Expansion of the genetic code via expansion of the genetic alphabet

Highlights

• An unnatural base pair has been developed and used as the foundation of semi-synthetic organisms.

• The semi-synthetic organisms can use the unnatural base pair to efficiently make unnatural proteins.

• Hydrophobic and packing interactions can replace hydrogen bonds to control base pairing.

Current methods to expand the genetic code enable site-specific incorporation of non-canonical amino acids (ncAAs) into proteins in eukaryotic and prokaryotic cells. However, current methods are limited by the number of codons possible, their orthogonality, and possibly their effects on protein synthesis and folding. An alternative approach relies on unnatural base pairs to create a virtually unlimited number of genuinely new codons that are efficiently translated and highly orthogonal because they direct ncAA incorporation using forces other than the complementary hydrogen bonds employed by their natural counterparts. This review outlines progress and achievements made towards developing a functional unnatural base pair and its use to generate semi-synthetic organisms with an expanded genetic alphabet that serves as the basis of an expanded genetic code.

Introduction

Biological diversity allows life to adapt to different environments, and over time, evolve new forms and functions. The source of this diversity is the variation within protein sequences provided by the twenty natural amino acids, variation that is encoded in an organism’s genome by the four natural DNA nucleotides. Although the functional diversity provided by the natural amino acids may be high, the vastness of sequence space dramatically limits what might actually be explored, and moreover, some functionality is simply not available. Nature’s use of cofactors for hydride transfer, redox activity, electrophilic bond formation and so on, attests to these limitations. Furthermore, with the increasing focus on developing proteins as therapeutics [1], these limitations are problematic, as the physiochemical diversity of the natural amino acids is dramatically restricted compared to that of the small molecule drugs designed by chemists. In principle, it should be possible to circumvent these limitations by expanding the genetic code to include additional, non-canonical amino acids (ncAAs) with desired physiochemical properties.

Almost 20 years ago, Peter Schultz increased the diversity available to living organisms by expanding the genetic code using the amber stop codon (UAG) to encode ncAAs in Escherichia coli [2^••,3^••]. This landmark accomplishment was achieved using a tRNA–amino acid tRNA synthetase (aaRS) pair from Methanococcus jannaschii, in which the tRNA was recoded to suppress the stop codon and the aaRS was evolved to charge the tRNA with an ncAA. This method of codon suppression has since been expanded to the other stop codons [4] and even quadruplet codons [5], as well as to the use of several other orthogonal tRNA–aaRS pairs (most notably the pyrrolysyl (Pyl) tRNA–synthetase pair from Methanosarcina barkeri/mazei [6,7,8^•,9]), broadening the scope of ncAAs that may be incorporated into proteins. These methods have already begun to revolutionize both chemical biology [10, 11, 12] and protein therapeutics [13].

Though these methods enable incorporation of up to two, different ncAAs in both prokaryotic [14] and eukaryotic [15] cells, the heterologous recoded tRNAs must compete with endogenous release factors (RFs), or in the case of quadruplet codons, normal decoding [16], which limits the efficiency and fidelity of ncAA incorporation. To eliminate competition with RF1, which recognizes the amber stop codon and terminates translation, efforts have been directed toward removal of many or all instances of the amber stop codon in the host genome [17,18] or modification of RF2 [19] to allow for the deletion of RF1. However, eukaryotes have only one release factor, and while it may be modified [20] it cannot be deleted, and with prokaryotes, deletion of RF1 results in greater mis-suppression of the amber stop codon by other tRNAs, which reduces the fidelity of ncAA incorporation [21]. Though Herculean efforts to further exploit codon redundancy to liberate natural codons for reassignment to ncAAs are underway [22], codon reassignment may be complicated by pleiotropic effects, as codons are not truly redundant, for example due to their effects on the rate of translation and protein folding [23]. In addition, codon reappropriation is limited by the challenges of large-scale genome engineering, especially for eukaryotes [24].

An alternative approach to natural codon reassignment is the creation of entirely new codons that are free of any natural function or constraint, and whose recognition at the ribosome is inherently more orthogonal. This may be accomplished through the creation of organisms that harbor a fifth and sixth nucleotide that form an unnatural base pair (UBP). Such semi-synthetic organisms (SSOs) would need to faithfully replicate DNA containing the UBP, efficiently transcribe it into mRNA and tRNA containing the unnatural nucleotides, and then efficiently decode unnatural codons with cognate unnatural anticodons. Such SSOs would have a virtually unlimited number of new codons to encode ncAAs.