Q1 What exactly is in-situ insertion?

Refers to using CRISPR/Cas9 technology in the genome The target gene’s native genomic location Specifically, an exogenous sequence—typically a tag (such as EGFP, mCherry, FLAG, HiBiT, or TurboID)—is precisely inserted so that the protein translated from this gene carries a built-in “traceable identity marker.”

It has three non-negotiable features: the insertion site is native (at the target gene locus), the promoter is endogenous (driven by the native promoter), and the copy number is defined (typically a single copy). If any one of these conditions is not met, it cannot be considered a true in situ insertion.

Q2 Why does it have to be “in situ”? Can’t it be inserted elsewhere?

Because gene expression depends not only on “how much” is produced, but also on “when, in which cells, and how much.” In other words, the genome is not a flat linear strand but a three-dimensional city: enhancers, silencers, chromatin accessibility, and topological domains all influence gene expression. The same DNA sequence can have entirely different outcomes when placed at different locations on the chromosome—this is known as the “position effect.”

Random integration is like tossing a lamp into some random corner of the city; in situ insertion is like putting the lamp back into its original socket. The former might be bright enough, but it won’t illuminate the place you want to see.

Q3 Plug N or C? How to choose?

To draw an analogy: when sewing a label onto a garment, should it be attached at the neckline (N end) or the hem (C end)? The answer is: wherever the garment must remain immobile.

Generally prioritized for consumers . Inserting the C-terminal tag before the stop codon ensures that the tag is translated along with the full-length transcript, minimizing risk.

But when there’s a “no-go zone” on the consumer side, you have to adjust the business side. If the C‑terminal region harbors critical functional motifs (such as PDZ‑binding motifs, transmembrane domains, or sites for farnesylation or palmitoylation), insertion could directly disrupt function. Conversely, if the N‑terminal region contains a signal peptide, N‑terminal insertion might interfere with protein secretion and subcellular localization, so experimental validation is also required.

Practical advice: First check UniProt/Pfam for domain annotations, then look for published literature on the same target; if neither is available, perform a transient transfection pilot experiment to test expression and localization at both the N- and C-termini. This is a scientific issue, not a technical one, and it needs to be clearly defined at the project initiation stage.

Q4 What’s the difference between in situ insertion, overexpression, and random integration?

One table makes it clear. Note: There is no single “better” option; the choice depends on which one best suits your scientific question.

Q5 How big of an item can fit inside?

In theory, there is no hard upper limit, but Efficiency decreases very noticeably with increasing length. : CRISPR/Cas9-HDR exhibits relatively high integration efficiency for fragments of 1–6 kb; however, once the fragment size exceeds approximately 7 kb, both efficiency and stability decline significantly.

• Short tags (consisting of a few to about a dozen amino acids): HiBiT (11 aa), FLAG (8 aa), HA (9 aa); these exhibit the highest efficiency and make it easiest to obtain homozygous mutants.
• Medium tags (approximately 0.7–2 kb): EGFP/mCherry, Luciferase, degradation tags (dTAG, AID), the mainstream range.
• Large tags (greater than 1 kb): TurboID, BioID, and large fragment functional modules require optimized strategies, such as AAV donors, P2A linkers, and split designs.

The one-sentence principle: Use just enough tags—smaller is better. Each additional kDa simultaneously reduces insertion efficiency and increases the risk of functional interference.

Q6 What is the complete process? Why does it often take two to three months?

The standard link looks like this:

1) Target and sgRNA screening

2) Design and synthesis of donor template (homology arms + tag)

3) Co-delivery of Cas9 RNP and donor (mostly by electroporation)

4) Pool-level assessment of editing efficiency

5) Monoclonal plating / sorting

6) Clonal expansion

7) Genotype identification (Junction PCR + Sanger sequencing)

8) Functional validation (expression level, subcellular localization)

A standard delivery includes one monoclonal hybridoma, with a typical turnaround time of 10–12 weeks. The biggest variable here is not the technical process but the cell doubling time—cells that grow slowly can add several weeks just to expand the clone.

Q7 Why is knock-in so much more difficult than knockout?

This is the key to understanding the entire technology. After Cas9 cleaves the DNA double strand, the cell has two repair pathways available. Competing for the same breaking point ：

• NHEJ (non-homologous end joining): fast, crude, active throughout the cell cycle, and can randomly add or delete a few bases, causing a frameshift—this is precisely the principle behind gene knockout. It is highly efficient.
• HDR (homology-directed repair): slow and precise, requiring a “blueprint” (donor template) to insert an exogenous sequence, and primarily active only during S/G2 phase—this is precisely the principle behind gene knock-in. Efficiency is low.

