A bulk cell heterozygous knock-in strategy for targeted protein degradation

Abstract

Targeted protein degradation using conditional degron tag (CDT) technology is a powerful method for rapidly degrading a protein of interest (POI) upon the addition of a degrader drug. A prerequisite for the temporally controlled degradation of an endogenous POI is the generation of homozygous knock-in cells with the degron tag integrated at either the N- or C-terminus of their gene loci. However, obtaining those homozygous knock-in cells often requires selecting many single-cell clones, as human cells typically exhibit low homology-directed repair (HDR) activities. Additionally, tagging a degron to an endogenous protein may inadvertently reduce protein expression, potentially affecting protein function even before the drug is administered. Here, we develop a method for generating degron-tagged knock-in cells that allows us to skip the laborious single-cell cloning. This method arose from our observation that most knock-in cells carry the degron tag only in one allele (heterozygous), while the other allele typically harbors a frameshift insertion/deletion. This observation allowed us to bypass the need for single-cell cloning. We validated our method by knocking in degron tags at the N-terminus of cytoplasmic dynein1 subunits or Adaptor Protein 2 (AP2) subunit. Our experiments confirmed the rapid degradation of these proteins and their functional inhibition in bulk cell populations. Additionally, to mitigate the reduced expression often associated with the degron tagging, we established a method to control expression levels by inserting a mini-promoter immediately upstream of the knock-in cassette. Our method simplifies the workflow for degron tag knock-ins and enhances the versatility of these valuable technologies.

Introduction

Conditional protein depletion via degron-mediated degradation is a powerful tool for investigating protein function in complex biological systems and can also serve as a therapeutic approach to modulate disease-associated proteins Pillow et al. (2020)^; Maneiro et al. (2020). Conditional degron technologies have been developed based on proteolysis-targeting chimeras (PROTACs) Nabet et al. (2018^{, 2020}); Buckley et al. (2015)^; Tovell et al. (2019) or molecular glue degraders Nishimura et al. (2009)^; Yesbolatova et al. (2020)^; Koduri et al. (2019)^; Sievers et al. (2018) that utilize E3 ubiquitin ligases, such as CRL4–Cereblon (CRBN) Nabet et al. (2018)^; Koduri et al. (2019)^; Sievers et al. (2018) and CRL2–Von Hippel-Lindau (VHL) Nabet et al. (2020)^; Buckley et al. (2015)^; Tovell et al. (2019). These include the degradation tag (dTAG) system Nabet et al. (2018), the Halo-PROTAC system Buckley et al. (2015), the Auxin-inducible degron (AID) Nishimura et al. (2009)/AID2 Yesbolatova et al. (2020) system, and the IKAROS family zinc finger 3 degron (IKZF3d) system Sievers et al. (2018). These degrons can confer rapid degradation kinetics upon drug addition and reversible recovery of protein levels after the drug removal Nabet et al. (2018)^; Nishimura et al. (2009)^; Yesbolatova et al. (2020)^; Bondeson et al. (2022). To investigate the function of endogenous proteins, the degron tag technologies are often combined with the CRISPR-Cas9 system to knock in the degron tag at either the 5’ or 3’ end of the gene of interest Yesbolatova et al. (2020)^; Damhofer et al. (2021)^; Ghetti et al. (2021)^; Abuhashem and Hadjantonakis (2022)^; Kurashina and Mizumoto (2023). The CRISPR-Cas9 system induces precise double-strand breaks (DSB) in the targeted DNA via an RNA-mediated process Jinek et al. (2012), and the DSBs are repaired mainly by the two pathways: nonhomologous end-joining (NHEJ) and homology-directed repair (HDR) Scully et al. (2019). NHEJ often creates insertions or deletions at the repair site, whereas HDR utilizes a donor template that has homology to the sequence surrounding the DSB region for accurate repair and is employed for knocking in the degron tags. Although the HDR/NHEJ ratio is largely affected by the gene locus and cell types, NHEJ was shown to be predominant for repairing the Cas9-induced DSBs in human cells Miyaoka et al. (2016). The low HDR activity makes it difficult to obtain the homozygous degron tag knock-in cells, and laborious single-cell cloning is often considered necessary to select the cells that harbor degron tags in both alleles Damhofer et al. (2021)^; Abuhashem and Hadjantonakis (2022). This limitation may restrict applicable cell lines to those with relatively high HDR activity. Several strategies that can improve the HDR/NHEJ ratio have been developed. These include synchronizing the cell cycle phase to S and G2 Lin et al. (2014)^; Wienert et al. (2020)^; Yang et al. (2016), when HDR normally occurs Takata et al. (1998), inhibiting NHEJ Cloarec-Ung et al. (2024)^; Maruyama et al. (2015)^; Riesenberg and Maricic (2018), and enhancing HDR Charpentier et al. (2018)^; Pinder et al. (2015)^; Song et al. (2016)^; Zhang et al. (2020). However, those approaches may not yield consistent results Gerlach et al. (2018) or may lead to genomic instability Cullot et al. (2024).

Another consideration not frequently discussed is the alternation of the expression level of the protein of interest through the insertion of a degron tag. To efficiently inhibit the function, the expression level of the degron-tagged protein should be high enough to maintain its function before adding the drug, and low enough to compromise its function upon the addition of the drug. The previous systematic study demonstrated using the ectopic expression system that the expression level of the protein of interest is largely affected by the specific degron tag (FKBP12^F36V, Halo-tag, AID/AID2, or IKZF3d) and the location (either N- or C-terminus) of the tag in the absence of the drug Bondeson et al. (2022). Similarly, the published protocol noted that the addition of FKBP12^F36V to the endogenous gene locus can lower the protein expression level in an unpredictable manner Damhofer et al. (2021). A versatile method to overcome this issue has not been established.

Results

The heterozygous dTAG knock-in system allows us to bypass single-cell cloning

We first determined the efficiency of obtaining homozygous degron tag knock-in in the immortalized retinal pigment epithelial cell (RPE1-hTERT), a genetically stable near-diploid cell Jiang et al. (1999), which is believed to have low HDR activity. The previous study required the utilization of two donor vectors that have different antibiotic resistance genes to obtain homozygous knock-in cells Miyamoto et al. (2015). Another study employed the NHEJ inhibitor to acquire a higher percentage of homozygous knock-in cells Ghetti et al. (2021). We transfected RPE1-hTERT cells expressing Blue Fluorescent Protein (BFP) tagged Cas9 Kanie et al. (2017) with a CRISPR RNA (crRNA) targeting near the start codon of a gene of interest (GOI) and a transactivating CRISPR RNA (tracrRNA) along with the donor cassette that consists of a left homology arm (LHA), a blasticidin resistance gene (BlastR), the self-cleaving peptides (P2A), V5-tag, FKBP12^F36V, and a right homology arm (RHA) (Figure 1A and Figure 1-figure supplement 1A–E). The dTAG system utilizes FKBP12^F36V with the heterobifunctional degraders that bind to both FKBP12^F36V and the E3 ubiquitin ligases Nabet et al. (2018^{, 2020}). Upon addition of the degrader drug, the FKBP12^F36V fusion protein can be quickly degraded Nabet et al. (2018^{, 2020}). Following the blasticidin selection and single-cell sorting, we extracted genomic DNA from each single-cell clone and amplified the region flanking the crRNA targeting region using polymerase chain reaction (PCR) to test the insertion of the knock-in cassette. As expected, all the single-cell clones that we analyzed had knock-in alleles. The analyses of five knock-in lines revealed that almost all, except four, had the knock-in cassette in only one allele, confirming the low success rate of generating homozygous knock-in cells (Figure 1B).

Figure 1. Generation and validation of the heterozygous knock-in cells harboring FKBP12^F36V tagged proteins.

(A) A diagram showing the design of the donor double-stranded DNA (dsDNA) for the FKBP12^F36V knock-in cells. LHA, Left Homology Arm; BlastR, blasticidin resistance gene; RHA, Right Homology Arm. (B) A table showing the number of single-cell clone RPE-BFP-Cas9 knock-in cells that have homozygous or heterozygous knock-in alleles and the percentage of homozygous knock-in cells. PCR amplified DNA surrounding the guide RNA targeting region of each Gene of Interest (GOI) was analyzed through DNA electrophoresis. (C–G) Immunoblot (IB) analysis of the indicated proteins in the heterozygous knock-in RPE-BFP-Cas9 cells expressing GOI fused to V5 and FKBP12^F36V. Cells grown to confluence were treated with either dimethyl sulfoxide (DMSO) or 1 μM dTAG-13 (drug) for 24 hours. Tubulin (IB: Tubulin) serves as a loading control. Molecular weights (kDa) estimated from a protein marker are indicated. The black asterisks indicate the endogenous proteins; The red asterisks indicate the fusion proteins; The blue arrowheads indicate the slower-migrating fusion proteins, which likely come from the ribosome read-through of the P2A peptide (i.e., blasticidin resistance gene-V5-FKBP12^F36V fusion proteins).

