Insights into substrate promiscuity of human seryl-tRNA synthetase

Seryl-tRNA synthetase (SerRS) attaches L-serine to the cognate serine tRNA (tRNA(Ser)) and the noncognate selenocysteine tRNA (tRNA(Sec)). The latter activity initiates the anabolic cycle of selenocysteine (Sec), proper decoding of an in-frame Sec UGA codon, and synthesis of selenoproteins across all domains of life. While the accuracy of SerRS is important for overall proteome integrity, it is its substrate promiscuity that is vital for the integrity of the selenoproteome. This raises a question as to what elements in the two tRNA species, harboring different anticodon sequences and adopting distinct folds, facilitate aminoacylation by a common aminoacyl-tRNA synthetase. We sought to answer this question by analyzing the ability of human cytosolic SerRS to bind and act on tRNA(Ser), tRNA(Sec), and 10 mutant and chimeric constructs in which elements of tRNA(Ser) were transposed onto tRNA(Sec). We show that human SerRS only subtly prefers tRNA(Ser) to tRNA(Sec), and that discrimination occurs at the level of the serylation reaction. Surprisingly, the tRNA mutants predicted to adopt either the 7/5 or 8/5 fold are poor SerRS substrates. In contrast, shortening of the acceptor arm of tRNA(Sec) by a single base pair yields an improved SerRS substrate that adopts an 8/4 fold. We suggest that an optimal tertiary arrangement of structural elements within tRNA(Sec) and tRNA(Ser) dictate their utility for serylation. We also speculate that the extended acceptor-T Psi C arm of tRNA(Sec) evolved as a compromise for productive binding to SerRS while remaining the major recognition element for other enzymes involved in Sec and selenoprotein synthesis.