An Evolved Methanomethylophilus alvus Pyrrolysyl-tRNA Synthetase/tRNA Pair Is Highly Active and Orthogonal in Mammalian Cells.

We recently characterized a new class of pyrrolysyl-tRNA synthetase (PylRS)/PyltRNA pairs from Methanomassiliicocales that are active and orthogonal in Escherichia coli. The aminoacyl-tRNA synthetases (aaRSs) of these pairs lack the N-terminal domain that is essential for tRNA recognition and in vivo activity in the Methanosarcina mazei ( Mm) PylRS but share a homologous active site with MmPylRS; this facilitates the transplantation of mutations discovered with existing PylRS systems into the new PylRS systems to reprogram their substrate specificity for the incorporation of noncanonical amino acids (ncAAs). Several of the new PylRS/PyltRNA pairs, or their evolved variants [including Methanomethylophilus alvus ( Ma) PylRS/ MaPyltRNA(6)CUA], are mutually orthogonal to the MmPylRS/ MmPyltRNA pair, and the active sites of the Mm pair and Ma pair can be diverged to enable the incorporation of distinct ncAAs in response to distinct codons via orthogonal translation in E. coli. Here we demonstrate that MaPylRS/ MaPyltRNA(6)CUA is orthogonal to the aaRSs and tRNAs in mammalian cells and directs efficient incorporation of ncAAs into proteins. Moreover, we confirm that the MaPylRS/ MaPyltRNA(6) and MmPylRS/ MmPyltRNA pairs are mutually orthogonal in mammalian cells and demonstrates that these pairs can be used to encode distinct ncAAs into a protein in mammalian cells. Thus, the MaPylRS/ MaPyltRNA(6)CUA pair provides an additional pair that is orthogonal in both E. coli and mammalian systems and is mutually orthogonal to the most widely used system for genetic code expansion. Our results provide a foundation for expanding the scope of genetic code expansion and may also facilitate strategies for proteome-wide ncAA tagging with mutually orthogonal systems.

Genetically encoding the site-specific co-translational incorporation of noncanonical amino acids into proteins in eukaryotic cells and animals has provided numerous strategies for imaging and controlling the functions of proteins in their native environment.[1] Extensions of these approaches have enabled the tagging and labeling of cell-specific proteomes via stochastic orthogonal recoding of translation (SORT).[2−5]

The incorporation of ncAAs into proteins relies on the development of orthogonal aminoacyl-tRNA synthetase (aaRS)/tRNA pairs: the orthogonal aaRS selectively recognizes its cognate orthogonal tRNA over endogenous tRNAs, and the orthogonal tRNA is a substrate for the orthogonal aaRS but a poor substrate for endogenous synthetases. Because the sets of endogenous synthetases and tRNAs differ between organisms, aaRS pairs that are orthogonal in one system are commonly not orthogonal in another.[1] For example, the Methanocaldococcus janaschii (Mj) TyrRS pair that has been extensively used for genetic code expansion in Escherichia coli (Ec) cannot be used for genetic code expansion in eukaryotic cells because it is not orthogonal with respect to endogenous eukaryotic aaRS/tRNA pairs.

The EcTyrRS/TyrtRNA[6,7] pair and EcLeuRS/LeutRNA pair[8] are orthogonal in eukaryotic cells, and variants of these pairs have been discovered, primarily by directed evolution in yeast or subsequent screening, that enable the incorporation of a range of ncAAs in eukaryotic systems. The PylRS/PyltRNA pair from Mm is commonly considered an ideal pair for genetic code expansion because it is orthogonal in both E. coli and eukaryotic cells and animals.[1] This has facilitated the discovery and characterization of MmPylRS variants that incorporate ncAAs in E. coli and the transfer of these variants to eukaryotic systems, thereby facilitating genetic code expansion in eukaryotic cells and animals. Recent work has demonstrated that an evolved SepRS/v1.0/SeptRNAv1.0 pair enables the efficient incorporation of phosphoserine and its nonhydrolyzable analogue,[9] can be further evolved to incorporate phosphothreonine in E. coli,[10] and is also orthogonal in mammalian cells.[11] The ability to incorporate ncAAs into proteins in mammalian cells has been further expanded by strategies that replace the genomically encoded EcTrpRS/TrptRNA pair in E. coli with the Saccharomyces cerevisiae (Sc) TrpRS/TrptRNA pair.[12−14] Because the ScTrpRS/TrptRNA is orthogonal in E. coli, suppressor derivatives of the EcTrpRS/TrptRNA pair can be introduced into the resulting E. coli strains and evolved for ncAA incorporation. The resulting EcTrpRS/TrptRNA pairs can then be used for genetic code expansion in mammalian cells, where they are orthogonal. Recent work has extended this strategy to the EcTyrRS/ TyrtRNA pair.[15] In several cases, aaRS/tRNA pairs that are active and orthogonal with respect to the endogenous aaRSs and tRNAs in mammalian cells have also been shown to be orthogonal with respect to other orthogonal pairs, creating “mutually orthogonal” pairs in mammalian cells.[16−18]