Figure 1—figure supplement 1. The genomic DNA sequence flanking the CRISPR-Cas9 targeting region for each gene of interest.

(A-E). The CRISPR-Cas9 targeting genomic DNA sequence, left/right homology arm (LHA/RHA) sequences used in knock-in cell generation, and partial Coding DNA Sequence (CDS) for human DYNC1H1 (A), DYNC1I2 (B), SAR1B (C), AP2S1 (D), and SEC23A (E). The half-arrows indicate 5’ to 3’ direction of the sequences.

To overcome the limitation of efficiently obtaining homozygous degron tag knock-in cells, we sought to develop a method that allows us to use heterozygous cells rather than boosting the chance of obtaining homozygous cells by increasing the HDR/NHEJ ratio. We predicted that when the degron cassette is introduced by HDR following a double-strand break induced by Cas9, the other allele may also be excised by the enzyme and repaired through NHEJ, leading to an insertion/deletion in that allele. To test our hypothesis, we analyzed the genomic DNA sequence of our bulk knock-in cells before the single-cell sorting using the Tracking of Indels by DEcomposition (TIDE) analysis Brinkman et al. (2014), which allows us to analyze the frequency of insertions/deletions in a mixed cell population. The analysis confirmed that the allele that did not have the knock-in cassette had either frame-shift insertion/deletion or its start codon was deleted in almost all cells in four out of five knock-in lines that we tested (Figure 1-figure supplement 2A–E).

Figure 1—figure supplement 2. Genomic DNA sequences of the non-knock-in allele in the heterozygous knock-in cells.

(A-E). The ICE CRISPR Analysis data showing insertion/deletion harbored in the non-knock-in allele of the heterozygous knock-in RPE-BFP-Cas9 cells expressing FKBP12^F36V tagged DYNC1H1(A), DYNC1I2 (B), SAR1B (C), AP2S1 (D), and SEC23A (E). The percentage of insertion/deletion was analyzed by the ICE CRISPR analysis tool following the purification of PCR-amplified genomic DNA. Indel % shows the percentage of sequences that are different from control samples; Model Fit (R²) shows how well the model fits the data; Knockout Score indicates the percentage of cells with either frameshift or +21bp insertion/deletion. Notably, Figure A shows that 8% of sequences harbor a 9-bp deletion that disrupts the start codon of DYNC1H1. (F). A table summarizing the size, an E3 ubiquitin ligase that is used for the degradation (mechanism), and a drug used in this study for each conditional degron tag.

As an exception, FKBP12^F36V-SAR1B knock-in lines showed 96% of the non-frame shifting minus 3 deletions (Figure 1-figure supplement 2C). To validate our knock-in bulk cells at the protein level and assess degradation efficiency across different Proteins of Interest (POIs), we performed immunoblots of the bulk knock-in cells using antibodies against the V5 epitope tag or the endogenous proteins (Figure 1C–G). Consistent with the genomic DNA analyses, the bands that correspond to the endogenous proteins were almost undetectable in the bulk knock-in cell lines (black asterisks in Figure 1D–G), suggesting that most of the heterozygous knock-in cells express only the degron fusion proteins. While the genomic DNA analyses showed the dominant 3bp in-frame deletion (Figure 1-figure supplement 2C), the expression of the endogenous SAR1B protein was greatly diminished in the FKBP12^F36V-SAR1B knock-in line (Figure 1E). Upon 24-hour treatment with the degrader drug, dTAG-13 Nabet et al. (2018), each FKBP12^F36V fusion protein was potently degraded to varying extents (Figure 1C–G). These results indicate that the dTAG system is effective in our bulk heterozygous knock-in cell lines, thereby enabling the bypassing of laborious single-cell cloning.

Comparison of three different degron tag systems in our heterozygous knock-in system

To verify the applicability of our new heterozygous knock-in system to different degron tag systems and compare the efficiency of the different systems, we expanded our focus to three popular degrons: FKBP12^F36V Nabet et al. (2018), IKZF3d (IKZF3 aa130–189) Sievers et al. (2018), and HaloTag Buckley et al. (2015). The three degrons utilize different degradation mechanisms (Figure 1-figure supplement 2F). The compounds dTAG-13 and pomalidomide (an IKZF3d degrader) recruit CRBN Nabet et al. (2018)^; Sievers et al. (2018), whereas the HaloPROTAC3 (a HaloTag degrader) binds to VHL Buckley et al. (2015). We generated a panel of nine knock-in bulk cell lines for three degrons, each crossed with three POIs, SAR1B, AP2S1, and SEC23A (Figure 2A). Immunoblot analyses showed that the endogenous proteins were efficiently removed in all three systems (Figure 2B–D), except for FKBP12^F36V-SAR1B (Figure 2B), which harbors a non-frameshifting 3 bp deletion (Figure 1-figure supplement 2C). While a band slightly upper-shifted than the endogenous AP2S1 was detected in the knock-in cells that harbor HaloTag-AP2S1 (blue asterisk in Figure 2C), genomic DNA analyses showed that the allele mainly has “frame-shifting mutations” (Figure 2-figure supplement 1A). Consistent with the previous study Bondeson et al. (2022), which used an overexpression system to compare different degron tags, our analyses showed that the basal expression level of the degron tag fusion proteins and their degradation efficiency upon the drug treatment greatly vary among the different degrons (Figure 2B–D). Interestingly, we consistently observed the expression level to be the highest with the HaloTag fusions, the lowest with the IKZF3d fusions, and the median with the FKBP12^F36V fusions. Notably, the expression levels of the fusion proteins are generally lower than those of endogenous proteins (compare endogenous proteins marked by black asterisks with the fusion proteins marked by red asterisks in Figure 2C and D), and SAR1B fusion proteins are undetectable with an antibody against the protein (Figure 2B). The degradation efficiency tended to be the highest with the IKZF3d fusions, and the HaloTag fusions had a slightly higher degradation rate than the FKBP12^F36V fusions (Figure 3-figure supplement 1 and 2). Additionally, we occasionally observed upward-shifted bands (blue arrowheads in Figure 1F–G and Figure 2C–D), which likely correspond to blasticidin-resistant gene fusion proteins resulting from ribosome read-through of the P2A peptide Liu et al. (2017). The degradation efficiency of the upshifted bands varied among the fusion proteins.

Figure 2. Generation and validation of the heterozygous knock-in cells harboring different degron-tagged proteins.

(A) A diagram illustrating the design of the donor double-stranded DNA (dsDNA) used for the generation of knock-in cells that have three different degron tags. The three degrons are FKBP12^F36V (for dTAG system), IKZF3d (IKZF3 aa130–189), and HaloTag7. (B-D) Immunoblot (IB) analysis of the indicated proteins in the heterozygous knock-in RPE-BFP-Cas9 cells expressing GOI fused to V5 and indicated degron (FKBP12^F36V, IKZF3d, and HaloTag). The cells grown to confluence were treated with either dimethyl sulfoxide (DMSO) or 1 μM of the indicated drug (drug1: dTAG-13, drug2: pomalidomide, and drug3: HaloPROTAC3) for 24 hours. Tubulin (IB: Tubulin) serves as a loading control. Molecular weights (kDa) estimated from a protein marker are indicated. The black asterisks indicate the endogenous proteins; The red asterisks indicate the fusion proteins; The blue asterisk indicates the band slightly shifted to the upper region than the endogenous AP2S1. The blue arrowheads indicate the slower-migrating fusion proteins, which likely come from the ribosome read-through of the P2A peptide (i.e., blasticidin resistance gene-V5-FKBP12^F36V fusion proteins).

Figure 2—figure supplement 1. Genomic DNA sequences of the non-knock-in allele in the heterozygous HaloTag-AP2S1 knock-in cells.

(A-E). The ICE CRISPR Analysis data showing insertion/deletion harbored in the non-knock-in allele of the heterozygous knock-in RPE-BFP-Cas9 cells expressing AP2S1 fused to HaloTag (Figure 2C). The percentage of insertion/deletion was analyzed by the ICE CRISPR analysis tool following the purification of PCR-amplified genomic DNA. Indel % shows the percentage of sequences that are different from control samples; Model Fit (R²) shows how well the model fits the data; Knockout Score indicates the percentage of cells with either frameshift or +21bp insertion/deletion. While the upper-shifted form of AP2S1 compared to the endogenous AP2S1 was observed in the immunoblot analysis (the blue asterisk in Figure 2C), the cells have dominant −2 deletions.

Figure 3—figure supplement 1. Quantification of the expression levels of the fusion proteins shown in Figure 3.

