Ribosome Subunit Stapling for Orthogonal Translation in E. coli

The creation of orthogonal large and small ribosomal subunits, which interact with each other but not with endogenous ribosomal subunits, would extend our capacity to create new functions in the ribosome by making the large subunit evolvable. To this end, we rationally designed a ribosomal RNA that covalently links the ribosome subunits via an RNA staple. The stapled ribosome is directed to an orthogonal mRNA, allowing the introduction of mutations into the large subunit that reduce orthogonal translation, but have minimal effects on cell growth. Our approach provides a promising route towards orthogonal subunit association, which may enable the evolution of key functional centers in the large subunit, including the peptidyl-transferase center, for unnatural polymer synthesis in cells.

The ribosome is al arge molecular machine,u niversally composed of two subunits,t hat decodes non-overlapping triplet codons in mRNAs for the encoded polymerization of amino acids into proteins. [1] Thesmall subunit, containing 16S rRNA, binds mRNAa nd decodes the interaction between codons on mRNAs and their cognate tRNAa nticodons,a nd the large subunit, containing 23S rRNA, facilitates many functions,i ncluding peptide bond formation. While natural translation encodes the polymerization of the canonical 20 amino acids,e xtensions of translation for the polymerization of unnatural building blocks will unlock routes to encode and evolve new classes of polymers.H owever,b ecause the ribosome is essential for proteome synthesis and many mutations in the ribosome are dominant-negative or lethal in the cell, [2] it is challenging to alter and evolve the natural ribosome for unnatural polymer synthesis in cells.
To address the challenge of creating an evolvable ribosome,w eh ave previously created orthogonal (O)-ribosome-O-mRNApairs ( Figure 1) in E. coli. [3] TheO-ribosome contains am utated anti-Shine-Dalgarno (ASD) sequence within its O-16S rRNA, enabling O-ribosomes to selectively and efficiently translate O-mRNAs bearing the orthogonal Shine-Dalgarno (O-SD) sequences.Likewise,O-mRNAs are not translated by endogenous ribosomes.Because the orthogonal ribosome,unlike the natural ribosome,isnot responsible for synthesizing the proteome,i ts O-16S rRNAm ay be evolved to perform new functions.W eh ave previously evolved ribo-X in which the decoding center,w ithin the O-16S rRNAo ft he orthogonal ribosome,n ol onger recognizes release factor 1, thereby enabling efficient incorporation of unnatural amino acids in response to the amber stop codon. [4] We have also evolved ribo-Q,which uses extended anticodon tRNAs to efficiently incorporate unnatural amino acids in response to diverse quadruplet codons,e nabling the sitespecific incorporation of multiple distinct unnatural amino acids into recombinant proteins. [5] Many key ribosomal functions,including interactions with tRNAs and elongation factors,peptide bond formation in the peptidyl-transferase center (PTC), and the folding and release of the nascent chain through the exit tunnel, [1,6] are mediated by 23S rRNAw ithin the large subunit. These functional centers cannot be evolved in the current orthogonal ribosome that uses the endogenous pool of large subunits,c ontaining 23S rRNA, in combination with the orthogonal small subunit, containing O-16S rRNA, to translate the O-mRNA( Figure 1). Creating an O-23S rRNAt hat assembles into an orthogonal large subunit and is specifically coupled to the orthogonal small subunit, containing O-16S rRNA, will enable the creation of orthogonal ribosomes in which both subunits are selectively recruited to an orthogonal message ( Figure 1). This will facilitate alteration and evolution of functional centers in the O-23S rRNAnot possible on the endogenous 23S rRNA.
Thel arge and small ribosomal subunits interact through non-covalent RNA-RNAi nteractions between 16S rRNA and 23S rRNAt hat bury approximately 6000 2 ,a nd these interactions are dynamically regulated through the trans- Creating an orthogonal2 3S rRNA that specifically functions with an orthogonal1 6S rRNA (that does not function with endogenous 23S rRNA), will enable an altered 23S rRNA to be insulatedf rom cellular translationa nd selectivelyu sed in orthogonalt ranslation (right). lation cycle. [7] Efforts to control non-covalent subunit interactions through rRNAmutagenesis have proved unsuccessful thus far. Here we investigate the creation of an orthogonal ribosome in which the O-16S rRNAiscovalently attached to a2 3S rRNAt ocreate afused rRNA( Figure 2A). Thef used rRNAa ssembles into an ew orthogonal ribosome that translates an orthogonal message and permits mutagenesis of the 23S rRNA.
We envisioned joining the two subunits by reorganizing the rrnB operon such that a23S rRNAwould be nested within the 16S rRNAa salarge insertion ( Figure 2B). We were encouraged by previous observations that in various organisms 16S rRNAs can exist in multiple fragments or with long insertions. [8] Moreover,t he 23S rRNAi st olerant to circular permutation, [9] indicating that it might be possible to circularly permute the 23S rRNAtoopen up new 5' and 3' termini at positions proximate to surface exposed features of the 16S rRNA, and then insert this permuted 23S at that site on the 16S,c onnected on both ends by an RNAl inker ( Figure 2C).
