|Home | About | Journals | Submit | Contact Us | Français|
Combinatorial libraries of peptoids (oligo-N-substituted glycines) have proven to be useful sources of protein ligands. Each unit of the peptoid oligomer is derived from 2-haloacetic acid and a primary amine. In order to increase the chemical diversity available in peptoid libraries, we demonstrate here that heterocyclic halomethyl carboxylic acids can be employed as backbone building blocks in the synthesis of peptoid-based oligomers. Optimized conditions are reported that allow the creation of large, high quality combinatorial libraries containing these units.
There is great interest in the discovery of potent and selective bioactive compounds for use in medicine and biology. In recent years, high-throughput screening has been a major contributor to the discovery of such species. These screening campaigns are carried out in a completely unbiased fashion if no structural information for the target protein is available to guide library design. In such cases, the more chemically diverse the library one screens, the better. One source of highly diverse peptidomimetic compounds is libraries of peptoids (N-substituted oligo-glycines)1. Peptoids have certain advantages over native peptides, including protease insensitivity and improved cell permeability2–3.
Each unit of a peptoid is put together in two steps4, the first being acylation of an existing amine with an activated ester of 2-bromo (or chloro5) acetic acid, followed by displacement of the halide with a primary amine. These are almost ideal molecules for combinatorial chemistry6, because the source of diversity is a primary amine, of which thousands are readily available. In addition, the structures of peptoids can be determined by tandem mass spectrometry, meaning that split and pool strategies7 can be used to create peptoid libraries without the requirement for encoding.
It would be desirable to develop peptoid analogues that retain the many favorable characteristics of this class of compounds but offer increased chemical diversity and greater conformational constraints. One strategy to accomplish this goal would be to replace the two-carbon haloacetic acid-derived fragment of peptoids with a more elaborate building block 8–9. Aromatic heterocycles are of particular interest as main chain building blocks. In addition to expanding the functionality in the main chain relative to a peptoid, these units would be expected to introduce conformational constraints in the oligomer. Indeed, two previous reports indicate that incorporation of a 1,5-disubstituted-1,2,3-triazole10 and a 1,3,5-triazine unit11 in the peptoid backbone resulted in conformationally biased oligomers. Moreover, heterocycles can also increase the aqueous solubility of oligomers relative to all-carbon units 8,12. The critical role of heterocycles in drug discovery is underscored by the fact that more than half of known small molecule drugs contain at least one heterocycle. Here we report novel heterocyclic halomethyl carboxylic acids that can be used in place of the halo acetate, allowing the creation of peptoid analogues in which a heterocycle is incorporated into the main chain. Optimized conditions for the creation of high quality combinatorial libraries are presented. Furthermore, we also show that molecules derived from a single bead of a one bead one compound library of heterocycle-modified peptoids can be characterized by mass spectrometry.
Our initial efforts were focused towards incorporation of oxazole and thiazole units in the peptoid backbone, deriving inspiration from variety of natural products that display these entities. Biosynthetic assembly of these heterocycles occurs via various post-translational modifications on serine, threonine and cysteine residues in their peptide precursors. Synthetic routes have been reported for the generation of oxazole and thiazole amino acid building blocks for combinatorial synthesis13–15 and several strategies to assemble these scaffolds on solid phase have also been reported 16. However none of these strategies are applicable to a sub-monomer, peptoid-like approach to combinatorial library synthesis.
Therefore, we created heterocycles-containing haloacids 1 & 2 (Figure 1). These heterocycles were synthesized using a modification of a previously published procedure17 as described in Figure 1. Briefly, serine or cysteine methyl ester hydrochloride was condensed with dichloroacetonitrile in the presence of sodium methoxide. The resulting oxazoline or thiazoline intermediate was allowed to undergo base-catalyzed elimination to yield the corresponding chloromethyl-azole methyl esters. Saponification at low temperature followed by acidification yielded the respective carboxylic acids. Synthesis of these heterocycles was performed on a 45 mmol scale with ~ 85% and ~ 45% overall yields for the thiazole and the oxazole building blocks respectively.