(A-C). Quantification of protein expression in the cells shown in Figure 3 B–D. The bands shown in the Immunoblot analysis using the V5-tag antibody were quantified. Fusion protein level (%) was calculated by normalizing the expression level of the drug-treated cells to the DMSO-treated cells. The data is combined from two independent experiments. Error bars represent Mean ± SEM

Figure 3—figure supplement 2. Quantification of the expression levels of the fusion proteins compared with those of the endogenous one.

(A-C). Quantification of protein expression in the cells shown in Figure 3B–D. The bands shown in the Immunoblot analysis using the SAR1B antibody (A), AP2S1 antibody (B), and SEC23A antibody (C) were quantified. Fusion protein level (%) was calculated by normalizing the expression level of the DMSO or drug-treated cells to the respective endogenous protein.

Modulation of the expression level of the degron fusion proteins through mini promoters

As described above, we often observed lower expression levels of the degron fusion proteins than those of endogenous proteins (Figure 2B–D). The lower expression may compromise the function of the proteins even before adding the degrader drugs. Therefore, we sought to develop an approach to modulate the expression of the fusion proteins, so that they can retain the function until the degradation is induced through the degrader drugs. To do this, we incorporated either an enhancer-less core sequence (from −71 to −1) of the cytomegalovirus promoter Boshart et al. (1985) (mini-CMV) or a promoter derived from the human EF-1 gene that lacks the first intron Kim et al. (1990) (short-EF) in front of the BlastR gene in the dsDNA donor cassette (Figure 3A). Including the original constructs without the exogenous promoters (Figure 2A), we generated a panel of nine heterozygous knock-in cell lines (three promoters cross three degrons) for each of the three different POIs. Immunoblot analyses showed that endogenous proteins were eliminated in 21 out of 27 cell lines (Figure 3B–D). Endogenous proteins were still expressed in the remaining six cell lines, likely because most of these cells have in-frame insertion/deletion (Figure 1-figure supplement 2C and Figure 3-figure supplement 3). The assay also revealed that the integration of short-EF promoter consistently provided higher expression than the constructs without the exogenous promoters, whereas enhancement of expression by the mini-CMV promoter was less pronounced and depended on the tags and POIs (compare “original” vs “mini-CMV” in each tag and POI in Figure 3B–D, Figure 3-figure supplement 1A–C, and Figure 3-figure supplement 2A–C). Notably, with the addition of a short-EF promoter, we were able to bring the expression level of SAR1B fused with HaloTag to a level closer to the endogenous one (Figure 3B and Figure 3-figure supplement 2A). Even with the short-EF promoter, we failed to achieve a comparable expression level to the endogenous proteins with the IKZF3d fusion proteins (Figure 3B–D and Figure 3-figure supplement 2A–C). When we added the degrader drugs, the expression levels of the fusion proteins regulated by the exogenous promoters diminished to a similar extent as those without the mini-promoters in all the fusion proteins that we tested (Figure 3B–D, Figure 3-figure supplement 1A–C, and Figure 3-figure supplement 2A–C).

Figure 3. Generation and validation of the heterozygous knock-in cells that express degron-tagged proteins regulated by different mini-promoters.

(A) A diagram showing the design of the donor double-stranded DNA (dsDNA) to incorporate the mini-promoter (mini-CMV promoter and short-EF promoter) in the degron-tag knock-in cells. (B-D) Immunoblot (IB) analysis of the indicated proteins in the heterozygous knock-in RPE-BFP-Cas9 cells that express the gene of interest fused to V5 and indicated degron (FKBP12^F36V, IKZF3d, and HaloTag under the control of the original promoter, mini-CMV promoter, or the short-EF promoter. The cells grown to confluence were treated with either dimethyl sulfoxide (DMSO) or 1 μM indicated drug (drug1: dTAG13, drug2: pomalidomide, and drug3: HaloPROTAC3) for 24 hours. Tubulin (IB: Tubulin) serves as a loading control. Molecular weights (kDa) estimated from a protein marker are indicated. The black asterisks indicate the endogenous proteins; The red asterisks indicate the fusion proteins; The blue asterisk indicates the band slightly shifted to the upper region than the endogenous AP2S1. Quantification data are available in Figure 3-figure supplement 1 and 2.

Figure 3—figure supplement 3. The ICE CRISPR Analysis data of the mini-promoter-regulated degron-tag knock-in cells.

(A-D). The ICE CRISPR Analysis data showing insertion/deletion harbored in the non-knock-in allele of the heterozygous knock-in RPE-BFP-Cas9 cells expressing FKBP12F36V-AP2S1 under the control of short-EF promoter (A), HaloTag-SEC23A under the control of short-EF promoter (B), FKBP12F36V-SEC23A under the control of mini-CMV (C), FKBP12^F36V-SEC23A under the control of short-EF promoter (D), and IKZF3d-SEC23A under the control of mini-CMV (E). The percentage of insertion/deletion was analyzed by the ICE CRISPR analysis tool following the purification of PCR-amplified genomic DNA. Indel % shows the percentage of sequences that are different from control samples; Model Fit (R²) shows how well the model fits the data; Knockout Score indicates the percentage of cells with either frameshift or +21bp insertion/deletion. Notably, the weak bands detected in the cells expressing IKZF3d-SEC23A under the control of a minimal CMV promoter (Figure 3D) cannot readily explain the dominant +1 insertion detected in E. A possible explanation for this is that most of the cells may have a huge deletion that prevented the primers from binding, thus the actual deletion was not reflected in the ICE analysis.

Together, the data suggest that we can control the basal expression levels of the degron fusion proteins through the mini-promoters without affecting their drug-induced degradation efficiency.

Functional validation of the heterozygous degron knock-in cells.

Conditional protein degradation allows for precise and controlled analysis of essential genes, enabling researchers to understand cellular functions by rapidly inducing the degradation of specific proteins. The kinetics of degradation can vary and may be unpredictable, depending on both the degron and the target proteins involved Bondeson et al. (2022). To test if the degrader drugs induce rapid removal of the proteins, and thus inhibit the function of the target proteins in our heterozygous knock-in cells, we chose three proteins that are known to be essential for cell division or growth: DYNC1H1 Raaijmakers et al. (2013), DYNC1I2 Raaijmakers et al. (2013), and AP2S1 Zimmerman et al. (2025) (Table 1).

Table 1.

The abbreviation of the POIs used in this study.

GOI	Name	MW (kDa)	UniprotID
DYNC1H1	dynein cytoplasmic 1 heavy chain 1	532	Q14204
DYNC1I2	dynein cytoplasmic 1 intermediate chain 2	68	Q8TCX1
SAR1B	secretion-associated ras-related GTPase 1B	22	Q9Y6B6
SEC23A	SEC23 homolog A, COPII component	82	Q15436
AP2S1	adaptor related protein complex 2 subunit sigma 1	17	P53680

Cytoplasmic dynein 1, a minus-end-directed motor complex, is essential for mitosis. It plays crucial roles in mitotic spindle organization Vaisberg et al. (1993)^; Faulkner et al. (2000), chromosome movement Sharp et al. (2000), and kinetochore protein transport Howell et al. (2001). It is also important for maintaining the structure and the localization of the Golgi apparatus Harada et al. (1998)^; Palmer et al. (2009). Depletion of dynein subunits, such as DYNC1H1 or DYNC1I2, has been shown to cause mitotic arrest Raaijmakers et al. (2013) and the dispersal of the Golgi apparatus Harada et al. (1998)^; Palmer et al. (2009).

We used the cells harboring either DYNC1H1 or DYNC1I2 fused with FKBP12^F36V (Figure 1C and D) as model cases to evaluate degradation kinetics and functional defects upon addition of the degrader drug, dTAG-13, in our heterozygous knock-in cell lines. We found that DYNC1H1 and DYNC1I2 degradation began within 1 hour after administering the degrader drug and reached the maximum efficiency within 3 hours for the DYNC1H1-fusions (Figure 4A and B) and 6 hours for the DYNC1I2-fusions (Figure 5A and B), respectively. The degradation effect persisted for 24 hours for both fusion proteins without replacing the media (Figure 4A–B and 5A–B). Next, we investigated the impact of dynein1 subunit depletion on mitotic progression and Golgi positioning. Although the expression level of the fusion proteins was 50% or 35% for DYNC1H1 and DYNC1I2 before adding the degrader drugs, consistent with the cells having only one knock-in allele (heterozygous), the cells did not show apparent mitotic arrest or the Golgi dispersal (Figure 4-figure supplement 1A and B). This suggests that the cells can still retain the function of cytoplasmic dynein with the reduced expression of the subunits and that the fusion proteins are functionally intact. Upon addition of the drug, we started to see a statistically significant increase in the mitotic index at 6 hours, and the percentage of the cells arrested at mitosis increased over the time course of drug treatment in both knock-in cell lines (Figure 4C–D and Figure 5C–D). Concurrently, the total number of cells per micrograph decreased over time in the cells treated with the degrader drug, suggesting that the cells depleted of DYNC1H1 likely detach from the plate following the mitotic arrest (Figure 4E). The total number of cells did not significantly decrease in the cells depleted of DYNC1I2 (Figure 5E), likely reflecting the milder effect of DYNC1I2 depletion compared with the DYNC1H1 removal. When we analyzed the Golgi structure using an anti-TGN46 antibody, which recognizes the TransGolgi network Lujan et al. (2024), we observed Golgi dispersal within 6 hours of the drug treatment in the DYNC1H1 line (Figure 4G); Golgi dispersal was observed with 3 hours of drug treatment in the DYNC1I2 line (Figure 5G). This suggests a more pronounced effect of DYNC1I2 depletion compared with the DYNC1H1 removal, likely due to the direct interaction of the dynein1 intermediate chain with the golgin160 for the Golgi positioning Yadav et al. (2012). These data suggest that the function of the dynein subunits is intact before the drug treatment and is quickly inhibited following the drug administration in both heterozygous knock-in cells. In contrast, the degrader drug dTAG-13 did not visibly affect the mitotic cell count or the Golgi structure of the parental cells (Figure 4-figure supplement 1A–B).

