Difference between revisions of "Team:Bielefeld-CeBiTec/Project/translational system/library and selection"

The incorporation of a non-canonical amino acid (ncAA) into the nascent polypeptide chain requires a tRNA/aminoacyl-tRNA synthetase (tRNA/aaRS) pair which is highly specific for each amino acid. Mechanistically the aaRS has to recognize, accept, and bind the amino acid and transfer it in the second step onto the tRNA. For our ncAA we want to develop a tRNA/aaRS pair which is specific for the ncAA as well as orthogonal to the canonical amino acid tRNA/aaRS pairs.Therefore a library of mutatedorthogonal tRNA/aaRS pairs is generated, based on the commonly used tyrosyl- and pyrrolysyl- tRNA/synthetases. For the adaption of the codon recognition and amino acid binding, the candidate aaRS undergoes numerous rounds of positive and negative selection. The selection results in a synthetase which can be expressed efficiently in Escherichia coli and is able to reliable incorporate specifically our ncAA.

There are several non-canonical amino acids which have been incorporated into peptides and proteins in E. coli . Among others, it was possible to incorporate ncAA with heavy atoms, keto and alkaline side chains and photo crosslinking capability(Zhang et al.,2005). For the binding of the ncAA, an adaption of the synthetase is necessary. This is done by introducing mutations in several regions of the associated synthetase, leading to a conformational change of the binding pocket and thus in the binding of a new amino acid. The tyrosine synthetase is the first orthogonal E. coli tRNA/aaRS pair generated from archaea and the best known so far for the incorporation of ncAAs (Wang et al,.2001). This tyrosine synthetase (TyrRS) has a small anticodon loop binding domain (Steer et al,.1999). Therefore, the change of the recognition in order to change the anticodon loop of its cognate tRNA to CUA is not as difficult as with a larger anticodon loop binding domain (Steer et al.,1999). The binding site is located deep inside a small pocket of TyrRS (Tian et al.,2004). Due to this it is assumed that already very small conformational changes result in a larger or smaller binding pocket with an altered binding interaction with the ligand. Therefore, in this protein, only few changes in the coding sequence should be necessary to incorporated ncAA. (Zhang et al.,2005). In detail, the TyrRS binding pocket is highly hydrophilic, preventing the binding of phenylalanine, whose structure is similar to the one of tyrosine (Goldgur et al.,1997). Changing the characteristic of the binding pocket to become more hydrophobic, could favor the binding of phenylalanine derivates (Goldgur et al.,1997). In the middle of the negative charged binding pocket (Zhang et al.,2005), there is an aspartic acid on position 158. The entrance of the binding pocket is positioned between two glutamic acids at positions 36 and 172. Upon tRNA binding, the aspartic amino acid Asp158 forms two hydrogen bonds with the glycine at position 34 (Zhang et al.,2005).

When the ligand tyrosine is bound, hydrogen bonds to Tyr32, Asp158, Glu36, Gln173, Tyr151 and Gln155 are formed, resulting in slight movements of the side chains within the tyrosine-binding pocket (Zhang et al.2005). Beside the binding pocket, other domains are also affected by the binding of tyrosine. For example, the loop 73-83, positioned at the entrance of the binding pocket, provides a hydrophobic lid over the binding pocket in reaction to the conformational changes. This is supposed to have the effect of separating the activated tyrosine from water during the catalytic reaction (Zhang et al.,2005).