We recently discovered that a new class of PylRS/PyltRNA pairs from Methanomassiliicocales are active and orthogonal in E. coli.[19,20] These pairs, unlike the MmPylRS/MmPyltRNA pair, lack the N-terminal domain of PylRS, which was previously thought to be essential for tRNA recognition and aminoacylation.[21,22] We showed that certain orthogonal pairs from this class are naturally mutually orthogonal to the MmPylRS;[19] this surprising result demonstrated that there is sufficient divergence between archeal PylRS/PyltRNA pairs to generate mutually orthogonal pairs for the same amino acid within a domain of life. We developed a number of exceptionally active and orthogonal pairs, including the MaPylRS/MaPyltRNA(6) pair, that are mutually orthogonal to the MmPylRS/MmPyltRNA pair by virtue of mutations introduced into the body of MaPyltRNA.[19] Moreover, we showed that, because the active sites of MmPylRS and MaPylRS share common substrate recognition determinants, we could transplant mutations that direct the selective incorporation of specific ncAAs from the MmPylRS active site to the MaPylRS active site to reprogram its substrate specificity.[19] Finally, we showed that by diverging the active sites of MaPylRS and MmPylRS to selectively recognize distinct substrates and altering the anticodons of MaPyltRNA(6) and MmPyltRNA to decode distinct codons in orthogonal translation, we could use these pairs in the same cell to direct the incorporation of two distinct ncAAs into a single polypeptide.[19] Here we show that the MaPylRS/MaPyltRNA(6) pair is highly active and orthogonal in mammalian cells, where it is also mutually orthogonal to the MmPylRS/MmPyltRNA pair (Figure ), and that derivatives of the two pairs can be used together to incorporate distinct amino acids into a protein in mammalian cells.

Figure 1

The evolved MaPylRS/MaPyltRNA(6) is orthogonal with respect to the endogenous E. coli aaRS/tRNA pairs as well as mutually orthogonal to the MmPylRS/MmPyltRNA(6) pair. Here we ask if this pair is also orthogonal and mutually orthogonal in mammalian cells. NTD (N-terminal domain).

We first demonstrated that the MaPylRS/MaPyltRNA(6)CUA pair is active in mammalian cells and that both MaPylRS and MaPyltRNA(6)CUA are orthogonal with respect to the endogenous tRNAs and the aaRSs in human cells. To this aim, we cloned coding sequences of MaPylRS and MaPyltRNA(6)CUA into a vector for mammalian expression;[23] we cloned MaPylRS under the EF1a promoter and four copies of the MaPyltRNA(6)CUA under the human U6 promoter into the same vector. We transiently co-transfected HEK293 cells with the MaPylRS/MaPyltRNA(6)CUA vector and an mCherry-TAG-GFP reporter[11] and cultured the resulting cells in the presence or absence of Nε-[(tert-butoxy)carbonyl]-l-lysine (BocK), a known substrate for this pair.[19] We measured the ratio of GFP to mCherry fluorescence by flow cytometry (Figure a and Figure S1) and fluorescence microscopy (Figure S2) and performed control experiments with the well-characterized MmPylRS/PyltRNACUA pair.

Figure 2

The MaPylRS/MaPyltRNA(6)CUA pair is active and orthogonal in mammalian cells. (a) MaPylRS and MmPylRS show comparable, BocK-dependent, readthrough of the amber stop codon with their cognate tRNAs. Data represent means ± the standard deviation from two biological replicates. (b) ESI-MS of purified sfGFP confirms quantitative incorporation of BocK via the MmPylRS/MmPyltRNACUA pair. (c) ESI-MS of purified sfGFP confirms quantitative incorporation of BocK via the MaPylRS/MaPyltRNA(6)CUA pair.

We observed minimal readthrough of the amber codon by the MaPylRS/MaPyltRNA(6)CUA pair in the absence of BocK (Figure a). This demonstrates that MaPyltRNA(6)CUA is orthogonal with respect to the aminoacyl-tRNA synthetases that are endogenous in human cells. Upon addition of 0.5 and 1 mM BocK, we observed substantial readthrough of the TAG codon by the MaPylRS/MaPyltRNA(6)CUA pair. The level of amber codon readthrough mediated by the MaPylRS/MaPyltRNA(6)CUA pair is comparable to that mediated by the highly active MmPylRS/MmPyltRNACUA pair (Figure a); this demonstrates that the MaPylRS/MaPyltRNA(6)CUA pair is highly active in mammalian cells.