Figure 4. The degradation kinetics and function assay of cytoplasmic dynein1 in the heterozygous knock-in cells expressing FKBP12^F36V-DYNC1H1.

(A) Immunoblot (IB) analysis of V5 (IB: V5) and DYNC1H1 (IB: DYNC1H1) in the heterozygous knock-in RPE-BFP-Cas9 cells expressing FKBP12^F36V tagged DYNC1H1. The cells grown to confluence were treated with either dimethyl sulfoxide (DMSO) or 1 μM dTAG-13 (drug) for the indicated time. Tubulin (IB: Tubulin) serves as a loading control. Molecular weights (kDa) estimated from a protein marker are indicated. (B) Quantification of the DYNC1H1 protein level shown in Figure 4A. DYNC1H1 protein level (%) was calculated by normalizing the expression level at each time point to that of the control cells at the same time point. The data were averaged from three independent experiments. Error bars represent Mean ± SEM. (C) Mitosis arrest assay in the heterozygous knock-in RPE-BFP-Cas9 cells expressing FKBP12^F36V tagged DYNC1H1. The cells were treated with either dimethyl sulfoxide (DMSO) or 1 μM dTAG-13 (drug) for the indicated time, fixed, and stained with DAPI. The representative image from each condition is shown. The red arrowheads indicate the mitotic cells. Scale bar: 50 μm (D) Quantification of the percentage of the mitotic cells shown in Figure 4C. The data were averaged from three independent experiments. Error bars represent ± SEM. Statistics obtained through comparing the drug-treated cells and the control at each time point by multiple unpaired t-tests. (E) Quantification of the total cell number of the cells per microscope image (0.56 mm²) from the experiments shown in Figure 4C. The data were averaged from three independent experiments. Error bars represent ± SEM. Statistics obtained through comparing the drug-treated cells and the control at each time point by multiple unpaired t-tests. (F) The Golgi dispersal assay in the heterozygous knock-in RPE-BFP-Cas9 cells expressing FKBP12^F36V tagged DYNC1H1. The cells were treated with either dimethyl sulfoxide (DMSO) or 1 μM dTAG-13 (drug) for the indicated time, fixed, and stained with anti-TGN46 antibody. Scale bar: 10 μm (G) Quantification of the percentage of the cells with dispersed Golgi in the cells shown in Figure 4F. The data were averaged from three independent experiments. Error bars represent ± SEM. Statistics obtained through comparing the drug-treated cells and the control at each time point by multiple unpaired t-tests. n.s., not significant; *p < 0.05, **p < 0.01, ***p < 0.001

Figure 5. The degradation kinetics and functionassay of cytoplasmic dynein1 in the heterozygous knock-in cells expressing FKBP12^F36V-DYNC1I2.

(A) Immunoblot (IB) analysis of V5 (IB: V5) and DYNC1I2 (IB: DYNC1I2) in the heterozygous knock-in RPE-BFP-Cas9 cells expressing FKBP12^F36V tagged DYNC1I2. The cells grown to confluence were treated with either dimethyl sulfoxide (DMSO) or 1 μM dTAG-13 (drug) for the indicated time. Tubulin (IB: Tubulin) serves as a loading control. Molecular weights (kDa) estimated from a protein marker are indicated. (B) Quantification of the DYNC1I2 protein level shown in Figure 5A. DYNC1I2 protein level (%) was calculated by normalizing the expression level at each time point to that of the control cells at the same time point. The data were averaged from three independent experiments. Error bars represent Mean ± SEM. (C) Mitosis arrest assay in the heterozygous knock-in RPE-BFP-Cas9 cells expressing FKBP12^F36V tagged DYNC1I2. The cells were treated with either dimethyl sulfoxide (DMSO) or 1 μM dTAG-13 (drug) for the indicated time, then fixed and stained with DAPI. The representative image from each condition is shown. Scale bar: 50 μm (D) Quantification of the percentage of the mitotic cells shown in Figure 5C. The data were averaged from three independent experiments. Error bars represent ± SEM. Statistics obtained through comparing the drug-treated cells and the control at 24 hours by an unpaired t-test. (E) Quantification of the total cell number of the cells per microscope image (0.56 mm²) from the experiments shown in Figure 5C. The data were averaged from three independent experiments. Error bars represent ± SEM. Statistics obtained through comparing the drug-treated cells and the control at 24 hours by an unpaired t-test. (F) The Golgi dispersal assay in the heterozygous knock-in RPE-BFP-Cas9 cells expressing FKBP12^F36V tagged DYNC1I2. The cells were treated with either dimethyl sulfoxide (DMSO) or 1 μM dTAG-13 (drug) for the indicated time, fixed, and stained with anti-TGN46 antibody and DAPI. Scale bar: 10 μm (G) Quantification of the percentage of the cells with dispersed Golgi in the cells shown in Figure 5F. The data were averaged from three independent experiments. Error bars represent ± SEM. Statistics obtained through comparing the drug-treated cells and the control at 24 hours by an unpaired t-test. n.s., not significant; *p < 0.05, **p < 0.01, ***p < 0.001

Figure 4—figure supplement 1. The control experiments for the mitosis arrest assay and the Golgi dispersal assay shown in Figure 4.

(A). Mitosis arrest assay in control (the parental RPE-BFP-Cas9 cells) and the heterozygous knock-in RPE-BFP-Cas9 cells expressing FKBP12^F36V-DYNC1H1 cells. The cells were stained with DAPI. Representative images of control (DMSO) or 1 μM dTAG-13 (drug)- treated cells at the indicated time point are shown. The red arrowheads indicate the mitotic cells. Scale bar: 50 μm (B). Golgi dispersal assay in control (the parental RPE-BFP-Cas9 cells) and the heterozygous knock-in cells expressing FKBP12^F36V-DYNC1H1. The cells were stained with the indicated antibodies. Representative images of control (DMSO) or 1 μM dTAG-13 (drug)- treated cells at the indicated time point are shown. Scale bar: 10 μm

As a second example, we tested the ability of the heterozygous knock-in cells that harbor HaloTag-AP2S1 to recapitulate the previously reported effect of AP2S1 loss on endocytosis Xu et al. (2024). AP2S1 is a subunit of the adaptor protein complex 2 (AP-2) Collins et al. (2002), a critical regulator for clathrin-mediated endocytosis Motley et al. (2003). AP2 binds to membrane protein cargos, initiates the assembly of clathrin-coated pits, and promotes internalization of the cargo proteins, which include transferrin receptors Xu et al. (2024)^; Motley et al. (2003). The loss of AP2S1 was shown to destabilize other subunits Zimmerman et al. (2025)^; Li et al. (2025) and inhibit endocytosis of the transferrin receptors Xu et al. (2024). In the heterozygous knock-in cell lines expressing HaloTag-AP2S1 under the control of the original promoter, the degradation of AP2S1-fusions induced by HaloPROTAC3, a degrader drug for HaloTag proteins, began one hour after degrader administration, reached its 90% degradation efficiency at three hours, and the effects of degradation lasted for 24 hours (Figure 6A and B). We then performed the transferrin internalization assay at 6 hours following the drug administration using the spinning disk confocal microscopy. In the cells treated with DMSO, the dotty patterns of fluorescent transferrin were readily observed in the cytoplasm (Figure 6C and D), indicating that transferrin was successfully internalized via endocytosis in the assay condition. In contrast, we could not detect similar dotty staining in the cytoplasm of the majority of the drug-treated cells (Figure 6C and D), suggesting that endocytosis was efficiently inhibited by AP2S1 depletion. To better visualize internalization of transferrin, we conducted the optical reassignment-based super-resolution imaging Azuma and Kei (2015) to provide a 3D view of the fluorescent transferrin signals. The axial view of the 3D reconstructed images more clearly showed that most of the fluorescent-transferrin signals were detected in the cytoplasm in DMSO-treated cells, whereas the signals remained largely on the plasma membrane in the drug-treated cells (see xz and yz views in the Figure 6E). Notably, when we analyzed another cell line that harbors AP2S1 fusion proteins, short-EF promoter-regulated FKBP12^F36V-AP2S1, the effect of the degrader drugs on endocytosis of transferrin was less pronounced (Figure 6-figure supplement 1A–B). The weaker phenotypes of those cells likely reflect the milder degradation efficiency of the FKBP12^F36V-AP2S1 fusion protein than the HaloTag-AP2S1 fusion (Figure 3C and Figure 3-figure supplement 2B). This highlights the importance of carefully evaluating both expression level and functional effects upon the degrader drug treatment for each POI fusion.

