The tRNA Elbow in Structure, Recognition and Evolution

First Round of Evaluation

Round 1: Reviewer 1 Report and Author Response

This review, clearly written, is based on recent results from studies on the bacterial T-box regulatory system performed by the authors and other groups. I feel that the information will reach a broad audience, among which people interested in tRNA structure and evolution as well as the evolution of the genetic system at large.

Response: We thank the reviewer for her/his positive comments and constructive suggestions. Responses to each of the specific comment are listed below.

Specific comments to the authors:

Line 72: The authors wrote: "In addition to stabilizing the L-shape of tRNA, the elbow is characterized...". The way this idea is expressed suggests that it might be possible to consider the L-shape independently of the "flat hydrophobic surface" of the elbow. Could the "raison d'être" (as the authors like to say) of the L-shape be connected to something else than this flat hydrophobic surface?

Response: The reviewer has improved on the presentation of our message that the flat hydrophobic surface is just one of the many facets of the L shape. It is in our opinion unlikely to be a requisite feature of the L shape and could have evolved fortuitously. The central bend of the tRNA is thought of as one of many geometrical solutions that allow simultaneous juxtapositions of the anticodons and 3’ ends of adjacent tRNAs in the P- and A-sites, to enable coded peptidyl transfer. In this sense, the L shape achieves a much more fundamental function than any local features of the elbow. Other interlocking mechanisms that do not involve a flat hydrophobic surface should presumably work at least to some extent. The fixation of this particular bending and locking mechanism happened to have been selected as one preferred solution. Thus we think the flat surface is a product rather than the raison d'être of the L shape.

We have added this discussion in the tRNA elbow in evolution section as follows:

“Geometrically speaking, the existence of the elbow was likely driven by the need to bend the tRNA, in order to simultaneously juxtapose the anticodons and acceptor ends of immediately adjacent tRNAs in the P- and A- sites [34,45]. Conceivably, the bending of tRNA structure to enable codon-directed peptidyl transfer could be achieved by other types of “elbows” that don’t involve the presentation of a flat, hydrophobic surface.”

Line 100: fascinating subject (no issue here)

Response: We appreciate the reviewer’s interest in the discussion here.

Line 171: "...the tRNA elbow, which can serve as tRNA identity elements in those cases". I have checked the pdb 1GAX structure (Val tRNA). It is not clear to me that the elbow is used as an identity element, i.e. an element that allows the enzyme to discrimitate this particular tRNA from other tRNA species. The enzyme is clearly taking advantage of this contact to achieve a higher stabilization. However, does it help the enzyme discriminate it from other tRNAs? The authors could clarify this point (the same with the other structure).

Response: We agree with the reviewer’s point and have revised the text to state the following:

“A smaller number of aaRSes (such as LeuRS, ValRS) also interact with the tRNA elbow, which can potentially serve as additional tRNA identity elements in those cases. Even when the elbow interactions don’t contribute to tRNA specificity, they would increase overall functional affinity, or avidity of the binding.”

Further, we have added a new example (new Figure 4c) of GatCAB where an elbow contact directly contributes to tRNA selectivity. We have added a new paragraph as follows:

Response: “In two fascinating examples, the archaeal and bacterial glutamine amidotransferases (the heterodimeric GatDE and heterotrimeric GatCAB, respectively), distinguish cognate substrate tRNAGln from noncognate tRNAGlu in part by employing shape complementarity between their tail domains and the smaller D-loops at the elbow of tRNAGln (Figure 4c)[18,19]. Even when the elbow interactions don’t contribute to tRNA specificity, they could increase overall functional affinity, or avidity of protein-tRNA association. This is exemplified by the helical domain of GatE, which binds the minor groove of the TSL near the elbow to provide an additional anchoring interaction but does not contribute to tRNA selectivity [18].”

Line 204: "It has been proposed that primordial tRNA-like molecules consisted only of the ASL and AAL, which function in decoding of mRNA codons and esterification to an amino acid, respectively". The authors could provide some references.

Response: We have provided new references for this point, as follows:

42. Sun, F.J. and Caetano-Anolles, G. (2008) The origin and evolution of tRNA inferred from phylogenetic analysis of structure. J Mol Evol 66, 21-35.

43. de Farias, S.T., do Rego, T.G., and Jose, M.V. (2014) Evolution of transfer RNA and the origin of the translation system. Front Genet 5, 303.

44. Weiner, A.M. and Maizels, N. (1987) tRNA-like structures tag the 3' ends of genomic RNA molecules for replication: implications for the origin of protein synthesis. Proc Natl Acad Sci U S A 84, 7383-7387.

Line 221: "A number of mitochondrial tRNAs lack the tRNA elbow structure. Consistent with this, the mitochondrial ribosomal L1 region and RNase Ps also lack the sequences that would form the interdigitated T-loops [42-45]. Similarly, there are no known T-boxes that operate within mitochondria."

This observation was partially discussed in ref [43] (as stated). Unless I am wrong, I think it was originally discussed in ref [25], in which case this reference could be listed with the other mentioned papers [42-45].

Response: We have corrected this issue and included the original reference.

