Donor templates for a knock-in

The knock-in you can honestly specify starts from how much new DNA the donor must carry, and from whether the cells are likely to use it. A 3-base substitution and a 2-kilobase cassette are not two sizes of the same reagent. This page compares single-stranded oligonucleotides, linear double-stranded DNA and plasmids as research donors. Why those molecules often lose to end joining is in homology-directed repair versus end joining. The nuclease that makes the break is in how CRISPR-Cas9 editing works in research.

Oligonucleotides, plasmids and the enzymes used to check junctions are classes in the molecular biology catalogue. The sequence specification belongs on the quote request.

What every donor class still has to do

A donor for homology-directed repair presents the new sequence flanked by arms that match the target genome closely enough for strand invasion or annealing. The arms are homology, not decoration. A mismatch between the arm and the parental allele, including a SNP you did not know the cells carried, lowers the chance the donor is used. Build arms from a named assembly on Ensembl and correct them if you have parental sequence.

The cut should land close to the new bases when the donor is an oligo. A planning habit for that distance is within about 10 to 20 base pairs. Longer gaps are sometimes successful and are a weaker plan. The donor should also prevent the repaired allele from being cut again, unless the edit itself destroys the PAM or the seed. Otherwise the knock-in is a brief intermediate on the way to an indel.

Random integration is the competing fate of any double-stranded donor that enters the nucleus. The cell does not owe you a targeted allele because the map was labelled knock-in. Genotyping has to ask for junctions the donor molecule cannot supply by itself.

Single-stranded oligos for small cargo

A single-stranded oligodeoxynucleotide fits point substitutions, small tags of a few to a few dozen amino acids only when the whole molecule stays within what synthesis can deliver cleanly, and short insertions or deletions. A widely used planning range is about 30 to 80 nucleotides of homology on each side of the edit. Laboratories sometimes start nearer 40 to 60 nucleotides and lengthen an arm only with a reason. Arms much below about 20 nucleotides are a common cause of an oligo that never appears in the locus. Total length is often kept under roughly 200 nucleotides because longer single strands accumulate synthesis errors and truncations. That ceiling is a manufacturing consideration, not a biological law. Follow the paper for your cell type if it has earned a different length.

Which strand to synthesise, the PAM strand or its complement, and whether the two arms should be equal, are variables with a literature. Some designs use an asymmetric oligo. Treat the strand as a choice copied from the method you are following, then test it. Do not invent a house rule from one gene.

End protection, such as a few phosphorothioate linkages, is an optional chemical class used to slow exonuclease attack. It is not required for every oligo, and it does not replace homology or a nearby cut. If you use it, record it. An oligo with modified ends and an oligo without them are not the same reagent in a comparison.

Single-stranded donors integrate at random less readily than plasmids, which is one reason they are preferred for small edits. They can still be copied into the genotyping PCR as contamination. Handle them as concentrated sequence, not as a harmless primer.

Linear double-stranded DNA

When the insert outgrows a clean oligo, a linear double-stranded donor is the next class. Typical cargos are longer tags, selection cassettes and inserts of several hundred bases up to a few kilobases, subject to how the fragment is prepared. Homology arms for this class are often planned from a few hundred base pairs up to about one kilobase on each side. Shorter arms sometimes work with CRISPR, because the break is so local, and classical gene targeting used still longer arms. State the range you planned, and treat the method paper for that insert size as the document that overrides a generic window.

Linear donors can be PCR products or restriction fragments. Either way the ends and the arm sequences need to be known, not assumed from a gel band. A PCR-amplified donor that used primers inside a plasmid backbone can drag unwanted backbone into the locus. Sequence the donor molecule itself before you blame the cells.

Double-stranded DNA at high dose is toxic to some cells. The amount is a titration the method paper owns. This page will not copy a nanogram table. If viability collapses and the few survivors are unedited, you have a dose or a purity problem, not evidence that homology-directed repair is absent from the species.

