I am working on an (expensive) synthetic construct, which happens to have many "repetitive" sequences within it that are integral to its function. Primarily, the two sequences that are worrying me are:

  • A long polyA (about 15 nucleotides).
  • A consecutive sequence of two nucleotides: GGGUUUGGUUGU, etc.

I've heard from many some of the transcription horrors they've been through because of similar sequences with similar nucleotides. However, some advise that transcribing it shouldn't be a problem.

How predictably can RNA polymerase transcribe these sorts of sequences in E.coli?


2 Answers 2


I don't think repetitive sequences are intrinsically problematic to an E coli. Bacteria also deal with things like polyadenylation. Basically, you have nothing to lose from trying -- E coli is one of the more robust expression platforms, and it's difficult to predict whether the sequence will fold into some tertiary structure (which can easily impede translation.)

So I think the predictability should be set at the baseline for E coli RNA polymerase error rate, which is about $10^{-5}$.


Yes, transcription of a polyA sequence will be unreliable: for a repetition of 15 times the same nucleotide, the rate of transcription errors will be huge.

For example in E. coli, this paper http://www.ncbi.nlm.nih.gov/pmc/articles/PMC331007/ gives some order of magnitude of transcription errors caused by repeats of nucleotides. According to their data, for a run of 11 A (so shorter than yours), their are more than one third of transcripts with a frame-shift (deletion or insertion of one nucleotide).

In addition you may have bigger problems than transcription errors: your construct will also be extremely prone to frame-shift mutations. For example according to the data on figure 3 of http://www.ncbi.nlm.nih.gov/pubmed/22991466 by extrapolating to a run of 15 A, you can expect an indel rate higher than 10^-2 per base pair per generations. Because you have 15 target nucleotides it means that more than 10% of the cells will have a frameshift mutation leading to your gene not being functional.

I do not have data for 'A consecutive sequence of two nucleotides'.

  • 1
    $\begingroup$ Welcome to Biology SE! Your answer is welcomed, but we prefer more detailed answers where things are explained, statements are supported with references. Simply posting links to articles may not necessarily the best practice (while it is a good way to support your claims). Please elaborate your answer by processing the papers you linked, extract useful and relevant information from them, use your on words, quote if necessary. Feel free to visit our help center on guidelines for posting good answers. In this form it is rather a comment than a full answer. $\endgroup$ Apr 14, 2015 at 12:23
  • $\begingroup$ I am not sure why 'This does not provide an answer'? I edited the post to add more precisions, but my answer remains the same: a polyA sequence of 15 nucleotides is a problem if it is 'integral to [the] function' as the author says. $\endgroup$
    – Ant
    Apr 14, 2015 at 14:21
  • $\begingroup$ I've seen mRNA delivery papers that make their mRNA using a plasmid with up to 150 As on the 3' end of the gene so they don't have to add the tail through a separate enzymatic reaction. However, being on the 3' end would probably prevent frameshift problems. I don't know how they get the As into the plasmid, I tried to order a 100 A DNA oligo from IDT so I could try to add it to a plasmid, but they wouldn't synthesize anything longer than a 30 As in a row. $\endgroup$
    – user137
    Apr 14, 2015 at 15:08
  • $\begingroup$ @AliceD 's comment was before you edited your answer which by the was looks much better! $\endgroup$ Apr 14, 2015 at 20:53
  • $\begingroup$ Great answer. +1 thanks for editing. Writing up the answer and answering the question is something else here at Bio.SE :) $\endgroup$
    – AliceD
    Apr 14, 2015 at 21:04

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .