 13.8 Hz). Peptide Syntheses: The peptides for b-BAp3 and t-BAp 5 were synthesized using standard Fmoc-solid phase synthesis techniques. After cleavage, the peptide was precipitated from the crude orange oil with chilled ether and then was further purified by RPHPLC. b-BAp 3 : The reaction yielded 16.0 mg (0.0312 mmol) of peptide as a white powder (20%). t-BAp 5 : The reaction yielded I 1.0 mg of peptide as an off-white powder (17%). Pentafluorophenyl acetate: A solution ofpentafluorophenol and acetic anhydride (0.5 eq.) in CH3CN was added to a solution of DCC (0.5 eq.) in CH3CN under anhydrous conditions. After stirring for 2 hours at rt, the product was isolated by silica gel column chromatography to yield 121 mg of a white solid (67%). 1,3-diol-diene: The ADMET-acid compound (0.0782 mmol) was dissolved in anhydrous DCM. EDC (2.0 eq.) and DMAP (0.1 eq.) were added to the solution. After stirring for IO minutes, a solution of 2-hydroxymethyl-1,3-propanediol (2.0 eq.) in DMF was added. After stirring for 3 hours at r.t., the product was isolated by column chromatography in I: I EtOAc:Hex (35%). 63 V. References 1. Springsteen, G., Wang, B., Tetrahedron, 2002, 58, 5291-5300. 2. Dale, J., J. Chem. Soc., 1961, 922-&. 3. Bachelier, N.; Verchere, J. F., Polyhedron, 1995, 14, 2009-2017. 4. Yang, W. Q.; Gao, X. M.; Springsteen, G.; Wang, B. H., Tetrahedron Lett., 2002, 43, 6339-6342. 5. Gravel, M.; Thompson, K. A.; Zak, M.; Berube, C.; Hall, D. G., J. Org. Chem., 2002, 67, 3-15. 6. Carboni, B.; Pourbaix, C.; Carreaux, F.; Deleuze, H.; Maillard, B., Tetrahedron Lett., 1999,40, 7979-7983. 7. Armstrong, S. K., J. Chem. Soc., Perkin Trans. 1, 1998, 371-388. 8. Shelkov, R.; Nahmany, M.; Melman, A., Org. Biomol. Chem., 2004, 2, 397-401. 9. Kisfaludy, L.; Mohacsi, T.; Low, M.; Drexler, F., J. Org. Chem., 1979, 44, 654-656. 64