Figure 6. The degradation kinetics and function assay of AP2S1 in the heterozygous knock-in cells expressing HaloTag-AP2S1.

(A) Immunoblot (IB) analysis of V5 (IB: V5) and AP2S1 (IB: AP2S1) in the heterozygous knock-in RPE-BFP-Cas9 cells expressing HaloTag-AP2S1. The cells grown to confluencewere treated with either dimethyl sulfoxide (DMSO) or 1 μM HaloPROTAC3 (drug) for the indicated time. Tubulin (IB: Tubulin) serves as a loading control. Molecular weights (kDa) estimated from a protein marker are indicated. (B) Quantification of the AP2S1 protein level shown in Figure 6A. AP2S1 protein level (%) was calculated by normalizing the expression level at each time point to that of the control cells at the same time point. The data were averaged from three independent experiments. Error bars represent Mean ± SEM. (C) Transferrin internalization assays in the heterozygous knock-in RPE-BFP-Cas9 cells expressing HaloTag-AP2S1. The cells were treated with either dimethyl sulfoxide (DMSO) or 1 μM HaloPROTAC3 (drug) for 6 hours and incubated with transferrin-Alexa Fluor 488. Live-cell imaging was performed on a spinning disk confocal microscope. The individual image is from a representative z-slice. Scale bar: 10 μm (D) Quantification of the percentage of the cells with internalized transferrin shown in Figure 6C. The data were averaged from three independent experiments. Error bars represent ± SEM. Statistics obtained through comparing the drug-treated cells and the control by an unpaired t-test. ***p < 0.001 (E) 3D Super-resolution live-cell imaging was used to assess transferrin internalization assays in the condition described in Figure 6C. The representative 100x SoRa super-resolution z-slice images with orthogonal views are shown. For easier visualization, we have intensively enlarged the z-thickness of the orthogonal views. Scale bar for the x & y-axis: 10 μm; scale bar for the z-axis: 2 μm.

Figure 6—figure supplement 1. Transferrin internalization assays in the heterozygous knock-in cells expressing FKBP12^F36V-AP2S1 under the control of short-EF promoter.

Tra (A). Transferrin internalization assays in the heterozygous knock-in RPE-BFP-Cas9 cells expressing short EF promoter-regulated FKBP12^F36V-AP2S1. The cells were treated with either dimethyl sulfoxide (DMSO) or 1 μM HaloPROTAC3 (drug) for 6 hours and incubated with transferrin-Alexa Fluor 488. Live-cell imaging was performed on a spinning disk confocal microscope. The individual image is from a representative z-slice. Scale bar: 10 μm (B). Quantification of the percentage of the cells with internalized transferrin shown in Figure 6-figure supplement 1A. The data is from a representative experiment.

Nevertheless, all the functional validation tests confirmed that the heterozygous knock-in cells generated through our strategy can be used to test the function of essential genes.

Discussion

Conditional degron tag systems are important tools to investigate the function of proteins through rapid degradation of them; however, two major issues restrict their application: 1) difficulty in obtaining homozygous knock-in cells that express degron tag fusion proteins, and 2) adding a degron tag may alter the expression level of the protein even in the absence of the degrader drug. In the current work, we confirm the presence of the two problems through the generation of a good number of the degron-tag knock-in cells. Only 4 of 183 single-cell clones that we analyzed had the knock-in cassette in both alleles (Figure 1B), indicating the low yield of homozygous knock-in cells. While the extent varied across different degron tags and proteins of interest, the expression levels of the degron-tag fusion proteins were often lower than those of endogenous proteins (Figure 1C–G and 2B–D). Reduced expression of fusion proteins has been reported even in knock-in cells that express the fusion proteins from their endogenous gene locus Cho et al. (2022)^; Taylor et al. (2025), likely because foreign tags alter mRNA stability, protein folding, and protein interaction. To mitigate these problems and improve the applicability of these valuable tools, we report here a strategy to efficiently and rapidly obtain the degron tag knock-in cells that can be used for functional assays. Our method was established based on the observation that the allele that does not have the knock-in cassette is edited by Cas9 in virtually all the cells that we analyzed, and the majority of the non-knock-in alleles had frame-shift insertion/deletion (Figure 1-figure supplement 2 and Figure 2-figure supplement 1). Consistent with the frame-shift insertion/deletion, immunoblot analyses failed to detect the bands that correspond to the endogenous proteins in 23 out of 29 knock-in cells that we created (Figure 1C–D and Figure 3A–C). This allowed us to skip single-cell cloning and use the bulk cells for functional assays. Subsequent analyses confirmed that the function of the target proteins is rapidly and efficiently inhibited upon addition of the degrader drugs in bulk knock-in cells (Figures 4–6). We also introduce a method to mitigate the reduction in the basal expression level of degron-tagged POIs by inserting minimal promoters at the beginning of the donor cassette (Figure 3). We tested two different mini-promoters and found that the short-EF promoter consistently provides higher expression than the one without mini-promoters, whereas the effect of minimal CMV varies among POIs and degron tags (Figure 3, Figure 3-figure supplement 1 and 2).

While our approach has multiple advantages over commonly used methods that select homozygous clones for degron tag knock-in, there are several potential issues. In the subsequent sections, we will discuss the pros and cons of our methodology and a possible strategy that can be used for further improvement.

Efficiency of obtaining homozygous knock-in cells.

Consistent with the previously reported difficulty in obtaining RPE cells that harbor a knock-in cassette in both alleles Ghetti et al. (2021)^; Miyamoto et al. (2015), our results showed that the chances of gaining homozygous knock-in cells are very low (2.2% as shown in Figure 1B). In contrast, a recent study reported a much higher percentage of homozygous knock-in RPE cells, while the efficiency is affected by GOIs and crRNA sequences Philip et al. (2025). There are several methodological differences in the generation of knock-in cells between our report and the previous one. To create knock-in cells, we transfected RPE cells stably expressing BFP-Cas9 with a 1:500 ratio of donor dsDNA/guide RNA, whereas the previous study co-transfected RPE cells with a vector containing both Cas9 and guide RNAs with a plasmid carrying a donor template. Several studies suggest that the stoichiometry ratio of Cas9, guide RNA, and donor dsDNA is an essential factor influencing HDR-mediated knock-in efficiency Lin et al. (2014)^; Shirk et al. (2026)^; Chenouard et al. (2023)^; Schumann et al. (2015). While the intracellular concentrations of those components are difficult to predict in our current approach, stoichiometry optimization may further enhance HDR activity and improve the efficiency of successfully generating knock-in cells. Another notable difference is the size of the 5’ and 3’ homology arms. We adopted the previously established protocol, which uses 40 base pair (bp) homology arms flanking a donor gene cassette Ghetti et al. (2021), whereas the other study incorporated 500–1000 bp homology arms into a donor plasmid Philip et al. (2025). The short homology arm method can speed up the cell line generation process, as the homology arm can be added to the pair of PCR primers when the donor template is generated, thus eliminating the need for molecular cloning. The difference in the homology arm length may explain the distinct percentage of homozygous knock-in cells between our study and the previous study Philip et al. (2025), as shown previously Li et al. (2014)^; Zhang et al. (2017)^; Chai et al. (2024). While our approach does not require obtaining homozygous cells, increasing the length of the homology arm by synthesizing longer primers may further improve the efficiency to successfully generate knock-in cells via higher HDR activity. Notably, however, evidence suggests that the optimal stoichiometric ratio of Cas9, guide RNA, and donor dsDNA may play a dominant role in determining the HDR-mediated knock-in efficiency, potentially overriding the effects of homology arm length on the integration efficiency Shirk et al. (2026). In addition to optimizing these ratios, several other strategies may be considered in future studies to improve the HDR knock-in efficiency. We utilized dsDNA donor templates for the HDR because they are commonly used for the larger inserts and are easy to prepare Jin et al. (2025), though they often suffer from lower knock-in efficiency Renaud et al. (2016), higher cellular toxicity Shy et al. (2023), and the risk of random integration Roth et al. (2018). In contrast, single-strand DNA (ssDNA) templates often provide higher HDR efficiency Renaud et al. (2016), lower cellular toxicity Shy et al. (2023), and lower random integration Roth et al. (2018), but they are typically limited to the smaller inserts and can be challenging to prepare Velangani et al. (2026). A notable exception is circular single-stranded DNA (cssDNA), generated by M13 phage, which can reach a size up to 9 kb Xie et al. (2025)^; Ducani et al. (2013)^; Shepherd et al. (2019). CssDNA can be used as a vector to obtain single-stranded templates for HDR. Studies suggest that cssDNA serves as an HDR donor with superior efficiency for CRISPR-based editing compared to linear ssDNA Iyer et al. (2022)^; Letort et al. (2025). Even with the optimal stoichiometric ratio of Cas9, guide RNA, and an ideal format of donor template, it is challenging to ensure the donor template is present at the double-strand break (DSB) site. To bridge these gaps, novel strategies have been developed, such as covalently tethering the DNA donor template to the Cas9-guide RNA ribonucleoprotein (RNP) to ensure the presence of the donor DNA at the DSB site with the RNP complex Aird et al. (2018)^; Savic et al. (2018). Additionally, the use of CRISPR lipid nanoparticle–spherical nucleic acids (CRISPR LNP–SNAs) for the delivery of CRISPR gene editing components has shown promise in enhancing cellular uptake and improving HDR editing efficiency Han et al. (2025). Use of these new technologies may further improve the efficacy of our strategy.

