I am engineering a set of genetic sequences and have come across a surprisingly basic point of confusion that seems to have fallen through the cracks regarding coding sequences.

The standard start codon is AUG, and since this is an RNA sequence, I would thus expect that the start of the coding sequence would then be TAC, since DNA is complementary to the mRNA transcript taken from it. Wikipedia and various introductory texts confirm this expectation, e.g., in wikipedia's article on the central dogma:

The ribosome reads the mRNA triplet codons, usually beginning with an AUG

Here's another example: in the Khan academy introduction to the genetic code, the codon illustration clearly shows an AUG in the mRNA sequence.

The coding sequences that I am composing into my design, however, are all listed as starting with ATG, rather than TAC, which would be the complement of AUG. I am certain the coding sequences are not wrong, since some are very well established parts like BBa_E0040.

Something seems to have gotten confused or elided in many standard explanations, then. Why do coding sequences start with ATG and not TAC?


This image from the Kahn Academy article 'Overview of transcription' might help:

Overview of central dogma

Essentially, the sense/coding strand of the DNA encodes the sequence that is transcribed. The RNA polymerase binds to the antisense/template strand, for which the code is indeed TAC, but when it then transcribes this strand it is again complemented, giving us the AUG that is recognised by initiator tRNAMet (also known as tRNAfmet), as shown in this diagram from the Kahn academy article on 'tRNA and Ribosomes'.

Met tRNA binding

(Note that this diagram is actually showing Met in the middle of a sequence, rather than the start. fMET binding looks slightly different.

  • $\begingroup$ Thank you, that is very clear. I'm surprised that this is not more clearly presented in more places, since everything I'd found either said nothing about which strand or else used terms like "template" and "non-template" that weren't clear about how they related to promoter location strands. $\endgroup$ – jakebeal Mar 3 at 14:07
  • $\begingroup$ Yes! I think this is something that is intuitive to people who've been through the standard Biology training so it's often not explicitly explained to folks approaching SynBio from different angles. It's good to have a clear question here so that people can find it in the future. $\endgroup$ – Noah Sprent Mar 3 at 14:15

When your start codon is 5'-AUG-3', it's clear that the template DNA strand should be 5'-CAT-3'

If you look at Noah's illustration, it's also clear that the RNA sequence is basically the same as the sequence of the non-template strand, and runs in the same direction.

People generally refer to the mRNA not in terms of its RNA nucleotides, but in terms of the cDNA which it could be converted to. But they'll reported it in the direction that makes sense for talking about RNA.

  • $\begingroup$ I'm afraid that your answer contains the sort of mirror-language ambiguities that caused my confusion in the first place. Which cDNA to convert it to, the template or the non-template? From @NoahSprent's answer, I now know that the template is the opposite from the strand with the promoter, but if I didn't have that answer to read, I would still be confused. $\endgroup$ – jakebeal Mar 3 at 20:17
  • $\begingroup$ cDNA is double stranded molecule, so it contains the template and non-template strands. $\endgroup$ – swbarnes2 Mar 3 at 20:37
  • $\begingroup$ Yes, of course: that's why there's the ambiguity that I was finding confusing. $\endgroup$ – jakebeal Mar 3 at 20:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.