Evolution of synthetic polymers

Abstract

A strategy for the enrichment of a DNA template that encodes a functionalized PNA oligomer is discussed. The method relies on iterated cycles of chemical translation (of the template into PNA), selection (for function), and amplification (of the survivors). Potential restrictions and future perspectives are considered.

Throughout the last four billion years, nature afforded an enormous variety of different highly specialized and functionalized organisms. Each of them represents the result of a long lasting evolutionary process, in which the organism's biopolymers, namely proteins, nucleic acids and carbohydrates, survived all previous events of selection for functionality. Scientists have adopted the features of evolution to the laboratory in order to produce functionalized biopolymers such as DNA-aptamers or peptides.^1–2 Synthesis of a biopolymer library is followed by selection for binding affinity towards a target molecule and amplification of the survivors in recurring cycles. With this approach, rarely represented polymers from the library can be enriched over an acceptable time range without the need of knowing about specific binding interactions in detail. Quite the contrary, investigation of the resulting biopolymer-target complexes may give new insights into structural requirements for a successful binding event. Restricted by the need of using enzymes such as polymerases or ribosomes for the polymerization step, in vitro evolutionary methods have only been applied to biological polymers or slightly modified derivatives, since their monomers are the only accepted substrates (reviewed in ref. 3).

Brudno, Liu and coworkers recently published an approach that overcomes this constraint and exemplifies a strategy to accumulate a functionalized synthetic polymer such as a peptide nucleic acid (PNA) from a library of PNA oligomers by iterative cycles of chemical translation, selection and amplification.⁴ PNA was chosen because it fulfills several requirements needed for an in vitro evolution system. First, it binds with high affinity and in a sequence specific manner to its complementary DNA strand,⁵ which therefore can act as a template encoding the type and order of PNA building blocks being polymerized.⁶ Second, PNA is known to assume secondary and tertiary structure in its single stranded form making the naturally folded polymer a potential, aptamer-like candidate for interactions with biological target molecules.⁷ Furthermore, it has been shown that DNA-directed reactions between PNA-based oligomers proceed with high sequence fidelity.^8–12 Finally, PNA is not a substrate of any known enzyme and accordingly a biostable species, that can be successfully delivered into living cells making it a promising prospect for drug design.¹³

The crucial step is chemical translation of a DNA template sequence into a PNA sequence. The authors drew upon reductive amination, which has first been used by Lynn in the template-catalyzed oligonucleotide ligation,^14–16 to polymerize PNA-pentamer building blocks. They identified a set of twelve different PNA-pentamers, which showed to polymerize with similar efficiency on their encoding single stranded DNA templates. This is important, because the synthesized polymers should later on be picked by selection for functionality and not for synthetic discrimination or preference. It was found that the building blocks must agree in pyrimidine content as well as number of guanines or cytosines because these parameters correlate with melting temperatures of the corresponding heteroduplexes. A library containing ∼4.3 • 10⁸ different DNA template sequences was synthesized on a single column using repetitive mixtures of mono- and dinucleotide phosphoramidites. Each sequence encoded a 40 base pair segment with eight of twelve different consecutive codons (pentamers) in a row. The library contained a spiked-in positive control DNA-template 1, which was complementary to a unique biotinylated PNA-pentamer and represented statistically only each millionth oligomer. This template contained a singular recognition sequence for cleavage by the restriction enzyme MspI, which was introduced in order to enable specific detection of products that originated from chemical translation, selection and amplification of the positive control by enzymatic digestion and subsequent gel analysis. In the initial step, two hairpins were ligated to the 3′- and 5′-ends of the DNA templates catalyzed by T4 DNA-ligase (Fig. 1). The 5′-hairpins of 2 and the other library members have free amino groups that allow covalent attachment via C-terminal PNA aldehydes. In this event, the information encoded in the DNA part is linked with a certain phenotype expressed by the PNA part. The approach bears resemblance to mRNA display where RNA-template and translated polypeptide are covalently linked.¹⁷ In the presence of the 12 PNA-pentamers, the unique biotinylated PNA building block, and NaBH₃CN, template-instructed PNA-polymerization was carried out for 1 h. To displace the synthesized PNA-polymer in 3 and other non-biotinylated constructs from the coding strand and allow folding into secondary and tertiary structures, a special DNA-polymerase (Herculase II) was added. This polymerase was capable of stripping off PNA from PNA-DNA heteroduplexes by generating double stranded DNA. Displacement of the PNA is favored due to a gain in the number of Watson-Crick base pairs of the formed DNA-duplex, which is synthesized into the loop region of the 5′-hairpin. The reaction products such as 4 were subjected to selection for binding to streptavidin coated magnetic particles. Since only the positive control translates into a polymer carrying a biotinylated N-terminus 5, other polymers were washed away in this step. This procedure was followed by AvaI restriction digestion, releasing the DNA template 6 from the selected PNA. Subsequent template amplification was carried out by PCR using a biotinylated primer for the non-template strand to enable strand isolation after binding to streptavidin-coated beads. The non-biotinylated strand 1 was liberated into solution by NaOH-mediated denaturation and used for another round of chemical translation, displacement, selection and amplification. To monitor the fraction of positive controls in the amplification product after each cycle the library was analyzed by gel electrophoresis before and after restriction enzyme cleavage with MspI. The reduced molecular weight of products 7 processed by MspI distinguishes positive control products from other library members. After six iterations, gel electrophoresis proved the positive control to be the major component of the amplification product amounting to a more than 1,000,000-fold enrichment of this template in the library.

Figure 1.

A DNA-library containing a positive control in a 10⁻⁶:1 ratio that uniquely translates into a biotinylated PNA-pentamer (red) was subjected to 1. Hairpin-ligation mediated by T4 ligase, 2. Sequence specific PNA-polymerization (chemical translation), 3. Displacement of the PNA-polymer by DNA-polymerization retaining the covalent polymer-template linkage, 4. Selection for streptavidin-binding to magnetic particles, 5. Cleavage from the solid support by AvaI restriction digestion, 6. Amplification of the selection survivors by PCR using a biotinylated primer for the non-template strand, 7. Immobilization of the amplification products and 8. Strand separation by eluting the template strand with NaOH. The fraction of positive controls within the amplification products was monitored by MspI restriction digestion after each cycle.

This work is a proof of principle which demonstrates that a template encoding a functionalized synthetic PNA polymer can be enriched from a population of sequences in iterated rounds of chemical translation, selection and amplification. Nevertheless it remains to be shown that this concept also works for selection of PNA polymers that function via the properties of its sequence and not due to biotin-streptavidin recognition, one of the strongest noncovalent interactions in nature with a K_d in the femtomolar range.¹⁸ Even after strand displacement the PNA is still bound to a single stranded DNA of more than 20 nucleotides in length, that could induce other than native three dimensional folding structures of the polymer and influence the outcome of the selection event. Another potential restriction lies in the nature of the polymer that is synthesized. Up to now, only a few non Watson-Crick interactions are known for PNA with other biomolecules (example in ref. 19). On the other hand, the presented method may offer the fascinating opportunity to screen for such currently unknown interactions. It can be imagined that the method enables the discovery of new potent PNA-based binders and catalysts, especially with respect to the possibility of polymerizing side chain modified PNA building blocks that allow the introduction of other functional groups to the PNA backbone.^20–21

Commentary to: Brudno Y, Birnbaum ME, Kleiner RE, Liu DR. An in vitro translation, selection and amplification system for peptide nucleic acids. Nat Chem Biol. 2010;6:148–155. doi: 10.1038/nchembio.280.