So, knock-in is essentially competing with NHEJ for the job. Common approaches to increase HDR efficiency include:

• Small-molecule regulation: DNA-PK inhibitors (AZD7648, NU7026 / NU7441) suppress NHEJ; the RAD51 activator RS-1 enhances HDR. Reported fold increases across different systems range from 2-fold to several dozen-fold.
• Cell cycle synchronization: Release cells after arresting them at the M/G2 boundary with nocodazole (typically 200 ng/mL for 16–18 h) to increase the proportion of cells entering the HDR window.
• Low-temperature treatment Incubate at 32°C for 24–48 h; the operational cost is extremely low, making it worth a try.
• Protein engineering: CtIP-Cas9 fusion and Geminin-Cas9 enable cell cycle–specific expression.
• Reducing edit warring: Apoptosis is reduced through sgRNA optimization or inhibition of p53 function (e.g., GSE56).

Q8 How do you choose the donor template? How long should the homology arms be?

The choice of donor template is the step most easily overlooked—and most prone to error—in knock-in experiments.

• Longer is not always better: when the combined length of the homology arms exceeds approximately 4 kb, HDR efficiency decreases.
• The distance of the nick is critical: the distance between the Cas9 nick and the insertion site should ideally not exceed 10 bp; beyond 10 bp, efficiency declines sharply.
• Preventing secondary cleavage: After repair is complete, introduce a silent mutation in the PAM or protospacer region; otherwise, Cas9 will cleave the newly repaired site again.
• Random integration rate: The random integration rate of ssDNA is 10–100 times lower than that of dsDNA; however, ssDNA is inefficient for large fragments, so a trade-off must be considered.
• Purity equals toxicity: dsDNA must be purified by column chromatography and endotoxin-free—endotoxin is the primary cause of cytotoxicity.

Q9 Why is it necessary to screen for monoclonal antibodies? Can’t we use a pool instead?

This is because the editing process is “cell-autonomous.” Within the same pool, there are unedited, heterozygous, homozygous, NHEJ‑induced frameshift‑mutant, and randomly integrated cells, resulting in completely heterogeneous allelic compositions—making it impossible to reproduce the experimental data in follow-up studies.

Exception : For primary cells, or cells with limited passage potential and extremely low clonogenic efficiency, pooling at the polyclonal level can be performed; however, this approach is suitable only for qualitative trend analysis and not for quantitative measurements. This point must be clearly communicated at the project initiation stage; otherwise, interpreting the data later may prove challenging.

The double low probability: the knock-in efficiency is already low, and it must be further compounded by the monoclonal formation rate (which in many cell lines falls below 10% at low density). Multiplying these two small probabilities means that the number of correct clones obtained may be only in the single digits, or even zero. This is also the fundamental reason why the cost of in situ insertion remains so high.

Q10 Sequencing indicates that the insertion was successful, so why is there not a single band on the WB?

This is the most classic hidden pitfall of in-place insertion: Successful genomic insertion does not necessarily mean the protein is detected. 。

The reason is straightforward: the expression level of an in‑situ insert is entirely driven by the endogenous gene. If the target gene is already lowly expressed in that cell, the tag’s signal will be “muted” along with it. It’s like speaking in a quiet room—the volume is too low to carry.

Three Steps to Avoid Pitfalls ：

① Before project initiation, first determine the endogenous expression level of the target gene in the target cell line (by Western blot or from public databases).

② If expression levels are too low, switch to a high-expression cell line (e.g., HEK293T, HeLa) to improve success rates at the source.

③ Use high-sensitivity tags; the detection sensitivity of HiBiT/NanoLuc is much higher than that of WB, and small tags cause less disruption to protein function.

Q11 What is the “cheat” called HiBiT? Why is everyone using it?

HiBiT is a small peptide tag consisting of only 11 amino acids. It does not fluoresce on its own, but upon binding with LgBiT (17.6 kDa) with extremely high affinity (Kd ≈ 700 pM), it will Reengineered bright NanoLuc luciferase — More than 10 times brighter than firefly luciferase.

• extremely small :11 aa, causing the least disruption to protein structure and function, with the highest CRISPR insertion efficiency.
• Extremely sensitive: The linear range exceeds seven orders of magnitude, and the detection limit can reach 10⁻¹⁹ mol.
• One-step method: Add reagents, mix thoroughly, and read the results—all in just a few minutes, with no antibodies or elution steps required.
• Quantifiable: Suitable for degradation kinetics, protein stability, and high-throughput screening.

Q12 Will the label damage the protein?