Plasmids, and the integration you did not want

A circular plasmid can carry large cassettes and arms that are inconvenient to amplify as a linear fragment. In CRISPR-era targeting plasmids, a common planning window is about 0.5 to 2 kilobases of homology on each side. Designs inherited from older gene-targeting vectors sometimes use longer arms. Again these are planning ranges. A paper that succeeded with 800 base pairs in one cell type has not proved that length for your locus.

Circularity changes the outcome mix. Some methods linearise the donor. Some do not. Linearisation can change how readily the molecule integrates at random versus how it recombines at the break. Follow one method and say which you did. A plasmid that remains circular in the well for days is also a persistent source of PCR contamination and of continued expression if it carried a marker.

The selection cassette is the feature that misleads clone picking. Resistance or fluorescence shows that the cassette was retained somewhere. Random integrants express markers. Targeted integrants express markers. Untargeted cells that were shielded by a neighbour can look positive for a while. Junction genotyping, with one primer outside an arm, is mandatory before a plasmid-derived clone is called a knock-in. Selecting edited clones without fooling yourself walks through that trap. Plasmid construction itself, before any cell sees the DNA, follows the same discipline as plasmid cloning from insert to colony.

Adeno-associated viral donors are a further class: a single-stranded genome with a packaging limit, useful in some primary cells, and a biosafety review of their own. They are not an oligo and not a naked plasmid. Mention them in the plan only if the institutional approval covers them.

Junctions that tell a convenient story

Design genotyping primers in NCBI Primer-BLAST so you know what else they might bind. For a knock-in, one primer sits in the insert or across the new junction and the mate sits in genomic DNA outside the arm. Run the donor molecule alone as a control. It must not produce that amplicon. Run the unedited parent. It must not produce it either. A positive in both edited cells and the donor-only tube means the assay cannot see targeting.

Sequence across both junctions for a large insert. One correct junction and a second broken one is a partial insertion, which can still make a marker and still fail the protein you meant to tag. Small oligo edits need the whole interval sequenced, because a blocking mutation and the intended mutation can segregate if more than one repair event occurred.

Silent changes you added to block recutting are part of the allele. Omitting them from the figure legend is how a later group "corrects" a clone that was already not wild type at those bases.

Containment of a concentrated sequence

Donor DNA is a contaminant with a map. Store it away from genotyping benches. The physical separation used for PCR setup applies with extra force, because a successful knock-in PCR and a contaminated one look the same. Institutional biosafety covers the cells and, if you use a viral donor, the vector. This comparison is not approval to build a clinical vector, and it is not a procedure for editing embryos or the germline. A cassette that includes a resistance gene or a recombinase should be named in the risk review. The DNA is not inert because it is synthetic.

Paperwork, humidity and peeled labels

Order the full sequence, not a gene name plus the words homology arms. The enquiry and the synthesis sheet should state arm lengths, the blocking changes, the strand for an oligo, and the assembly. On arrival, check that the length you received can contain the arms you drew. A truncated synthesis will not confess itself in the cell. In humid weather, tube labels fail. Write the donor identifier into the notebook and on the cap. A freezer that warmed in a power cut is a smaller risk for DNA oligos than for protein or RNA, and a relabelled box is still a risk. Do not aliquot two donors into identical unmarked tubes because the map is "in the folder".

What a knock-in enquiry should contain

State the insert sequence, the arm lengths you are planning, the donor class, the distance from the guide cut to the edit, and the cell type. Say whether a selection cassette is included and how junctions will be sequenced. The CRISPR validation sequencing enquiry reference can be used to discuss that genotyping. It is an enquiry reference, not a statement that a knock-in service is operated. Plasmid work sits in the context of the molecular biology pathway. Ask whether a quotation is possible for the oligonucleotide or the cloning reagents. Do not ask for a knock-in as if it were a catalogue item with a guaranteed allele.