Line 226: some T-box riboswitches such as the ileS T-boxes from Actinomyces, are truncated at the top of Stem I so that they no longer contain the interdigitated T-loops canonically employed for tRNA elbow recognition [46].

"no longer" is ambiguous here (given the discussion that follows).

Response: “No longer” has been removed, to avoid the impression of an evolutionary connection. 

Round 1: Reviewer 2 Report and Author Response

The tRNA elbow structure is a unique structural motif highly conserved in almost all tRNAs, except for animal mitochondrial tRNAs. Intriguingly, this structural motif is specifically recognized by some functional RNAs like 23S rRNA, RNase P RNA, and T-box riboswitch. Recent studies mainly done by authors’ group revealed molecular basis of its recognition by RNAs as well as by proteins. In fact, this review paper is an excellent contribution for deeper understanding of characteristic features of the tRNA elbow in terms of structure, recognition, and evolution. I really enjoyed reading it. I have only minor comments given below.

Response: We thank the reviewer for her/his positive comments and detailed suggestions. Responses to each of the specific comment are listed below.

 In lines 13 and 32, it says “elongator tRNAs”, but the elbow is also conserved in initiator tRNAs. I suggest to remove “elongator”.

Response: The “elongator” qualifier has been removed in these two locations.

As an example for protein that recognizes the elbow structure, I recommend authors to add Yuki’s work on Thermotoga maritima GatCAB which recognizes the elbow structure of Glu-tRNAGln.

Response: We thank the reviewer for an excellent suggestion. We have now added this example with a new figure panel (Figure 4c), two new entries to Table 1, and a new discussion on the elbow acting as an identity element for transamidation enzymes. Yuki’s work, among the two references here, does not appear to provide a PDB entry with a traced tail domain. So a homologous structure of the bacterial GatCAB is used instead to generate a new figure panel. We have added the following text.

 “In two fascinating examples, the archaeal and bacterial glutamine amidotransferases (the heterodimeric GatDE and heterotrimeric GatCAB, respectively), distinguish cognate substrate tRNAGln from noncognate tRNAGlu in part by employing shape complementarity between their tail domains and the smaller D-loops at the elbow of tRNAGln (Figure 4c)[18,19]. Even when the elbow interactions don’t contribute to tRNA specificity, they could increase overall functional affinity, or avidity of protein-tRNA association. This is exemplified by the helical domain of GatE, which binds the minor groove of the TSL near the elbow to provide an additional anchoring interaction but does not contribute to tRNA selectivity [18].”

Line 19. Should be riboswitches


Line 58. Should be dihydrouridine


Line 81. Phe should be superscript.


Response: Corrected

Line 90. “There” can be “Thus”?

Response: Although both words should work well here, in this particular context, “there” was used to accentuate the location (A, P, E sites) where tRNAs are encased by both RNA and protein components of the ribosome.

Lines 98 and 100. Should be 23S


Response: Corrected

Line 99. As no indication for Figure 2D in text, it can be cited here?

Response: We thank the reviewer for the suggestion and implemented it here.

Lines 117. Explain TSL-AAL.

Response: To increase clarity, we now changed it to “TSL (T-stem-loop)- Acceptor Stem”, rather than TSL-AAL.

Lines 169, 181, 216, and 237. Better to use aaRSs?

Response: We changed all the aaRSes to aaRSs.

Line 176, As no indication for Figure 4C in text, it can be cited here?


Response: We appreciate the reviewer suggestion and have added it here.

Lines 199 and 201. lambda-form

Response: Corrected


Line 221. “mitochondrial” should be “animal mitochonarial”, because yeast mithochondrial tRNAs have canonical elbow structures.

Response: We thank the reviewer for pointing out this fact and have made the correction.


In Line 237, tRNA-modification enzymes, but in line 162, 182, and 195, it is described as tRNA-modifying enzymes. Unify the nomenclature.

Response: We have now used tRNA modification enzymes throughout the text.

In Figure 1B. Describe the species of tRNA. Remove “t” in residue numbering.

Response: We have now added species of the tRNA as yeast tRNAPhe. The “t” notation for tRNA has been removed to improve clarity through all the figures. 


In Figure 1C-E. It’s helpful for readers to put numbers on residues which are discussed in main text.

Response: We have now added the numbers.

In Figure 2, explain nomenclatures for “T-Loop1”, “T-Loop2”, tT-Loop”, “tD-Loop”

Response: The “t” preceding T- or D-loops was intended to denote tRNA, and have been all removed now.

Figure 3B appears earlier than Figure 3A.

Response: In line 99, which is before Figure 3b, we added a citation to Figure 3a together with Figure 2d, thus resolving the issue. It now reads:

 “…the deacylated tRNA is moved to the E site where 23S rRNA from a mobile element of the large ribosomal subunit termed the L1 stalk holds the tRNA by the elbow (Figure 2d & 3a)”


In legend for Figure 3, Gly should be superscript.


In legend for Figure 4, RNA modyfing enzymes should be replaced with tRNA binding proteins. Leu, Val, Thr, and Trp should be superscript.

Response: Corrected