Yes, especially for large tags (such as GFP and TurboID). Typical manifestations include decreased expression levels, abnormal subcellular localization, loss of function, and aberrant protein aggregation.

Four mitigation pathways:

① Insert a flexible linker (e.g., (GGGGS)×3) between the target gene and the tag to reduce steric hindrance;

② Replace with a smaller tag (use HiBiT instead of GFP);

③ Swap the insertion ends (swap N and C);

④ Adjust the linker length—too short causes steric hindrance, while too long makes it prone to cleavage.

Q13 How can we rule out false positives caused by “the plug is inserted, but in the wrong location”?

The most dangerous outcome is not “no insertion,” but rather “insertion, yet in the wrong location.” Plasmid donors—especially those carrying resistance genes—can integrate randomly elsewhere in the genome. Such clones may pass resistance selection and even initial PCR screening if primers are poorly designed, but their genotypes will be incorrect.

Four lines of defense for authentication:

① Double-Ended Junction PCR One primer anneals to a genomic sequence outside the homologous arm, and the other anneals within the inserted fragment; both the 5′ and 3′ ends must be verified. Only targeted integration will yield an amplicon of the expected size.

② Sanger sequencing Confirm that the junctions, linkers, and border sequences are accurate.

③ Southern blot or ddPCR Determine the copy number to rule out multiple copies or tandem insertions.

④ Give priority to adopting ssODN or AAV donor , the random integration rate was significantly lower than that of plasmids.

Q14 Are heterozygotes sufficient? Why are homozygotes so difficult to obtain?

• In most scenarios, heterozygosity is sufficient. For subcellular localization and interaction studies, a heterozygous genotype is perfectly adequate, and it also retains one wild-type allele, resulting in lower functional risk.
• Scenarios Requiring Homozygotes : Rigorous quantitative comparisons, dose-response studies, and degradation assays that require all proteins to be tagged.
• What’s so hard about it? : This amounts to a “double hit” at the same locus, compounded by variables such as insert size, effects on protein function, and cellular ploidy (polyploid cells may have three or more alleles). In some cells, due to polyploidy or homozygous lethality, obtaining a true homozygote is simply not feasible.

Q15 Why is phase separation research indispensable?

Because liquid–liquid phase separation (LLPS) is important for Extremely sensitive to protein concentration —When the concentration becomes high enough, proteins that normally shouldn’t undergo phase separation can be “forced” to aggregate into condensates. In other words, if you conduct experiments using cells with overexpressed proteins, you may observe… manufacture The phenomenon you want to study.

So there is an unwritten rule in the field of phase separation: Conclusions must be based on evidence at the endogenous level.

Q16 Why is in situ recommended for proximity labeling (TurboID / BioID)?

The logic behind proximity labeling is to “equip the bait protein with a paint‑spraying machine”—all proteins within the enzyme’s radius of action get biotinylated. Once the bait is overexpressed, three things happen:

① Expression levels far exceed physiological norms, labeling even non-physiological neighbors and causing background signals to surge;

② The bait is mislocalized and mapped to a completely incorrect compartment, resulting in systematic false positives;

③ The copy number is variable, samples are not comparable between batches, and the data cannot be replicated.

Q17 Why is it also needed for protein degradation / PROTAC / target validation?

Degradation assays measure changes in protein abundance. However, overexpression causes the protein pool to far exceed the endogenous level—so even if 80% appears to be degraded, the actual remaining amount remains above physiological levels, completely distorting the degradation curve.

After in situ insertion of HiBiT, it can be used directly with Quantitative endogenous protein abundance from luminescence values , conducting degradation kinetics (DC50, Dmax), protein stabilization (e.g., HIF1A), target degradation (e.g., BRD4), and more, while also being naturally compatible with high-throughput screening using 96- or 384-well plates.

The same logic applies to systems such as dTAG and AID (auxin‑inducible degradation)—these techniques inherently require attaching a degradation tag to the endogenous protein; otherwise, they lose their purpose.

Q18 When should you not use in-place insertion?

• Large-scale screening: It requires hundreds or thousands of genes to be manipulated simultaneously, using CRISPRa or a shotgun overexpression library; do not perform in situ experiments.
• The target gene is barely expressed in the cells used. : First switch the cell line; otherwise, even if the construct is successfully built, it will not be detectable (see Q10).
• It is necessary to stably express the heterologous gene in a “safe site.” : Use safe harbor sites for targeted knock-in (e.g., Rosa26 or AAVS1) to achieve more stable and controllable expression. (See also: Service Guide | Liman Bio Site-specific and In situ Knock-in Stable Cell Line Construction Service )