We used high-resolution structures of E. coli ribosomes [7a, 10] and phylogenetic variation [11] in rRNAs equence to identify regions of 23S rRNAand 16S rRNAthat come close in space,a nd may be tolerant to insertion (Supporting Information, Figure S1 A). This analysis identified Helix 101 (H101) on the 23S and helix 44 (h44) on the 16S as an excellent pair of sites to test our strategy (Figure S1 B). These helices come into close contact (3 nm) in ribosome structures, [7a, 10] and are tolerant to insertions as judged by their natural phylogenetic variation [11b] and previous genetic engineering. [8b] Moreover,t hese helices are distal from the corridor through which tRNAs transit and elongation factors dock (Figure S1 B). Taking ar ational structure-based approach, we opted to circularly permute 23S at H101 and insert it within 16S,a tt he terminal loop of h44 ( Figure 2C). We linked the 16S and 23S sequences via the J5/J5a region from the Tetrahymena group Is elf-splicing intron (Figure 2C), an RNAhinge that can toggle between an extended and "U-turning" form. [12] This "stapled" ribosome rDNAwas synthesized by overlap extension PCR ( Figure S2, Table S1), cloned into ap RSF plasmid following an inducible P tac promoter, and given an orthogonal ASD (O-ASD) via sitedirected mutagenesis. [3] We refer to the resulting construct as pRSF-O-ribo(h44H101).
Because the unusual topology of the O-ribo(h44H101) rRNAc ould complicate ribosome folding and assembly pathways, [13] it was critical to ascertain the extent to which pRSF-O-ribo(h44H101) produces af ull-length rRNAt hat persists in vivo.T oa ddress this question we probed RNA extracted from E. coli expressing O-ribo(h44H101) by northern blot using ab iotinylated probe specific to the O-ASD sequence of the orthogonal ribosome ( Figure 3A). We detected as ingle band at 4500 nt, demonstrating that the major species bearing an O-ASD,i nc ells transformed with pRSF-O-ribo(h44H101), is the full length O-ribo(h44H101) rRNA. These data suggest that translation of O-mRNAs in  cells bearing pRSF-O-ribo(h44H101) results from the activity of the stapled ribosome.I nc ontrol experiments,R NA extracted from cells expressing the orthogonal ribosome from pRSF-O-Ribo (a plasmid with the same copy number, encoding orthogonal ribosomes under the same promoter, but with wild-type operon topology) was probed in an orthern blot with the O-ASD-specific probe.I nt his experiment we detected ab and at 1500 nt, as expected for the 16S rRNA ( Figure 3A), and the intensity of this band was approximately four times that of the band detected for O-ribo(h44H101) rRNA( Figure 3B). These data suggest that either the Oribo(h44H101) rRNAi sn ot transcribed as efficiently as the rrnB operon with native topology and/or af raction of the transcript does not assemble correctly and is ultimately degraded.
To test the activity of O-ribo(h44H101) in protein translation we co-transformed pRSF O-ribo(h44H101) and an Ocat reporter in which ac hloramphenicol acetyltransferase gene (cat) is downstream of an O-SD site for ribosome binding. [3,4] Tables S2, S3). In contrast, when cells are provided with pRSF-O-ribo(h44H101) and O-cat they grow robustly on Cm concentrations up to 70 mgmL À1 ( Figure 3C,T able S4), indicating that pRSF-O-ribo(h44H101)d irects the synthesis of ribosomes that specifically translate the orthogonal message.T he activity of O-ribo(h44H101) in the assay is lower than that of O-ribosomes with independent subunits produced from astandard operon, which confer Cm resistance up to 200 mgmL À1 ,b ut not 300 mgmL À1 ,i no ur assay ( Figure S4, Table S5). We further demonstrated the activity of O-ribo-(h44H101) in an independent assay by measuring its ability to translate O-luciferase (a luciferase gene expressed from an orthogonal ribosome binding site), [4] as measured by aluciferase activity assay ( Figure S5). This led to results that are quantitatively consistent with our observations in the chloramphenicol resistance assay.
These data indicate that the mutations do not have asubstantial dominant-negative effect on cellular translation in the stapled ribosome,c onsistent with the large subunit of O-ribo(h44H101) being functionally insulated from the endogenous small subunit ( Figure 4D).

Angewandte Chemie
In conclusion, we have described the rational, structurebased design of as tapled orthogonal ribosome.O ur design inserts acircularly permuted 23S rDNAinto the 16S rDNAat sites determined by structural and phylogenetic analysis,and uses an RNAh inge to staple the two subunits and facilitate subunit association and disassembly.Our results indicate that the stapled orthogonal ribosome allows the effects of mutations in 23S rRNAt ob es pecifically coupled to translation of an orthogonal message and insulated from endogenous translation. Future work will focus on optimizing the activity of our rationally designed stapled ribosome,and fully characterizing the extent to which orthogonality in subunit association (Figure 1) may be achieved through the stapling of ribosome subunits.W ea nticipate that the development of stapled orthogonal ribosomes may further extend orthogonal translation, and enable further progress on the genetically encoded synthesis of unnatural polymers in cells.