Advantages of our approach

An apparent benefit of our methodology is the ability to bypass single-cell cloning. This not only speeds up the generation of knock-in lines but also mitigates the impact of clone-specific artifacts. The previous study underscored the presence of heterogeneity in supposedly homogenous cell lines Westermann et al. (2022), and avoiding the selection of clonal populations would increase the reproducibility of biological experiments. Second, the use of heterozygous cells that carry a knock-in cassette as well as the null GOI allele for functional assays may have an additional advantage. One functional copy of the heterozygous cell is often enough to maintain the function of the gene Bartha et al. (2018), yet the function is more easily disrupted by the degrader drug, as the expression level is already half compared to homozygous cells (as shown in Figure 4B, 5B, and 6B). We also addressed the issue in which degron tags themselves decrease the expression levels of fusion proteins by inserting minimal promoters upstream of the GOIs. We showed that minimal promoters can indeed increase fusion protein expression, and that the extent of this increase varies widely across different degron tags and GOIs (Figure 3). This suggests that we need to test combinations of tags and mini-promoters to find the optimal expression level of the POI, at which the fusion proteins can maintain physiological function in the absence of the drug, but cannot do so upon degradation by the degrader drug. Given that our method saves time by skipping the single-cell cloning step for creating degron-tag knock-ins, it allows users to test multiple combinations to find the “sweet spot”.

Limitations and potential modifications for improvement

One disadvantage of our current approach is that it is limited to N-terminal tagging, as we need to knock out the non-knock-in allele. Our strategy is not applicable if the degron tag needs to be fused to the C-terminus of the POI. Secondly, the decreased basal expression level of the degron-tag protein (Figure 1D–F and 2B–C) could be due to our bicistronic donor construct that contains a selection marker (blasticidin resistance gene) followed by P2A and degron-tag. The self-cleaving peptide P2A induces ribosomal skipping during translation, creating two proteins from a single open reading frame Donnelly et al. (2001). It was shown that the expression level of the second gene is lower than the first gene, possibly because of the ribosome fall-off Liu et al. (2017)^; Donnelly et al. (2001). Therefore, removing the selection marker and P2A site after successful knock-in via a Cre-loxP system may be an alternative strategy to alleviate the reduced expression of degron-tagged fusion proteins. However, collecting cells in which the selection marker and P2A are cleaved off requires an additional strategy, such as single-cell sorting or fusing the selection marker to GFP to sort GFP-negative cells. The addition of GFP to the knock-in construct may further decrease the homologous recombination efficiency, as the size of the donor DNA increases. Thirdly, although controlling the basal expression level of a degron-fusion protein by inserting a mini-promoter has proven effective (Figure 3), the mini-promoter may alter dynamic gene regulation. While the short-EF promoter and minimal CMV promoter that we used are considered core promoters Ede et al. (2016)^; Li et al. (2023), binding sites for basal transcription factors (e.g., TFIID) Kadonaga (2012)^; Danino et al. (2015), the core promoters can also affect the size of transcriptional bursts, which are primarily regulated by enhancers Larsson et al. (2019). Thus, gene expression kinetics may be changed through the mini-promoter insertion. Lastly, although we validated this system in near-diploid RPE1-hTERT cells, testing this methodology across a broader range of cell lines remains a priority to demonstrate its robustness across diverse genetic backgrounds.

Recommended strategy for generating degron-tag knock-in cells.

In addition to establishing the heterozygous knock-in degron-tag system, we compared three widely used degron tag systems: the dTAG system, the Halo-PROTAC system, and the IKZF3d system. The results showed that the expression levels and degradation efficiencies of FKBP12^F36V and HaloTag are comparable and vary among GOIs, while the expression of the fusion proteins is often lower than the intrinsic expression levels of the GOIs. In contrast, IKZF3d-tagged proteins often exhibit poor expression levels relative to the endogenous protein (Figure 2, Figure 3, and Figure 3-figure supplement 1 and 2). Additionally, as mentioned above, inserting a short EF promoter can ameliorate the reduced expression of degron-fusion proteins. With these noted, the recommended strategy for generating heterozygous knock-in cell lines is to make 4 cell lines per GOI, each expressing two different degron tags (FKBP12^F36V and HaloTag) under the control of the original or short EF promoter. Each fusion protein likely has a different expression level and degradation efficiency, allowing users to find the best conditions for functional assays.

Conclusion

By shifting from the single-cell cloning to the heterozygous approach, our method addresses the labor-intensive nature of the traditional approach for generating degron tag knock-in cells. The method also reduces phenotypic variability often associated with clonal isolation and provides a strategy to mitigate the instability of the degron-tag fusion proteins. Thus, our new method expands the applicability and scalability of the valuable degron-tag technology.

Methods and Materials

Vector construction

The donor template vectors that served as templates for PCR reactions to generate donor double-stranded DNA (donor dsDNA) were generated in this study and are available from Addgene (#251478, #251479, #251480, #251481, #251482, #251483, #251484, #251485, and #251486). They contain the donor cassette that consists of a mini-promoter (optional), a blasticidin resistance gene, a V5 epitope tag, a specific conditional degron tag, and a rigid linker (EAAAK₃). The V5-degron tag-rigid linker cassette (Addgene plasmids #185775, #185774, #185772) and blasticidin resistance gene taken from pWPXLd/LAP-C/blast/long EF/DEST Kanie et al. (2025), mini-promoters taken from pCS2+ vector and pWPXLd/LAPC/blast/short EF/DEST Kanie et al. (2025) were amplified by polymerase chain reaction (PCR) and assembled into either pUC19 or pENTR221 using Gibson assembly master mix (E2611S, NEB). The insert sequence was validated at Eurofins Genomics by Sanger Sequencing. Sequence and maps for all the vectors are available through Addgene (#87079).

Cell line and Cell culture

RPE cells stably expressing Blue Fluorescent Protein (BFP)-Cas9 (RPE-BFP-Cas9) Kanie et al. (2017), and 293T cells were grown in DMEM/F-12 (12400–024, Invitrogen) supplemented with 10% FBS (100–106, Gemini), 1xGlutaMax (35050–061, Gibco), 100 U/mL Penicillin-Streptomycin (15140122, Thermo Fisher Scientific) at 37 °C in 5% CO₂. Both cell lines were authenticated via a short-tandem-repeat-based test. The authentication was performed by the MTCRO-COBRE Cell Line Authentication Core of the University of Oklahoma Health Science Center. Mycoplasma negativity of the original cell lines (RPE-BFP-Cas9 and 293T) grown in antibiotic-free media was confirmed by a PCR-based test (G238, Applied Biological Materials).

Compounds

FKBP12 PROTAC dTAG-13 (HY-114421, MedChemExpress), Pomalidomide (HY-10984, MedChemExpress), HaloPROTAC3 (AOB36136, Aobious), were dissolved in dimethyl sulfoxide (DMSO) at a stock concentration of 1 mM and stored in −20°C. Working solutions were prepared by adding stock solution to the culture media and typically contained 0.1% of DMSO.

gRNA preparation

crRNAs were designed using the ‘Custom AltR CRISPR-Cas9 guide RNA’ tool and ordered through IDT. crRNA and tracrRNA (IDT) were resuspended in Nuclease Free Duplex Buffer (IDT) to a final concentration of 100 μM on ice. To prepare guide RNA, 5 μL of 100 μM crRNA and 5 μL of 100 μM tracrRNA were mixed, heated at 95 °C for 5 min, and cooled down to room temperature for 5 min. The annealed gRNA was used for transfection immediately. The crRNA sequence for each GOI is listed in Table 2.