In nature, one orthogonal aminoacyl-tRNA synthetase pair incorporates the 22^nd proteinogenic amino acid pyrrolysine (Pyl). Interestingly, Pyl is assigned to the amber stop codon (UAG). Thus, it is incorporated naturally through an endogenous nonsense codon suppression system(Burke et al., 1998; Paul et al., 2000). It was discovered in the archaeal order of Methanosarcinales (Srinivasan et al., 2002). Due to this specific property, the aminoacyl-tRNA synthetase/tRNA^Pyl pair can be used to expand the genetic code. The pyrrolysyl synthetase (PylRS) is a class II synthetase with a highly conserved class II catalytic core at the C-terminus of around 270 amino acids with a Rossmann fold for ATP-binding. PylRS forms an obligate dimer with an active site in each subunit (Kavran et al.,, 2007; Nozawa et al., 2009). Interestingly, the N-terminus of the synthetase shows no significant similarity to any known protein domains and its length is variable. The N-terminus leads to aggregation of full-length PylRS due to its high insoluble properties (Yanagisawa et al., 2006; Jiang and Krzycki, 2012). Therefore a truncated version containing the catalytic core of the synthetase is used (residues 185-454) (Kavran et al., 2007; Yanagisawa et al., 2008). In Methanosarcina mazei the catalytic core of PylRS has a deep hydrophobic pocket where A302, L305, Y306, L309, N346, C348 and W417 form a large cavity for the binding of the Pyl side chain ((4R,5R)-4-methyl-pyrroline-5-carboxylate) (Kavran et al., 2007; Yanagisawa et al., 2008a). These hydrophobic interactions are relatively non-specific. Given that PylRS does not have an editing domain for hydrolysation of misacylated tRNAs, this synthetase shows high substrate side chain promiscuity (Wan et al., 2014). On top of that, Pyl is quite large in comparison to the other endogenous canonical amino acids, thus there was no selection pressure for PylRS to stringently recognize Pyl(Wan et al., 2014). Therefore structurally similar and dissimilar non-canonical amino acids were shown to be incorporated using PylRS (Mukai et al., 2008; Polycarpo et al., 2006; Yanagisawa et al., 2008b). The residues Y384 and N346 in the active site of PylRS form hydrogen bond interactions with the α-amine of Pyl (Yanagisawa et al., 2008a; Kavran et al., 2007). But PylRS also has a high tolerance for α-hydroxy acids (Kobayashi et al., 2009; Li et al., 2012).

Most interestingly, PylRS does not show a direct synthetase-anticodon interaction as other synthetases do. Even a mutation at the anticodon of tRNA^Pyl does not influence aminoacylation activity of PylRS (Ambrogelly et al., 2007; Yanagisawa et al., 2008a). This characteristic can be used to incorporate non-canonical amino acids using unnatural base pairs. The only drawback using PylRS is its low translational efficiency and catalytic activity compared to other synthetases (Guo et al., 2014). Because in nature PylRS is rarely and selectively expressed, a high catalytic and translational efficiency is not needed (Srinivasan et al., 2002). Every change in the native PylRS does lower the catalytic and translational activity (Guo et al., 2014).

As shown, it is possible to tune tRNA synthetases into incorporating ncAA. We want to achieve this adaptation process by creating a library of tRNA synthetases and obtain the tRNA synthetase accepting our ncAA by a selection mechanism. To pass as many mutated synthetases as possible through the cycle of the two selections for the adaption of the specific ncAA, a large library of different synthetase versions has to be created. The large library of random variants of sequences can be generated by side specific mutagenesis using randomized primers with the NNK scheme to avoid the incorporation of all stop codons, but one (Sieber et al.,2015). N codes for the bases A,C,G,T and K codes for the bases G and T. This leads to the use of 32 codons, canonical amino acid as well as one stop is encoded through the NNK scheme (Yuval et al.,2011). By comparison to other randomization, the NNK scheme has a relatively low ratio between the most common and rarest coded amino acid (3/32 vs. 1/32) (Yuval et al.,2011). By the randomization of more than one position, numerous more sequence variants are possible. If one position is randomized, 130 variants of the sequences are needed to attain a 0.99 probability of discovering the best variant (Yuval et al.,2011). At the same time, for the randomization of three positions, there are statistically 102,478 variants needed to discover the best variant. In this context, the probability of full codon coverage changes from 0.82 % for one randomizes position to 3.25 • 10^-36 % for three randomized position (Yuval et al.,2011). Regarding the fact that not the full coverage is essential, but rather the discovery of the best variant, the NNK scheme is a solid method for the generation of a synthetase library for the selection of the best ncAA binding site.