To demonstrate that MaPylRS is functionally orthogonal in human cells, we co-transfected plasmids encoding the MaPylRS/MaPyltRNA(6)CUA pair and GFP(150TAG)His6 and cultured the cells in the presence of 1 mM BocK. ESI-MS of the resulting GFP gives the expected mass (Figure b) and is indistinguishable from a control in which we used the MmPylRS/MmPyltRNACUA pair to incorporate BocK into GFP(150TAG)His6 (Figure c). Taken together, our experiments reveal that the MaPylRS/MaPyltRNA(6)CUA pair is a highly active and orthogonal pair in mammalian cells.

Next, we aimed to demonstrate that the MmPylRS/MmPyltRNACUA pair and MaPylRS/MaPyltRNA(6)CUA pair are mutually orthogonal in their aminoacylation specificity when expressed in mammalian cells. We swapped the 4× U6-tRNA cassette between the Mm and Ma expression vectors producing a plasmid containing the MmPylRS/MaPyltRNA(6)CUA pair and a plasmid containing the MaPylRS/MmPyltRNACUA pair. We co-transfected the mCherry-TAG-GFP reporter with each aaRS/tRNA combinations and cultured the cells in the presence and absence of BocK. We compared the readthrough of the amber stop codon by these noncognate pairs to that mediated by the Mm- and Ma-derived cognate pairs (Figure ). Our data show that the noncognate MmPylRS/MaPyltRNA(6)CUA and MaPylRS/MmPyltRNACUA pairs lead to minimal readthrough of the amber stop codon (Figure a), while the cognate pairs lead to efficient amber suppression (Figure ). This demonstrates that these two PylRS/PyltRNACUA pairs are mutually orthogonal in mammalian cells.

Figure 3

The MaPylRS/MaPyltRNA(6)CUA pair is mutually orthogonal to the MmPylRS/MmPyltRNACUA pair in mammalian cells. MaPylRS and MmPylRS show comparable, BocK-dependent aminoacylation of their cognate tRNA and minimal cross-aminoacylation of the noncognate PyltRNACUA. Data represent means ± the standard deviation from two biological replicates.

Next we differentiated the active sites of MaPylRS and MmPylRS such that they selectively recognize distinct substrates. By screening a collection of MaPylRS mutants for non-natural substrate specificity in E. coli, we discovered a variant of MaPylRS, MaPylRS(mut), that incorporates 3-methyl-l-histidine (Me-His) but not BocK; this synthetase contains L121M, L125I, Y126F, M129A, and V168F mutations. We find that MmPylRS directs the incorporation of BocK but not Me-His; these specificities are maintained in mammalian cells (Figure S3a). Finally, we performed a double incorporation of BocK and Me-His into GFP(101TGA,150TAG) using the MmPylRS/MmPyltRNACAU and MaPylRS(mut)/MaPyltRNA(6)CUA pairs. Production of full-length protein was dependent on addition of both ncAAs (Figure S3b), consistent with the site-specific incorporation of both amino acids into GFP in mammalian cells.

We have demonstrated that the MaPylRS/MaPyltRNA(6)CUA pair is orthogonal with respect to the synthetases and tRNAs present in mammalian cells and is highly active in mammalian cells. As the MaPylRS/MaPyltRNA(6)CUA pair is also orthogonal in E. coli and we have demonstrated that mutations discovered in the active site of MmPylRS or MbPylRS can be transplanted into MaPylRS to reprogram its substrate specificity,[19] it will be possible to rapidly expand the substrate scope of the MaPylRS/MaPyltRNA(6)CUA pair. Moreover, as MaPylRS is a single-domain protein that lacks the poorly soluble N-terminal domain of MmPylRS, it may be even more amenable to directed evolution in E. coli than the MmPylRS system. Our results demonstrate that MaPylRS mutants discovered and characterized in E. coli will be of direct utility in mammalian cells.

Finally, we have confirmed that the mutual orthogonality of the MmPylRS/MmPyltRNA pair and MaPylRS/MaPyltRNA(6) pair, which we have characterized in E. coli, is maintained in mammalian cells and shown that these pairs can be used together for (unoptimized) double incorporation. These are therefore the first mutually orthogonal pairs in which each pair is itself orthogonal in both E. coli and eukaryotic systems. We anticipate that this foundational advance will facilitate multiplexed proteome labeling.[2−5] Combinations of the advances reported herein, together with strategies for creating additional blank codons and increasing the efficiency of multisite ncAA incorporation in mammalian cells, may facilitate the site-specific incorporation of diverse, and currently inaccessible, combinations of ncAAs into proteins in mammalian cells.[16−18]