Table 2.

crRNA sequence used in this study.

GOI	crRNAs (5’ to 3’)
DYNC1H1	TCCCGAGCGCGACACCATGT
DYNC1I2	GTCAGACAAAAGTGAATTAA
SAR1B	GTAAATCCAATCAAATATGA
SEC23A	CTCCACAATGACAACCTATT
AP2S1	GATGAGGATAAAGCGGATCT

Homology-directed repair (HDR) donor double-strand DNA (donor dsDNA) preparation

Donor dsDNAs carrying Left Homology Arm (LHA)-promoter (optional)-blasticidin resistance-P2A-V5-degron-Right Homology Arm (RHA) cassette were generated by PCR using a pair of ~60 bp oligonucleotides that consists of 40 bp of homology arm and ~20 bp that anneals with the donor DNA cassette, the donor template vector mentioned above as a template, and Q5 High-Fidelity DNA Polymerase (M0491S, New England BioLabs). The PCR products were separated by DNA electrophoresis and purified using the Zymoclean Gel DNA Recovery Kit (D4002, Zymo Research) according to the manufacturer’s instructions. The concentrations of the recovered Donor dsDNAs were measured using a NanoDrop. The primers used for amplifying each donor dsDNA are listed in Table 3. The primer sequence consists of the homology arm (shown in lowercase, with the optimized codon shown in red), followed by the Kozak sequence (optional, shown in Bold), and the annealing sequence (shown in uppercase) to the vectors containing the donor template.

Table 3.

List of the PCR primers for donor dsDNA preparation.

Name	Sequence (5’ to 3’)
DYNC1H1- LHA (orig)	cgccttctcatcgctcctggaaggtcccgagcgcgacaccGGCAGCATGGCCAAGCCTTTGTCTCAAGAA
DYNC1H1- LHA (mCMV)	cgccttctcatcgctcctggaaggtcccgagcgcgacaccACGTTCCATTGACGTAAATGGGCG
DYNC1H1- LHA (sEF)	cgccttctcatcgctcctggaaggtcccgagcgcgacaccGTCGACGATAAGCTTTGCAAAGATGGATAAA
DYNC1H1- RHA	cggccgagccgtcctcgccgccgccgcccccgggctccgaACTGACGGATCCAGCCTT
DYNC1I2- LHA (orig)	tatatgtttctaaggtcacaaacatgtcagacaaaagtgaGCCAGCATGGCCAAGCCTTTGTCTCAAGAA
DYNC1I2- LHA (mCMV)	tatatgtttctaaggtcacaaacatgtcagacaaaagtgaACGTTCCATTGACGTAAATGGGCG
DYNC1I2- LHA (sEF)	tatatgtttctaaggtcacaaacatgtcagacaaaagtgaGTCGACGATAAGCTTTGCAAAGATGGATAAA
DYNC1I2- RHA	gggccagtcgctgcttcttacgttccaactcagcctttaaACTGACGGATCCAGCCTT
SAR1B-LHA (orig)	cttttttttcagaccctatcagatttggatatgtccttcaGCCAGCATGGGATCCGCCAAGCCTTTGTCTCAAGAA
SAR1B-LHA (mCMV)	cttttttttcagaccctatcagatttggatatgtccttcaACGTTCCATTGACGTAAATGGGCG
SAR1B-LHA (sEF)	cttttttttcagaccctatcagatttggatatgtccttcaGTCGACGATAAGCTTTGCAAAGATGGATAAA
SAR1B-RHA	tgtagcacactgctgaaaccactgtaaatccaatcaaataACTGACGGATCCAGCCTT
SEC23A-LHA (orig)	tttgatacagaaataagaatcaaactccacaatgacaaccATGGCCAAGCCTTTGTCTCAAGAA
SEC23A-LHA (mCMV)	tttgatacagaaataagaatcaaactccacaatgacaaccACGTTCCATTGACGTAAATGGGCG
SEC23A-LHA (sEF)	tttgatacagaaataagaatcaaactccacaatgacaaccGTCGACGATAAGCTTTGCAAAGATGGATAAA
SEC23A-RHA	ctccatctcgttcttcattttgttgaatgaattccaaatacgtcgtACTGACGGATCCAGCCTT
AP2S1-LHA (orig)	gcctctctcacttcctccttctcctcctgctgcccccagaGCCAGCATGGCCAAGCCTTTGTCTCAAGAA
AP2S1-LHA (mCMV)	gcctctctcacttcctccttctcctcctgctgcccccagaACGTTCCATTGACGTAAATGGGCG
AP2S1-LHA (sEF)	gcctctctcacttcctccttctcctcctgctgcccccagaGTCGACGATAAGCTTTGCAAAGATGGATAAA
AP2S1-RHA	cgcgtcttgcctgcccggttctggatgaggataaagcggaACTGACGGATCCAGCCTT

Generation of stable knock-in cell lines expressing a conditional degron tag

RPE-BFP-Cas9 cells were plated into a 24-well plate (FB012929, Fisherbrand) at a density of 5 × 10⁴ cells per well and grown for 24 hours. Before transfection, two separate mixtures were prepared in a 1.5 ml microcentrifuge tube (05–408-129, Fisher Scientific): Mixture 1 [1.5 μl lipofectamine3000 (L3000001, Invitrogen) in 23.5 μl OPTI-MEM (31985062, Gibco)] and Mixture 2 (30 fmol Donor dsDNA, 15 pmol gRNA, and 1 μl P3000 in 24 μl OPTI-MEM). Both mixtures were incubated at room temperature for 5 minutes. Mixture 2 was then added to Mixture 1, mixed thoroughly by tapping the tube, followed by a brief centrifugation, and was incubated at room temperature for 15 minutes. The transfection mixture was added to the RPE-BFP-Cas9 cells and incubated for 72 hours in 5% CO₂ incubator at 37°C. The cell lines were then selected with 10 μg/ml blasticidin (14499, Cayman Chemical) until the negative control cells (without infection) reached complete mortality. The selected cells were expanded and subjected to immunoblotting, immunofluorescence, or genomic PCR analyses to determine knock-in/knock-out efficiency.

Genomic DNA analyses

Genomic DNA used for genomic PCR was prepared by lysing 5 × 10⁵ cells in lysis buffer [10 mM Tris-HCl (pH 7.6), 50 mM NaCl, 6.25 mM MgCl₂, 0.045% NP40, and 0.45% Tween-20] with Proteinase K (FEREO0491, Thermo Scientific, 200 μg/ml final concentration) treatment for 1 hour at 56 °C. Proteinase K was subsequently inactivated by heating the samples to 95°C for 15 minutes. The resulting lysate was used directly as a template for genomic PCR. Genomic PCR was performed using the template described above, primers listed in Table 4, and DreamTaq DNA Polymerase (EP0702, Thermo Scientific). For genotyping shown in Figure 1B, the PCR products were run on a 1% agarose gel in 0.5x TAE buffer containing 0.1 μg/ml ethidium bromide at a constant voltage of 100V until the dye front had migrated across approximately two-thirds of the gel length, and the DNA bands were visualized on EpiChemII Darkroom (UVP). For the TIDE analyses, the PCR products were separated by a 1% agarose gel in 0.5x TAE buffer containing MaestroSafe (MR-031203, Transilluminators), and the bands corresponding to the non-knock-in allele were visualized on Blue Light Transilluminator (NEB-SLB-01W, Transilluminator.com) and excised. The PCR products were purified using the Zymoclean Gel DNA Recovery Kit (D4002, Zymo Research) according to the manufacturer’s instructions and Sanger-sequenced by Eurofins Genomics. The sequences were analyzed using the Inference of CRISPR Edits (ICE CRISPR Analysis, 2025, v3.0. EditCo Bio) to determine the percentage of insertions/deletions.

Table 4.

PCR primers for genotyping used in this study.

Name	Sequence (5’ to 3’)
DYNC1H1-geno-F	GCGCCGAGACTACAAGTCCC
DYNC1H1-geno-R	GCAGACAAAGCGACCCCATTGT
DYNC1I2-geno-F	ACTCATGGCCTGAAGCAGTTCT
DYNC1I2-geno-R	ACAGTGTTATGTCTGCAGTCAGA
SAR1B-geno-F	GCCCATGACCACACCCGGTT
SAR1B-geno-R	AGAAAACTCAGGCTAAGGTGAAAAAAGTAATTTGCC
SEC23A-geno-F	GGCCCAGCTGGTCCTGATTT
SEC23A-geno-R	GGCTGCCACAGGAACAACCATT
AP2S1-geno-F	GTCACTCCACAGGGGTGCTGGA
AP2S1-geno-R	AGCAAGCAAGCTCAAAGCAGGT

Single-Cell Clonal Isolation

Monoclonal cell lines were established from bulk populations of degron tag knock-in cells. Briefly, cells were dissociated using 0.05% Trypsin-EDTA (25300062, Thermo Fisher Scientific) and resuspended in PBS. To remove aggregates, the suspension was passed through a 35 μm nylon mesh. Single-cell sorting was performed on a FACSAria IIIu equipped with a 100 μm nozzle, and doublet depletion was performed using FSC-H vs. FSC-A and SSC-H vs. SSC-A gating. Individual cells were sorted into 96-well plates (12–556-008, Fisher Scientific) containing 100 μL of a 1:1 mixture of fresh DMEM/F12 containing 10% FBS and the conditioned media. The conditioned media were harvested from knock-in bulk cells cultured in DMEM/F-12 containing 10% FBS for 48 hours and filtered through a 0.22 μm PES membrane (431117, Corning).