With these building blocks in hand, we turned to deriving optimized conditions for inserting them into a peptoid chain. A five-residue peptoid (Figure 2) was assembled on polystyrene resin functionalized with the rink amide linker. We examined several conditions for carboxylic acid coupling and amine displacement (summarized in Table 1). In each case, the resin beads were subjected to a chloranil test18 to determine the completeness of the coupling reaction. Coupling of halomethyl oxazole and thiazole carboxylic acids at room temperature for three hours with DIC as the coupling agent, both in the absence and presence of additives such as HOAt were found to be inefficient. Subsequently, we tested the utility of standard microwave-assisted peptoid synthesis conditions19 (1M acid + 1M DIC in DMF with irradiation at 100W for 15 seconds) for the coupling of azole heterocycles 1 and 2 to the N-terminus of this chain. Unfortunately, these conditions also resulted in inefficient coupling as indicated by the chloranil test. The sluggish nature of this coupling can be attributed to the much lower reactivity of the activated esters of these carboxylic acids relative to those obtained from bromo- or chloroacetic acid. Optimal coupling of these heterocyclic carboxylic acids could only be achieved by performing the DIC-mediated couplings at 70 °C, employing 10 equivalents each of 1 or 2 along with 12 equivalents of DIC for a period of five hours. Performing the same reaction within a temperature-controlled microwave reactor (Biotage Inc.) at 80 °C for 20 minutes with 2 minutes of pre-activation was found to decrease the reaction time considerably.
The amine coupling reaction also proved to be more difficult than is the case with standard peptoid synthesis. This is not surprising given the reduced electrophilicity of the chloro-substituted carbons in the heterocyles relative to a carbon adjacent to carbonyl center. Nonetheless, we found that essentially quantitative displacement of the chloride could be achieved by treating the beads with primary amine (1M in DMF) at 80 °C for 12 hours or at 100°C for 30 minutes in the microwave reactor. In general, microwave-assisted couplings and amine displacement reactions generated the resulting oligomers with a greater purity in comparison with thermal conditions. After an additional standard peptoid residue was added to the chain as shown in Figure 2, the molecules were cleaved from the resin and analyzed by HPLC and MALDI mass spectrometry (Figure 3).
We then wanted to evaluate our ability to insert these heterocyclic building blocks into oligomeric chains in the context of one bead one compound (OBOC) libraries created by split and pool synthesis where the library members can still be sequenced by tandem MS unambiguously. The MS sequencing profiles of representative azole-modified peptoids, shown in Figure 3, exhibit a fragmentation pattern somewhat different from the fragmentation trends observed with regular peptoids and peptides. We observed that fragmentation of the amide bond at the C-terminal side of the heterocycle was suppressed, resulting in a complete absence or a low intensity of the corresponding y and b fragments resulting from that bond breakage. This unusual fragmentation trend can potentially complicate the sequencing of these oligomers. This means that the heterocycle-containing monomer and the one on the C-terminal side of it must be read as a single unit in the MS/MS spectrum. Thus, one must employ a different set of amines with distinct molecular weights to functionalize the azole sub-monomer in order to assign unequivocally the side chains on the heterocycles unit and the one before it from the fragmentation spectrum. Bearing this in mind, we synthesized a library of heptamers containing a constant four-residue peptoid linker appended to a peptide dimer constructed by cysteine and 6-aminohexanoic acid at the C-terminal. The variable three residues were diversified by using four different amines where the sixth position was further diversified by substituting oxazole or thiazole sub-monomer in place of the usual bromoacetic acid, thus resulting in a library of 128 compounds (Figure 4). The oxazole and thiazole-containing units were inserted using the optimized conditions mentioned above while all of the peptoid residues were inserted using standard microwave-assisted conditions. Twenty beads were picked at random in a 96-well plate for quality analysis of this library and each bead was subjected to cleavage with trifluoroacetic acid. The post-cleavage MS analysis displayed clean mass spectra for all compounds. MS-MS fragmentation of the major peak obtained in each case enabled the sequencing of all oligomers obtained from aforementioned 20 beads, where seven sequences appeared more than once. The design template, amine building blocks utilized for each variable position and the mass spectral profile of representative members of this library are shown in Figures 4 and 55.
