Base and Prime Editing: Changing DNA Without a Planned Double-Strand Break
Conventional cytosine and adenine base editors link a deaminase to a DNA-targeting protein and act within an editing window. Prime editors copy a template carried on a guide RNA. Compare their intended changes, bystander edits, off-target risks and repair-dependent outcomes.
What a learner can do afterwards
- State the intended substitutions of conventional cytosine and adenine base editors and the changes prime editing can specify
- Explain the editing window and why bystander edits can appear beside the intended change
- Distinguish on-target unwanted outcomes from off-target changes and explain why editor comparisons need matched conditions and assays
1 · Read
Earlier editing lessons showed that repair affects outcomes. Conventional base editors link a deaminase to a DNA-targeting Cas protein, often a nickase that cuts one strand. A cytosine base editor, or CBE, aims for C to T. An adenine base editor, or ABE, aims for A to G. These changes describe the edited strand; the opposite-strand descriptions are G to A and T to C. Other base-editor systems can have different menus.
An editing window contains positions where the chosen editor can act. If a target C and another C both lie inside it, either may change. An unwanted neighbouring change is a bystander edit. A different guide may exclude that neighbour, but its required nearby DNA motif and window placement must permit this. Some targets have no suitable alternative guide. Conversion is possible, not guaranteed.
A Cas9 prime editor uses a nickase and reverse transcriptase, which copies RNA into DNA. The guide's targeting sequence pairs with the target DNA strand. The nickase cuts the other, non-target strand. Reverse transcriptase extends it using the extended guide's template. Cells can retain or reject this edited DNA. Prime editing can specify all 12 single-base substitutions and small insertions or deletions, with variable efficiency. Some versions also nick the other strand.
Classify location and outcome separately. A neighbouring base change at the intended locus is an on-target bystander. An indel there is on-target; whether it is wanted depends on the requested edit. A change at another genomic site is off-target. Some deaminase editors can also affect RNA. Prime editors and nucleases can have unwanted outcomes. Compare the same cells with suitable assays; no editor class has a universal precision or speed ranking.
Cellular processing can retain or reverse editing intermediates. In some prime-editing designs, an opposite-strand nick can favour the intended sequence but also increase unwanted outcomes. Endpoint sequencing alone cannot tell whether an undetected edit never formed or formed and was lost. Failed delivery, low editor activity and repair are possible causes. An edited fraction may also fall below the assay's detection limit.
Match the requested change to the editor and target site. Conventional CBE and ABE menus exclude C to A, a transversion; prime editing is a candidate for that change. A nuclease cut alone specifies no replacement letter and can yield varied repair products. A nuclease plus a donor can support precise editing, but that is a different design. Avoiding a planned double-strand break does not make indel risk zero.
Match the change to the editor and target site. Check bystanders, off-target changes and repair outcomes instead of assuming one editor is always most precise.
Readings and flashcards written for this learning dataset. Linked background videos: How CRISPR lets you edit DNA - Andrea M. Henle, from TED-Ed, and Introduction to genetic engineering, from Khan Academy.
More to read
Read the conversion menu on a stated strand
Write down the requested change and the strand it describes before choosing an editor. Conventional cytosine deaminase editors aim for C to T. Conventional adenine deaminase editors aim for A to G. Reading the opposite strand describes these changes as G to A and T to C. These menus do not cover every base-editing system.
Next inspect the window of positions accessible to the chosen editor. A second suitable base in that window may change along with the target. That is a bystander opportunity, not a guaranteed outcome. Moving the guide helps only if a suitable nearby motif and window placement are available. A target without a suitable guide may need another editing design.
Classify unwanted changes before comparing editors
Classify where a change occurred separately from whether it was requested. A neighbouring base conversion at the intended locus is an on-target bystander. An insertion or deletion there is also on-target. It is unwanted only if it differs from the requested outcome; prime editing can deliberately specify small insertions and deletions. A DNA change at another genomic site is off-target. Some deaminase editors can also change RNA, which requires a different measurement.
Avoiding an intended double-strand break can reduce certain repair-related problems. It does not remove every unwanted outcome. Prime editing can also yield unintended products, and different nuclease designs have different risks. Compare the intended edit and unwanted products under matched conditions, using assays suited to each outcome. Results from different targets or detection limits do not establish a universal precision ranking.
What an editing endpoint can tell you
Prime editing supplies an RNA template on an extended guide. In Cas9-based prime editing, the nickase cuts the DNA strand not paired with the guide's targeting sequence. Reverse transcriptase extends that strand using the RNA template. The cell must process the edited segment before the intended sequence is retained. This mechanism can specify all single-base substitutions and small insertions or deletions, but successful editing still depends on the design and cells.
Some versions add a nick in the opposite DNA strand to favour retention of the edited sequence. That can also increase unwanted outcomes. An endpoint showing only the original sequence cannot reveal the full history. Delivery may have failed, editing activity may have been low, or an intermediate may have been reversed. A small edited fraction could also be below detection. Further measurements are needed to distinguish these explanations.
Explore more
2 · Watch
Take it off screen
Where it sits
8 questions wait behind this lesson, each with its answer explained. Every answer feeds the sky: stars light as they are learned, and dim when it is time to come back.