Immunoblot

For immunoblotting, cells were grown to confluent in a 6-well plate and lysed in 100 μl of IGEPAL CA-630 lysis buffer {25 mM HEPES-NaOH [pH 7.4], 150 mM NaCl, and 0.3% IGEPAL CA-630 (I8896–100ML, SIGMA)} containing 1% Halt^™ Protease Inhibitor Cocktail (87785, Thermo Scientific) followed by clarification of the lysate by centrifugation at 15,000 g at 4°C for 10 min. 72.5 μl of the clarified lysates were then mixed with 25 μl of 4xLaemmli sample buffer (200 mM Tris-HCl [pH 6.8], 8% SDS, 40% glycerol, and 0.02% Bromophenol Blue) and 2.5 μl of 2-mercaptoethanol (M3148, SIGMA), and incubated at 95°C for 5 min. Proteins were separated in 10%, 12%, or top 10%-bottom 12% half-half Bis-Tris protein gels in SDS-PAGE running buffer (25 mM Tris, 192 mM Glycine, 0.1% SDS, pH 8.3), then transferred onto Immobilon^™-FL PVDF Transfer Membranes (IPFL00010, MilliporeSigma) in Towbin Buffer (25 mM Tris, 192 mM glycine, pH 8.3). Membranes were incubated in TrueBlack^® WB Blocking Buffer (23013A-500ML, Biotium) for 30 min at room temperature and then probed overnight at 4°C with the primary antibody listed in Table 5, diluted in TrueBlack^® WB Antibody Diluent (23013B-1L, Biotium). Next, membranes were washed 3 × 5 min in TBST buffer (20 mM Tris, 150 mM NaCl, 0.1% Tween 20, pH 7.5) at room temperature, incubated with the secondary antibodies listed in Table 5 diluted in TrueBlack^® WB Antibody Diluent for 30 min at room temperature, then washed 3 × 5 min in TBST buffer. Membranes were scanned on a Sapphire FL biomolecular imager (Azure Biosystems), and proteins were detected at wavelengths 685 and 784 nm. Western blots were quantified by either Fiji or AzureSpot Pro (Azure Biosystems).

Table 5.

List of antibodies used in this study.

Antigen	Supplier	Cat#	Dilution
V5	Cell Signaling	13202S	1:1000 (WB)
DYNC1H1	Proteintech	12345-1-AP	1:500 (WB); 1:1000 (IF)
DYNC1I2	Proteintech	12219-1-AP	1:1000 (WB)
SAR1B	Proteintech	22292-1-AP	1:1000 (WB)
SEC23A	Abcam	ab179811	1:1000 (WB)
AP2S1	Proteintech	15634-1-AP	1:1000 (WB)
α-Tubulin	Santa Cruz	sc-32293	1:500 (WB); 1:1000 (IF)
TGN46	BioRad	AHP500GT	1:1000 (IF)
CF680 Goat Anti-Mouse	Biotium	20065-1	1:20000 (WB)
CF770 Goat Anti-Rabbit	Biotium	20078-1	1:20000 (WB)
Alexa Fluor 488 Donkey anti-Sheep IgG	Invitrogen	A11015	1:2000 (IF)
Alexa Fluor 568 Goat Anti-Rabbit IgG	Invitrogen	A11036	1:2000 (IF)
Alexa Fluor 647 Goat Anti-Mouse IgG1	Invitrogen	A21240	1:2000 (IF)

Induction of protein degradation by the degrader drugs for immunoblot analyses

Cells were plated in a 6-well plate (12–556-004, Thermo Scientific) at a density of 2 × 10⁵ cells/well and grown for 48 hr. After the cells were treated with the indicated compound or DMSO for 0-, 1-, 2-, 3-, 6-, 9-, 12-, and 24-hr, all cell lysates were prepared simultaneously for immunoblotting.

Immunofluorescence

For wide-field microscopy, cells were grown on acid-washed 12 mm #1.5 round coverslips (71861–057, Electron Microscopy Sciences) and fixed in 4% paraformaldehyde (15710, Electron Microscopy Sciences) diluted in phosphate-buffered saline (PBS) for 15 min at room temperature. The primary antibodies used for immunofluorescence are listed in Table 5. After permeabilization in immunofluorescence (IF) buffer [3% bovine serum albumin (BP9703100, Fisher Scientific), 0.02% sodium azide (71448–16, Fisher Scientific), and 0.1% IGEPAL^® CA-630 in PBS] for 30 min at room temperature, cells were incubated with primary antibody listed in Table 5 in IF buffer for 3–4 hours at room temperature, followed by rinsing with IF buffer five times. The samples were then incubated with fluorescent dye-labeled secondary antibodies (listed in Table 5) in IF buffer for 1 hour at room temperature, followed by rinsing with IF buffer five times. After nuclear staining with 4’,6-diamidino-2-phenylindole (DAPI) (D9542–1MG, SIGMA) in IF buffer at a final concentration of 0.5 μg/ml, coverslips were mounted with Fluoromount-G (0100–01, SouthernBiotech) onto glass slides (22–339-411, Epredia). Images were acquired on a Nikon Ti2 inverted fluorescence microscope equipped with a Fusion BT camera (Model: C15440–20UP, Hamamatsu Photonics), 405 nm, 488 nm, 561 nm, and 642 nm lasers, and CSU-W1 SoRa Confocal Scanner Unit. A 20x NA0.75 planapochromat objective lens (MRD00205, Nikon) with the spinning disk confocal was used for mitosis arrest assays. A 60x NA1.4 plan-apochromat Lambda objective lens (MRD01605, Nikon) with a spinning-disk confocal was used for Golgi dispersal and transferrin internalization analyses. A 100x NA1.49 SR HP apochromat TIRF objective lens (MRD01995, Nikon) with the SoRa module was used for super-resolution transferrin internalization analysis. All the raw image data are available through BioImage Archive (Accession ID: S-BIAD2849).

Mitosis arrest assays and Golgi dispersal assays

Cells were plated into a 6-well plate at a density of 2 × 10⁵ cells/well and grown for 48 hr. After treatment with the indicated compound or DMSO for 3-, 6-, 9-, 12-, 24-hr, all the cells were fixed at the same time in 4% paraformaldehyde diluted in PBS. After permeabilization, the cells used for mitotic arrest assays were stained with DAPI in IF buffer at a final concentration of 0.5 μg/ml for 5 min. The cells used for Golgi dispersal assays were stained with either anti-TGN46 antibody only, or anti-TGN46 antibody combined with anti-DYNC1H1 antibody, and anti-α-Tubulin. At least six images from different fields per sample were captured for typical analysis. Typically, roughly 200 cells were analyzed per experiment. The exact number of cells we analyzed in each sample is available in the corresponding Source data in the BioImage Archive (Accession ID: S-BIAD2849). The percentages of mitotically arrested cells and cells with dispersed Golgi were manually counted using the NIS-Elements software (Nikon).

Transferrin internalization assays

Cells were plated into an 8-well chambered coverslip (80807–90, iBidi) at a density of 5 × 10⁴ cells/well and grown for 48 hours. Cells were treated with HaloPROTAC3 or DMSO for 6 hours and were kept on ice for 10 min. After the cold treatment, the cells were incubated with 25 μg/ml Transferrin Alexa Fluor^™ 488 conjugate (T13342, Invitrogen) in complete media without phenol red (11–039-021, Gibco) at 37°C for 15 minutes and washed with PBS 3 times. All the 60x spinning disk confocal live cell images were captured in the CSU-W1 mode to show the inhibition of the transferrin internalization. The 100x SoRa super-resolution Z-stack live cell images were captured in the SoRa mode with the 2.8x magnifier with the 0.175 μm z-increments to ensure Nyquist sampling. Images were then computationally deconvolved using the NIS-Element software.

Data availability

All data generated or analyzed during this study are included in the manuscript and supporting files; source data files have been provided for all figures. All vectors have been deposited in Addgene (#87079). All immunoblot and immunofluorescent data have been deposited at BioImage Archive (Accession ID: S-BIAD2849).