SeqBench

In-Frame or Not: Checking a Fusion Before You Express It

7 min read · Updated November 11, 2026

A fusion construct is two coding sequences joined in one reading frame, and the junction is the only place it can go wrong. Everything downstream of a one-base error is not a tagged protein with a small defect; it is a different peptide that usually stops early.

The arithmetic to check it is a modulus. What is worth understanding is why the check is done against the sequence rather than the map, and what it still leaves open.

The offset rule

If the tag begins at the CDS's first base, its offset is 0 and it is in frame. If it begins at the second base, the offset is 1, which is not a multiple of three, and the tag's codons are read against a frame the ribosome is not using. The test is simply offset mod 3 = 0.

Note that in-frame is not the same as correct. A tag that starts at offset 3 shares the reading frame and is translated, but there are now three bases of the target's codons in front of it, and the fusion is not what the diagram showed. The modulus catches the frame; it does not catch a junction that is three bases off.

Worked: one base, two proteins

Take a short construct that begins ATG, then six histidines as CAT CAT CAT CAT CAT CAT, then a coding stretch, then TAA. Translating it gives MHHHHHHAEVLFEGS and a stop — a clean N-terminal His fusion.

Insert a single A immediately after the ATG. The same bases now read in a different frame: MTSSSSC, then a stop, then more out-of-frame peptide. The tag is still there in the DNA — six CAT codons are still six CAT codons in the sequence — and what is translated is seven residues of nothing useful.

This is why a junction is checked in the sequence and not on the map. An annotated map that draws a tag box next to a CDS is describing an intention; the reading frame is a fact about the bases, and one inserted base during a cloning step moves it.

The checks that belong with it

  • A stop codon between the fusion partners. A tag followed by a stop codon and then your gene is a tag that is expressed and discarded; the target is not fused at all.
  • A linker, and whether it is in frame with both sides. Gly-Ser linkers are common because they are flexible and rarely structured, and they still have to be a multiple of three.
  • The cleavage site's frame, if there is one. TEV is ENLYFQG, thrombin is LVPRGS; both are peptide sequences and both shift the frame if they start at the wrong base.
  • The stop codon at the end, and whether the tag is C-terminal and therefore after it. A C-terminal tag placed past the stop is not translated.

What a sequence check cannot settle

Expression, solubility and folding are not properties a reading frame determines. An in-frame fusion can be insoluble, toxic, or cleaved by host proteases at the linker.

The construct in the tube is also not the construct in the file. A frameshift is exactly the class of error introduced during cloning rather than during design, so the check that matters is the one run on the sequence you read back from the colony — not the one run on the plasmid map you started from.

And a fusion that translates correctly can still be the wrong molecule: a duplicated tag, a partial digest that left two linkers, or an internal start codon that produces the untagged protein alongside the tagged one. A gel shows the sizes; only sequencing assigns them.

FAQ

How do I check whether my fusion is in frame?

Take the position of the tag's first base within the coding sequence, counting the A of the start codon as position 1. If (position − 1) is divisible by 3, the tag shares the reading frame. Then translate the junction and confirm there is no stop codon between the partners.

What happens if a His tag is out of frame?

The tag's codons are read in the wrong frame, so the translated peptide is unrelated to the tag and usually hits a stop codon within a few residues. The tag is still present in the DNA and absent from the protein, which is why an anti-His blot comes up blank on a construct that looks correct on the map.

Does the linker between a tag and my protein need to be a multiple of three?

Yes, every part of the junction does. Gly-Ser linkers are common because they are flexible, and they still have to preserve the frame; a linker of the wrong length shifts everything after it.

My construct is in frame on the map but the protein is the wrong size. Why?

Because the map is a drawing and the clone is the sequence. Check the junction in the read-back sequence for an inserted or deleted base, and check for a stop codon or an internal start codon that would produce a truncated or untagged product.

Related references

Related tools

Related guides

Helpful?