Incorporation of unnatural amino-acids is a powerful and versatile new tool for modifying and selectively tagging proteins of interest. The technique has been shown to be applicable to C. elegans by Greiss and Chen (2011), utilizing a transgene expressing an amber-suppressing tRNA together with a cognate pyrrolysl-tRNA synthetase from Methanosarcina.

We wished to apply this technique to investigate srf-5, a gene affecting surface antigenicity and bacterial infection susceptibility, which we have found to encode a predicted small secreted protein. The reference srf-5 allele, ct115, causes an early Trp-to-opal nonsense mutation and behaves as a genetic null. We reasoned that it should be possible to rescue the mutant phenotype by introducing a transgene carrying a Trp-to-amber mutation at the same position, and crossing this transgene onto an amber-suppressing genetic background such as sup-5. Crossing the same transgene onto the unnatural-amino-acid-inserting background developed by Greiss and Chen (2011), and providing a suitable unnatural amino-acid substrate, should then result in synthesis of SRF-5* (that is, SRF-5 with an unnatural amino acid instead of Trp). If SRF-5* is as functional as wildtype SRF-5, then it should also rescue the srf-5 mutant.

Surprisingly, we found that the srf-5(amber) transgene was able to rescue the srf-5 mutant in the absence of any amber suppression, when present in a multicopy extrachromosomal array. The same effect was seen when the original opal nonsense mutant, ct115, was introduced as a multicopy extrachromosomal array. We hypothesized that transcription from the multiple copies leads to elevated srf-5 mRNA levels, and consequently to some read-through of either stop codon. We tested this interpretation by crossing srf-5(ct115) with smg-2(e2008), a mutant defective in NMD (Nonsense Mediated Decay), and found that smg-2 also partly suppressed the Srf-5 phenotype. The ct115 opal (UGA) codon therefore appears to be a leaky stop. It may be unusually leaky, or else SRF-5 may be required only in very small amounts. An alternative explanation, that the amber fragment encoded by srf-5(ct115) has some functional activity, seems unlikely because this fragment is small (<35 amino-acid residues).

Using amber transgenes for unnatural amino-acid incorporation should work for other genes, and may still work for srf-5 itself. However, the read-through effect that we report here may sometimes complicate matters.