This encouraging outcome of incorporation of oxazole and thiazole building blocks into the peptoid backbone prompted us to develop other heterocyclic scaffolds. With 2,4-disubstituted azoles in hand that orient the carboxamide and amine substitution in a 1,3 configuration in the peptoid backbone, we wanted to develop heterocyclic building blocks that can be used to exhibit 1,4 and 1,2 connectivities. To this end, we synthesized the pyrazine 3 and furan 4 building blocks as shown in Figure 6. These building blocks were synthesized from inexpensive commercially available methyl substituted carboxylic acids via NBS-mediated free radical bromination. Briefly, 5-methylpyrazine-2-carboxylic acid was converted to its corresponding methyl ester that was subsequently treated with NBS in presence of AIBN as the radical initiator. The corresponding bromomethyl carboxylate ester was saponified at low temperature to generate the required pyrazine building block 3 in an overall yield of ~ 43% from the starting carboxylic acid. We envisioned utilizing the same route for generating the furan building block 4. However methyl 2-bromomethyl-3-furoate20 could not be saponified at ambient temperature, and we feared that this intermediate would be hydrolyzed to the corresponding alcohol at higher temperatures. Instead, this building block was obtained directly by radical bromination of 2-methyl-3-furoic acid with a yield of ~ 53%. The efficacy of these new building blocks as sub-monomers for incorporation into the peptoid backbone was investigated by utilizing the microwave-assisted conditions established for azoles earlier. The HPLC profiles of the resulting oligomers indicated that the microwave-assisted acylation and amine displacement conditions developed earlier for the azole building blocks can be utilized to incorporate the pyrazine and furan building blocks efficiently as well.
The synthesis of four heterocyclic halomethyl carboxylic acids and establishment of conditions that can be employed to append these building blocks onto a growing peptoid chain expands our ability to synthesize fairly diverse peptoid-derived libraries. Design of a complex library that showcases the incorporation of more than one thiazole, pyrazine and furan building blocks in the backbone is shown in Figure 7. Bromoacetic acid was also included as a building block at positions where heterocycles were incorporated to increase the overall theoretical diversity to 16,384 compounds in this library. This library was assembled on polystyrene Rink amide resin (500 microns, Rapp Polymere) and the overall quality of this library was assessed by cleaving the oligomers from twenty four beads in a 96-well format as described earlier. Out of these, compounds obtained from twenty one beads were sequenced unambiguously, indicating that roughly 87% of this library will yield the desired sequences. The MS spectra for representative members are shown in Figure 8.
Diverse oligomeric libraries that display a variety of backbone scaffolds and side-chain appendages are a promising source of bioactive ligands. Generation of such libraries in high quality is often challenging, primarily because the chemical reactions utilized to generate diverse scaffolds are unable to generate the resulting oligomers in high purity21. In this publication, we have reported the development of heterocyclic halomethyl carboxylic acids as building blocks for generation of diverse backbone scaffolds in peptoid-derived oligomers. We have also demonstrated that these halomethyl heterocyclic carboxylic acids can be conveniently utilized to construct peptoid-based libraries via a sub-monomer approach through microwave-assisted amide bond formation and subsequent nucleophilic substitution with primary amines. Furthermore, this strategy has been utilized to generate peptoid-derived libraries in the OBOC format where the library members were sequenced by tandem mass spectrometry. To expand the scope of this approach, development of other heterocyclic building blocks is currently in progress that can further enhance the backbone diversity of oligomeric libraries. These OBOC libraries will be screened with different biological targets to mine for structurally diverse ligands.
A full description of the materials and methods can be found in the Supplementary Material.
This work was supported by a grant from the National Institutes of Health (1RO1 GM 090294) and a contract from the National Heart Lung and Blood Institute (NO1-HV-00242).